Software development is a discipline that is almost as old as the history of computers. With the advent of the Internet and all of its related technologies, software development has been on high demand. But, and especially in SME (small and medium enterprise), this was not accompanied with a comparable effort to develop a set of sustainable and standardized activities of project management, which lead to increasing inefficiencies and costs.
Given the actual economic situation, it makes sense to engage in an effort to reduce said inefficiencies and rising costs. For that end, this work will analyze the current state of software development’s project management processes on a Portuguese SME, along with its problems and inefficiencies in an effort to create a standardized model to manage software development, with special attention given to critical success factors in an agile software development environment, while using the best practices in process modeling. This work also aims to create guidelines to correctly integrate these changes in the existing IS structure of a company.
Original Title
Master's Thesis - Standardization of project management processes
Software development is a discipline that is almost as old as the history of computers. With the advent of the Internet and all of its related technologies, software development has been on high demand. But, and especially in SME (small and medium enterprise), this was not accompanied with a comparable effort to develop a set of sustainable and standardized activities of project management, which lead to increasing inefficiencies and costs.
Given the actual economic situation, it makes sense to engage in an effort to reduce said inefficiencies and rising costs. For that end, this work will analyze the current state of software development’s project management processes on a Portuguese SME, along with its problems and inefficiencies in an effort to create a standardized model to manage software development, with special attention given to critical success factors in an agile software development environment, while using the best practices in process modeling. This work also aims to create guidelines to correctly integrate these changes in the existing IS structure of a company.
Software development is a discipline that is almost as old as the history of computers. With the advent of the Internet and all of its related technologies, software development has been on high demand. But, and especially in SME (small and medium enterprise), this was not accompanied with a comparable effort to develop a set of sustainable and standardized activities of project management, which lead to increasing inefficiencies and costs.
Given the actual economic situation, it makes sense to engage in an effort to reduce said inefficiencies and rising costs. For that end, this work will analyze the current state of software development’s project management processes on a Portuguese SME, along with its problems and inefficiencies in an effort to create a standardized model to manage software development, with special attention given to critical success factors in an agile software development environment, while using the best practices in process modeling. This work also aims to create guidelines to correctly integrate these changes in the existing IS structure of a company.
Subttulo TTULO Nome completo do Candidato Subttulo Dissertao / Trabalho de Projeto / Relatrio de Estgio apresentada(o) como requisito parcial para obteno do grau de Mestre em Estatstica e Gesto de Informao
TTULO Nome completo do Candidato Subttulo Dissertao / Trabalho de Projeto / Relatrio de Estgio apresentada(o) como requisito parcial para obteno do grau de Mestre em Estatstica e Gesto de Informao
Standardization of project management processes Miguel Augusto Martins Pereira An inside view on changes to achieve efficiency Work Project presented as partial requirement to obtain the Masters degree in Information Management
ll lnsLlLuLo Superlor de LsLaLlsLlca e CesLo de lnformao unlversldade nova de Llsboa
"#$%&$'&()$#(*% *+ ,'*-./# 0$%$1.0.%# ,'*/.""." by Mlguel AugusLo MarLlns erelra
Work ro[ecL presenLed as parLlal requlremenL Lo obLaln Lhe MasLer's degree ln lnformaLlon ManagemenL, speclallzaLlon ln ManagemenL of lnformaLlon SysLems and 1echnologles
CoordlnaLor: Ales opovlc
november 2013 lll &.&(/$#(*%" 1o Lhe ones who are noL among us anymore.
lv $/2%*34.&1.0.%#" 1here are a number of persons LhaL l would llke Lo acknowledge for Lhelr help ln wrlLlng Lhls work. WlLhouL Lhem, Lhls work would noL be posslble. loremosL, l would llke Lo express my graLlLude Lo my advlsor, Ales opovlc for hls guldance and asslsLance ln deflnlng Lhe guldellnes of Lhls work LhaL helped sLeer lL for a Llmely concluslon. Secondly, l also would llke Lo acknowledge my deepesL and slncere appreclaLlon Lo Lhe supporL glven by Susana Cllvelra, noL only for Lhe endless nlghLs spenL proof- readlng Lhls work and Lhe much appreclaLed lnpuL ln every sLep of Lhls work, buL also for Lhe backup glven on every aspecL of normal llfe, helplng ln maklng Lhls work whaL lL ls Loday. l also wlsh Lo Lhank my famlly, speclflcally my moLher, faLher, grandparenLs and slbllngs, buL also all Lhe oLher famlly members LhaL help me ln Lhls long perlod necessary Lo obLaln Lhe masLer's degree. A speclal recognlLlon Lo my co-workers aL u8l LhaL were klnd Lo glven Lhelr Llme ln answerlng my quesLlons, namely: Claudlo Calvo, aulo MrLlres and uora 8allo. Also l would llke Lo Lhank ulogo Cabral and Lhe Melos, SaleLe and 8lcardo for Lhelr help, proof-readlng and valuable lnpuLs. Llkewlse, l would llke Lo glve a speclal appreclaLlon Lo my Leam-members LhaL gave an lmporLanL feedback by means of Lhelr dally work. LasLly, l would llke Lo send my appreclaLlon Lo Lhe frlends LhaL read my work ln lLs several sLages and gave an lmporLanL conLrlbuLe Lo lL, buL also Lo Lhe frlends LhaL helped [usL by llsLenlng.
v "500$'6 1he prlmary focus of Lhls work ls Lo provlde a framework of sLandardlzed pro[ecL managemenL processes developed Lhrough Lhe analysls of Lhe processes of a orLuguese Small and Medlum LnLerprlse (SML), ln order Lo help address cosL consumlng lssues LhaL lead Lo lnefflclency on a company. AfLer Lhe ldenLlflcaLlon of Lhese lssues occurrlng ln Lhe sub[ecL company, Lhe ob[ecLlve ls Lo deLermlne Lhe besL pracLlces of pro[ecL managemenL based on a LheoreLlcal background, ln order Lo develop a sLandard ln pro[ecL managemenL acLlvlLles. 1o accompllsh Lhls, Lhe pro[ecL managemenL workflow ls re-modeled, whlle keeplng measuremenL acLlvlLles Lo ascerLaln lLs efflclency and valldlLy. 1he flnal parL of Lhls work ls a proof of concepL on how Lo lnLegraLe Lhese workflows ln an exlsLlng lnformaLlon SysLem sLrucLure and also lL glves a serles of guldellnes on how Lo lnLroduce Lhese changes on a company.
vl $7"#'$/# SofLware developmenL ls a dlsclpllne LhaL ls almosL as old as Lhe hlsLory of compuLers. WlLh Lhe advenL of Lhe lnLerneL and all of lLs relaLed Lechnologles, sofLware developmenL has been on hlgh demand. 8uL, and especlally ln SML (small and medlum enLerprlse), Lhls was noL accompanled wlLh a comparable efforL Lo develop a seL of susLalnable and sLandardlzed acLlvlLles of pro[ecL managemenL, whlch lead Lo lncreaslng lnefflclencles and cosLs. Clven Lhe acLual economlc slLuaLlon, lL makes sense Lo engage ln an efforL Lo reduce sald lnefflclencles and rlslng cosLs. lor LhaL end, Lhls work wlll analyze Lhe currenL sLaLe of sofLware developmenL's pro[ecL managemenL processes on a orLuguese SML, along wlLh lLs problems and lnefflclencles ln an efforL Lo creaLe a sLandardlzed model Lo manage sofLware developmenL, wlLh speclal aLLenLlon glven Lo crlLlcal success facLors ln an aglle sofLware developmenL envlronmenL, whlle uslng Lhe besL pracLlces ln process modellng. 1hls work also alms Lo creaLe guldellnes Lo correcLly lnLegraLe Lhese changes ln Lhe exlsLlng lS sLrucLure of a company.
2.63*'&" SofLware developmenL, CrlLlcal success facLors (CSl), ro[ecL managemenL, Aglle, Scrum, 8uslness process managemenL (8M) besL pracLlces, ManagemenL of lnformaLlon sysLems (MlS) vll (%&.8 1. lnLroducLlon ................................................................................................................. 1 1.1. ConLexLuallzaLlon and problem ldenLlflcaLlon ...................................................... 1 1.2. Cb[ecLlves ueflnlLlon ............................................................................................ 3 1.3. 1he lmporLance of Lhe work ................................................................................. 3 1.4. Work's SLrucLure ................................................................................................... 4 2. 1heoreLlcal 8ackground ............................................................................................... 3 2.1. MeLhodology ........................................................................................................ 3 2.2. ro[ecL managemenL ............................................................................................ 8 2.2.1. CrlLlcal success facLors ................................................................................... 8 2.3. 8uslness process managemenL ........................................................................... 10 2.3.1. lmplemenLaLlon of Lhe 8M framework ..................................................... 11 2.3.2. CrlLlcal success facLors and relaLed besL pracLlce ........................................ 12 2.3.3. MeLrlcs for measurlng processes ................................................................. 13 2.4. lnformaLlon sysLems adopLlon ........................................................................... 18 2.4.1. AdopLlon of changes .................................................................................... 19 3. llndlngs ...................................................................................................................... 22 3.1. AbouL Lhe company ............................................................................................ 22 3.1.1. ro[ecL managemenL meLhodology aL u8l .................................................. 23 3.1.2. LnLerprlse ro[ecL ManagemenL ................................................................. 24 3.2. lnLervlews analysls .............................................................................................. 26 3.2.1. CommunlcaLlon, documenLaLlon and requlremenLs' assessmenL problems 26 3.2.2. Lack of conLrol on pro[ecL sLaLus and furLher analysls problems ................ 27 3.3. Modellng Lhe processes ...................................................................................... 28 3.3.1. ulscovery process ........................................................................................ 29 3.3.2. ueflnlLlon process ........................................................................................ 33 3.3.3. uevelopmenL process .................................................................................. 37 3.3.4. uemo process .............................................................................................. 43 3.3.3. uellvery process ........................................................................................... 46 3.4. CbLalnlng Lhe MeasuremenLs ............................................................................. 32 4. ulscusslon .................................................................................................................. 33 4.1. 1heoreLlcal lmpllcaLlons ..................................................................................... 33 vlll 4.2. racLlcal lmpllcaLlon ........................................................................................... 34 4.3. LlmlLaLlons and recommendaLlons ..................................................................... 33 4.4. luLure work ........................................................................................................ 37 3. Concluslons ................................................................................................................ 39 6. 8lbllography ............................................................................................................... 60 7. Appendlx .................................................................................................................... 63 7.1. lnLervlew LranscrlpLs ........................................................................................... 63 7.1.1. lnLervlew 1 ................................................................................................... 63 7.1.2. lnLervlew 2 ................................................................................................... 82 7.1.3. lnLervlew 3 ................................................................................................... 93
lx (0$1. (%&.8 lmage 1 - 8equlremenLs assessmenL from Moe eL al. (2012) .......................................... 9 lmage 2 - 8M framework ............................................................................................. 12 lmage 3 - CompleLe process workflow for a pro[ecL ...................................................... 29 lmage 4 - ulscovery process .......................................................................................... 30 lmage 3 - lnformaLlon sLrucLure of Lhe ulscovery process ............................................ 32 lmage 6 - ueflnlLlon process .......................................................................................... 33 lmage 7 - lnformaLlon sLrucLure of Lhe ueflnlLlon process ............................................ 36 lmage 8 - uevelopmenL process .................................................................................... 38 lmage 9 - 8urn-down charL ............................................................................................ 41 lmage 10 - lnformaLlon sLrucLure of Lhe uevelopmenL process .................................... 42 lmage 11 - uemo process .............................................................................................. 43 lmage 12 - lnformaLlon sLrucLure of Lhe uemo process ................................................ 43 lmage 13 - uellvery process ........................................................................................... 47 lmage 14 - lnformaLlon sLrucLure of Lhe uellvery process ............................................ 31
x #$74. (%&.8 1able 1 - CuesLlons for Lhe C ......................................................................................... 6 1able 2 - CrlLlcal success facLors for buslness process managemenL. ........................... 13 1able 3 - lacLors LhaL help measurlng processes .......................................................... 16 1able 4 - Common pro[ecL Lasks .................................................................................... 26 1able 3 - ro[ecL raLlng lndlcaLors .................................................................................. 31 1able 6 - lnformaLlon needed for a sLory ....................................................................... 33
xl 4("# *+ $77'.9($#(*%" 7,0: 8uslness rocess ManagemenL. An approach on managlng Lhe buslness processes of a company Lo lmprove Lhelr efflclency ln order Lo dellver beLLer value Lo Lhe cllenL. 7,0%: 8uslness rocess Model and noLaLlon. A noLaLlon of graphlcal represenLaLlon Lo model a workflow of a process. /'0: CusLomer 8elaLlonshlp ManagemenL. A serles of pracLlces supporLed by an lnformaLlon Lechnology lnfrasLrucLure LhaL manages Lhe currenL and fuLure cusLomers (Lhrough Lhe use of commerclal leads). /"+: CrlLlcal Success lacLor. necessary sLep Lo achleve success ln Lhe relaLed endeavor. .,0: LnLerprlse ro[ecL ManagemenL. An lnformaLlon plaLform Lo manage and monlLor Lhe company's pro[ecLs. (": lnformaLlon SysLem. An aggregaLlon Lerm LhaL encompasses lnformaLlon Lechnology, people and Lhelr processes. 0(": ManagemenL lnformaLlon SysLem. An lnformaLlon Lechnology sysLem Lo manage all Lhe componenLs of an lnformaLlon sysLem. *#.: Crlglnal 1lme LsLlmaLe. An esLlmaLlon glven by Lhe C for a funcLlonallLy of a pro[ecL. ,*: roducL Cwner. ln Scrum lL ls Lhe responslble Lo ensure LhaL Lhe pro[ecL dellvers Lhe lnLended producL Lo Lhe cllenL. '#.: 8emalnlng 1lme LsLlmaLe. An esLlmaLlon glven by Lhe developmenL Leam Lo an lssue of a sLory. "0.: Small and Medlum LnLerprlse. A deflnlLlon of a company followlng cerLaln parameLers llke number of employees and Lurnover or balance sheeL LoLal.
1 ;< (%#'*&5/#(*% ldeally, sLrucLured pro[ecL managemenL processes are essenLlal Lo correcLly manage a pro[ecL, however Lhe managemenL of Lhese processes ls Lyplcally overlooked ln favor of a qulcker expedlLlon of a pro[ecL's producL, especlally on Small and Medlum LnLerprlse (SML), where resources llke Llme, knowledge and money are noL as avallable as ln larger companles. neverLheless, for companles ln general and SMLs ln parLlcular, one way Lo reslsL Loday's economlc crlsls ls Lrough efflclenL work, meanlng LhaL Lhelr lnLernal processes should be properly sLrucLured and carefully allgned wlLh Lhelr (sLraLeglc) goals. 1he goal of Lhls work ls Lo provlde a framework LhaL sLandardlzes sofLware developmenL pro[ecL managemenL processes. 1o do so, an analysls was conducLed on a orLuguese SML whlch feaLured some problems on Lhe level of pro[ecL managemenL ln order Lo obLaln flndlngs LhaL supporL Lhe new proposed framework. A focus ls glven Lo SMLs slnce Lhey are Lhe bulk of noL only orLugal's enLerprlse fabrlc (lnL, 2010), buL also ln Lhe Lu (Wymenga, Spanlkova, 8arker, konlngs, & CanLon, 2012), and LogeLher Lhey consLlLuLe Lhe ma[or employer of Lhe global acLlve populaLlon and are vlewed as a ma[or source of lnnovaLlon on Lhelr flelds. ;<;< /*%#.8#5$4()$#(*% $%& ,'*74.0 (&.%#(+(/$#(*% ln order Lo obLaln Lhe basls Lo formulaLe a conslsLenL framework of sLandardlzed pro[ecL managemenL processes, an analysls of lssues concernlng pro[ecL managemenL was conducLed ln a orLuguese SML. 1he company, u8l, works ln Lhe fleld of sofLware developmenL and lLs sLrucLure can be ldenLlfled as a flaL sLrucLure (Llm, CrlfflLhs, & Sambrook, 2010) because Lhere ls a low formallzaLlon of Lhe buslness processes, low level of cenLrallzaLlon of declslon maklng and Lhere ls noL a culLure of documenLaLlon, l.e. lnserLlng Lhe lnpuL on Lhe sLaLus of a pro[ecL, or Lhe Llme spenL on each of Lhe Lasks of Lhe pro[ecL. ln Lhe beglnnlng, when Lhe company had few workers, Lhls was a manageable slLuaLlon. As Lhe workers were on Lhe same pro[ecL, all of Lhe lnformaLlon was (malnly) known by everyone. 8uL slnce Lhe company ls growlng ln slze, Lhe lnformaLlon Lends Lo be scaLLered, so some ad[usLmenLs ln Lerms of buslness processes have Lo be made, as many of Lhe processes are noL modeled or even sLrucLured. AddlLlonally, lL ls dlfflculL Lo dlssemlnaLe Lhe lnformaLlon due Lo a lack of hlerarchy on Lhe dlsLrlbuLlon of lnformaLlon.
2 WlLh Lhe company's growLh, Lhe managemenL declded Lo esLabllsh a common aglle pro[ecL managemenL meLhodology uslng Lhe Scrum framework and Lo use an LnLerprlse ro[ecL ManagemenL (LM) Lo help wlLh pro[ecL managemenL acLlvlLles. neverLheless, havlng Lhe Lools Lo do Lhe [ob ls only half Lhe work, as lL ls sLlll requlred Lo deLermlne Lhe correcL appllcaLlon of Lhese Lools Lo beLLer beneflL from Lhem. ln Lhls way, Lhe followlng lssues were ldenLlfled: ! 1here ls noL a common seL of processes Lo manage Lhe pro[ecLs. Lach pro[ecL manager, or roducL Cwner (C - whlch ls an equlvalenL role of a pro[ecL manager ln Lhe pro[ecL managemenL meLhodology adopLed by Lhe company) has lLs own meLhods, Laklng dlfferenL sLeps ln each pro[ecL, ! 1here ls noL a sLandard procedure Lo gaLher and sLore Lhe lnformaLlon relaLed wlLh Lhe pro[ecL. resenLly, Lhere are many programs avallable Lo be used for LhaL purpose, buL ln some slLuaLlons none ls used. 1here ls noL a cenLral sysLem Lo sLore all Lhe perLlnenL lnformaLlon of a pro[ecL LhaL ls needed Lo supporL lLs developmenL, ! 1he flnanclal and commerclal deparLmenLs do noL have a clear vlew of Lhe acLual sLaLus of Lhe pro[ecLs belng worked on, prevenLlng Lhe flnanclal deparLmenL Lo lssue on Llme paymenL requesLs and Lhe commerclal area Lo manage Lhe expecLaLlons of upcomlng cusLomers, ! ln Lhe operaLlonal area, Lhe developers also suffer wlLh Lhe lack of lnformaLlon abouL Lhe pro[ecL, because, slnce Lhey do noL have a clear plcLure Lhey cannoL be asserLlve and comply wlLh deadllnes, ! Whlle Lhere ls some degree of Lracklng Lhe Llme spenL on Lhe Lasks performed each day agalnsL prlor esLlmaLlons, Lhls lnformaLlon ls noL belng used Lo lmprove Lhe meLhods of plannlng and lLs esLlmaLlons. uue Lo Lhe lack of an esLabllshed seL of pro[ecL managemenL processes, a company can suffer from Lhe lssues ldenLlfled above whlch reflecL on Lhe company's performance and make Lhe developmenL process less effecLlve ln Lerms of Llme and resources. uslng Lhe daLa obLalned from u8l, lL ls Lhe auLhor's bellef LhaL, LogeLher wlLh an asserLlve LheoreLlcal background, lL ls posslble Lo devlse a good foundaLlon Lo bulld a sLandardlzed framework for managlng sofLware developmenL pro[ecLs and Lhus obLaln a seL of Lools and processes Lo ellmlnaLe or mlLlgaLe lssues occurrlng ln a company, resulLlng ln a beLLer accounLablllLy for Lhe cllenLs and for Lhe company lLself.
3 ;<=< *7-./#(9." &.+(%(#(*% Many of Lhe lssues prevlously polnLed ouL may be (somewhaL) addressed prompLly, buL LhaL should noL be Lhe norm whlle lmplemenLlng change. 1aklng Lhe example of whaL happened aL u8l, ln Lhe pasL, when a problem was found lL was resolved sLralghL away, whlle lL was poslLlve on a shorL Lerm Lo paLch a problem, noL Lhlnklng and deflnlng a sLraLegy on Lhe long Lerm [usL posLponed Lhe resoluLlon of Lhe problem lLself. As such, Lhe ob[ecLlves for Lhls work are Lo: ! uncover Lhe speclflc problems relaLed wlLh pro[ecL managemenL processes LhaL occur ln a company, ! ldenLlfy and model new pro[ecL managemenL processes Lo solve Lhe problems ldenLlfled, ! rovlde meLhods Lo allow Lhe managers, whlch are Lhe maln reclplenLs of Lhls work, Lo follow more closely Lhe sLaLus of a pro[ecL, ! Crganlze a common sLrucLure Lo sLore Lhe lnformaLlon relaLed Lo a pro[ecL, so lL may be avallable Lo whoever needs lL. CaLherlng from all of Lhese polnLs, Lhe overall ob[ecLlve ls Lo bulld a framework LhaL allows a company Lo lmprove Lhelr pro[ecL managemenL processes and provlde Lhe Lools and relevanL lnformaLlon noL only Lo Lhe pro[ecL managers, buL also Lo all of Lhe parLlclpanLs of a pro[ecL Lhus allowlng Lhem Lo be more effecLlve on Lhelr work. AlLhough Lhls work focuses on Lhe use of Scrum lL ls noL lLs ob[ecLlve Lo uphold Lhe use of a speclflc meLhodology, buL Lo defend a broader approach on lmplemenLlng aglle meLhodologles. ;<>< #?. (0,*'#$%/. *+ #?. 3*'2 8y employlng Lhe framework proposed ln Lhls work, a company can achleve: ! A unlfled seL of pro[ecL managemenL processes Lo allow all of Lhe company's managers Lo manage developmenL pro[ecL ln a coherenL fashlon and sLore all Lhe necessary lnformaLlon ln a cenLrallzed locaLlon, ! A clear plcLure of Lhe lnlLlal esLlmaLlons and Lhe real cosL ln Lerms of Llme used on Lhe pro[ecL, ! valldaLlon of Lhe work belng done ln Lhe pro[ecL Lhrough Lhe use of meLrlcs, ! CerLalnLy LhaL Lhe developmenL of producLs can be made wlLh accounLable and documenLable quallLy.
4 ;<@< 3*'2A" "#'5/#5'. 1he sLrucLure of Lhls work ls as follow: ln Lhe lnLroducLlon Lhe problems LhaL Lhls work ls almed Lo solve are ldenLlfled, as well as Lhe work's ob[ecLlves and goals. AfLerwards Lhe LheoreLlcal background secLlon ls presenLed, lncludlng Lhe meLhodology employed and a Lhorough presenLaLlon on Lhe research done for Lhls work. 1hls ls followed by Lhe presenLaLlon of lssues found Lhrough a seL of lnLervlews Lo Lhe Cs and a group of proposals Lo solve Lhe lssues ldenLlfled based on Lhe lnLervlews and lndependenL observaLlon. nexL lL ls presenLed a dlscusslon from a LheoreLlcal and pracLlcal polnL of vlew, LogeLher wlLh Lhe llmlLaLlons of Lhls work. Moreover, some addlLlonal recommendaLlons and some noLes of posslble works ln Lhe fuLure are made. AL Lhe end, concluslons abouL Lhls work are drawn LogeLher wlLh a dellberaLlon of Lhe paLh Lhls work proposes for companles.
3 =< #?.*'.#(/$4 7$/21'*5%& =<;< 0.#?*&*4*16 lor Lhls work, varlous meLhods of gaLherlng lnformaLlon were used. llrsLly observaLlon, slnce Lhe auLhor works on Lhe sub[ecL company of Lhls sLudy, day by day observaLlons prompLed several lnslghLs LhaL lead Lo Lhe nexL meLhod: lnLervlews. uslng Lhese observaLlons, a serles of quesLlons was devlsed (1able 1) and a serles of lnLervlews ensued wlLh Lhe Cs of pro[ecLs on Lhe company.
%BCDEF GBEHIJKLH *DMENIJOEH 1 Who supplles Lhe lnformaLlon relaLed Lo new pro[ecLs? 1o deLermlne where Lhe flow of lnformaLlon sLarLs. 2 WhaL ls Lhe Lyplcal workflow of a pro[ecL, from Lhe pro[ecL managemenL polnL of vlew? 1o deLermlne Lhe currenL workflow of a pro[ecL, from each of Lhe Cs ln order Lo assess Lhe common Lasks. 3 ls every pro[ecL LreaLed dlfferenLly, or do Lhey share some common Lasks / sLeps? 1o deLermlne lf Lhe C has some sLrucLure ln hls / her process. 4 lf yes, can you ldenLlfy Lhem? CbLaln Lhe sLrucLure. 3 Can Lhese common Lasks / sLeps be grouped by Lype of pro[ecL or by anoLher caLegory? ln poslLlve case, ls LhaL grouplng currenLly worklng? ueLermlne lf Lhere are Lasks LhaL can be grouped and deLermlne Lhe crlLerla Lo do so. 6 8esldes Lhe communlcaLlon LhaL occurs wlLh Lhe developers, how ls Lhe communlcaLlon wlLh Lhe oLher deparLmenLs of Lhe company made? 1o deLermlne how Lhe communlcaLlon, lf any, ls made. 7 WhaL are Lhe mosL common dlfflculLles LhaL you Lyplcally encounLer? 1o obLaln a llsL of common problems already ldenLlfled. 8 uo you have any suggesLlon Lo mlLlgaLe Lhem? 1ry Lo obLaln some soluLlons LhaL could be used ln Lhe flnal soluLlon. 9 uo you currenLly compare Lhe esLlmaLe Llmes done prevlously wlLh Lhe acLual Llme spenL on pro[ecL lssues? 1ry Lo deLermlne lf a comparlson ls made, and whaL are Lhe resulLs.
6 %BCDEF GBEHIJKLH *DMENIJOEH 10 Why or why noL? lL ls expecLed LhaL Lhe C does noL compare Lhe resulLs. 1hls wlll help Lo deLermlne Lhe cause and lmplemenL Lhe soluLlon. 11 uo you have any ldea of whlch are Lhe acLlvlLles belng consLanLly underesLlmaLed? 1ry Lo deLermlne lf Lhe C has any ldea whlch acLlvlLles are more problemaLlc. 12 ls Lhere any parameLer deflned LhaL allows deLermlnlng lf a pro[ecL ls a success or noL? 1o deLermlne lf Lhe C have any CSls for pro[ecLs. 13 lf so, whlch are Lhey? CbLaln Lhe llsL of CSl of Lhe C. 14 lf CSls are deflned, ls Lhere any ldea whlch are Lhe maln problems LhaL lead Lo a pro[ecL noL belng consldered a success? Also Lo Lry Lo deLermlne lf Lhere ls lnformaLlon on whlch of Lhe CSls are Lhe maln reasons for a fall pro[ecL. 13 And ls lL a problem LhaL occurs frequenLly? ls anyLhlng belng done Lo quench Lhls problem and why? ueLermlne soluLlons Lo Lhe problem, lf Lhere are problems ln Lhe currenL LnLerprlse ro[ecL ManagemenL (LM). 16 Were Lhere any CSls deflned for pro[ecLs? ls lL posslble Lo group Lhem? lurLher lnformaLlon gaLherlng Lo assess Lhe sLaLus of CSls of pro[ecLs. 17 1he new 1 LM ls already lmplemenLed. WhaL do you make of lL? Any suggesLlons Lo lmprove lL? Pave you glven some lnpuL when lL was lmplemenLed? CaLher lnformaLlon on Lhe lssues wlLh Lhe new LM sysLem. 1o deLermlne lf Lhe Cs where lncluded ln Lhe rollouL of Lhe sysLem. 18 ln whaL sLage of Lhe lmplemenLaLlon do you Lhlnk Lhe LM ls? 1o deLermlne Lhe level of lnvolvemenL of Lhe C ln Lhe lmplemenLaLlon of Lhe LM sysLem. 19 Cf Lhe Lasks you perform regularly, do you Lhlnk Lhey could be done more efflclenLly uslng Lhe LM? 1o deLermlne lf Lhe LM sysLem ls belng used Lo help Lhe C, lf lL flLs Lhe C's needs. 20 lf so, ls Lhere any exLra Lool LhaL you requlre Lo achleve a beLLer efflclency? ueLermlne Lhe needs of Lhe C ln hls / her dally Lasks. 1able 1 - CuesLlons for Lhe C
1 8ecenLly lmplemenLed aL Lhe Llme Lhe quesLlons were done.
7 AlLhough Lhe presenL llsL ls long, lL ls belleved LhaL lL was necessary Lo ascerLaln Lhe lnformaLlon needed Lo Lle up wlLh Lhe lnformaLlon gaLhered ln Lhe revlew of relevanL llLeraLure. Also, Lhe lnLervlews on whlch Lhese quesLlons were used were esLlmaLed Lo Lake abouL LhlrLy mlnuLes, a perlod LhaL ls consldered adequaLe. WlLh Lhe lnformaLlon gaLhered from Lhe lnLervlews lL was posslble Lo assess Lhe problems ln pro[ecL managemenL, from Lhe Cs polnL of vlew. 1hese resulLs were [uxLaposed and crossed examlned wlLh maLerlal gaLhered from anoLher meLhod: relaLed llLeraLure analysls from Lhe relevanL works regardlng besL pracLlces ln pro[ecL managemenL, namely crlLlcal success facLors Lo effecLlvely lmplemenL aglle sofLware developmenL pro[ecLs llke Scrum. 8y Lylng LogeLher Lhls lnformaLlon, lL was posslble Lo deLermlne whlch of Lhe ldenLlfled problems on Lhe model company should be focused for Lhls work. 8ecause Lhe maln problem ldenLlfled ln Lhe company was noL havlng a sLrucLured seL of processes, Lhe nexL loglcal sLep was Lhe research of correcL modellng of buslness processes Lhrough Lhe use of a 8M framework gulded by Lhe crlLlcal success facLors besL sulLed for Lhe company. AnoLher polnL of focus would be Lo acqulre Lhe necessary lnformaLlon Lo measure (Lhrough Lhe use of meLrlcs) Lhe effecLlveness of Lhe work belng done on each pro[ecL. llnally, Lhe besL pracLlces Lo lmplemenL and manage Lhese changes on Lhe lnformaLlon SysLem of Lhe company were revlewed, maklng Lhe lnformaLlon avallable Lo everyone who needs lL, when lL ls needed and followlng a common sLrucLure LhaL can be undersLandable for everyone ln every deparLmenL. As such lL ls posslble Lo dlvlde Lhls work ln slx separaLe phases: hase l - Analysls of relevanL llLeraLure for Lhe pro[ecL managemenL, buslness processes and managemenL of lnformaLlon sysLem Loplcs, hase ll - lormulaLlon of quesLlons, based on Lhe observaLlon on Lhe dally acLlvlLles of Lhe company regardlng pro[ecL managemenL, hase lll - lnLervlews wlLh Lhe company's pro[ecL managers (each aL a Llme), hase lv - Analysls of Lhe lnLervlews and [uxLaposlLlon wlLh relevanL lnformaLlon already gaLhered, hase v - ueslgn Lhe prevlously ldenLlfled key buslness processes and deLermlne Lhe relaLed meLrlcs, hase vl - Analysls on how Lo use Lhe LM Lo obLaln Lhe daLa of Lhe deslgned meLrlcs, how Lo sLrucLure Lhe lnformaLlon ouLpuLLed by Lhe processes and lf Lhe LM provldes Lhe necessary lnformaLlon ouL of Lhe box or lf some oLher lmplemenLaLlons
8 are requlred. 1hls wlll lead Lo a seL of guldellnes for Lhe correcL use of Lhe lS Lo sLore and consulL lnformaLlon relaLed wlLh a pro[ecL, ranglng from pro[ecL sLaLus Lo pro[ecL meLrlcs. =<=< ,'*-./# 0$%$1.0.%# Cne of Lhe cornersLones of Lhls work ls Lhe con[uncLlon beLween an analysls of crlLlcal success facLors of managlng sofLware developmenL pro[ecLs uslng an aglle meLhodology and an analysls of Lhe percelved problems occurrlng ln Lhe model company by Lhe producL owners. ln Lhe followlng chapLer, Lhls work wlll focus on undersLandlng Lhe role crlLlcal success facLors can have when managlng pro[ecLs especlally Lhe ones relevanL Lo Lhe problems Lhe sald company presenLs. =<=<;< /FJIJNPQ HBNNEHH RPNIKFH Chow and Cao (2008), clLlng 8ullen and 8ockarL (1981) descrlbe crlLlcal success facLors as Lhe llmlLed number of areas ln whlch saLlsfacLory resulLs wlll ensure successful compeLlLlve performance for Lhe lndlvldual, deparLmenL, or organlzaLlon". 1hls work's goal ls exacLly such: Lo dlscover whaL Lhe areas where more focus ln resolvlng lssues musL be glven, so Lhe processes of pro[ecL managemenL can be more efflclenL. Cf Lhe approaches LhaL could have been done aL Lhls polnL, lL was chosen Lo revlew a broad selecLlon of relevanL llLeraLure Lo ldenLlfy several facLors LhaL are plvoLal for a successful managemenL of pro[ecLs ln order Lo supporL Lhe lssues reporLed ln Lhe lnLervlews. lL ls worLh noLlng LhaL Lhe ldenLlfled facLors are conslsLenL wlLh Lhe prlnclples behlnd Lhe aglle manlfesLo (lowler & PlghsmlLh, 2001). !"!"#"#" $%&&'()*+,)%( .+*,%/0 CommunlcaLlon ls one of Lhe bases of Lhe aglle manlfesLo. 1hls facLor ls presenL ln many of Lhe selecLed llLeraLure. ln Sheffleld and LemeLayer (2013), communlcaLlon ls presenLed as an lmporLanL facLor LhaL leads Lo a successful lmplemenLaLlon of aglle pro[ecLs. 8uL ln Lhe sLudy of Mlsra, kumar, and kumar (2009), when adopLlng an aglle meLhodology on pro[ecL managemenL, communlcaLlon was noL deLermlned as an lmporLanL facLor llnked Lo success. neverLheless, for Chow and Cao (2008), communlcaLlon plays an lmporLanL role and was one of Lhe aLLrlbuLes ln Lhe pro[ecL managemenL process facLor LhaL appears ln Lhelr survey resulLs. CommunlcaLlng ls a key for passlng lmporLanL lnformaLlon and
9 knowledge ln a rlch aglle envlronmenL, granLlng a boosL ln Lhe posslblllLy of a successful lmplemenLaLlon of Lhe pro[ecL. As Lhls facLor ls one of Lhe problems ldenLlfled ln Lhe lnLervlews Lo Lhe Cs and has such hlgh lmporLance ln Lhe llLeraLure, a soluLlon wlll be proposed Lo solve Lhls lssue. !"!"#"!" 123')/2&2(,04 +00200&2(, 1he assessmenL of Lhe pro[ecL's requlremenLs ls an lmporLanL facLor ln every pro[ecL managemenL meLhodology. Powever, conLrary Lo older pro[ecL managemenL meLhodologles, Lhe lnvolvemenL of everyone ln Lhe pro[ecL ls recommended ln aglle meLhodologles, as Lhe shared experlence can help Lhe declslon maklng on lmporLanL sLeps of pro[ecLs. lor Moe, Aurum, and uyb (2012), noL lnvolvlng Lhe Leam prevenLs Lhe usage of Lhelr knowledge ln Lhe early sLages, leadlng Lhe C Lo make declslons unsupporLed, wlLhouL comprehenslve undersLandlng, whlch may have severe repercusslon ln Lhe developmenL of Lhe producL.
lmage 1 - 8equlremenLs assessmenL from Moe eL al. (2012)
As polnLed ouL ln Lhe lnLervlews, noL assesslng Lhe requlremenLs LogeLher as a whole (C and Lhe developmenL Leam) ls a rlsk LhaL makes Lhe pro[ecL's success much more dependenL on Lhe experlence and know-how of Lhe C, slnce he alone wlll be dolng much of Lhe sLraLeglc and LacLlcal declslons. !"!"#"5" 1)67, +&%'(, %. 8%*'&2(,+,)%( 1he aglle manlfesLo (lowler & PlghsmlLh, 2001) posLulaLes a Worklng sofLware over comprehenslve documenLaLlon", relylng on Lhe LaclL knowledge presenL on Lhe Leam's members and Lhelr lnLeracLlons Lo share Lhe knowledge raLher Lhan on expllclL
10 knowledge presenL on Lhe masslve documenLaLlon whlch ls defended on Lhe more LradlLlonal pro[ecL managemenL meLhodologles. neverLheless, documenLaLlon ls sLlll very much necessary, as polnLed ouL ln Lhe lnLervlews, aL Lhe level of pro[ecL managemenL and also aL Lhe level of developmenL. 1hls lasL case wlll only be referred marglnally, as Lhe focus of Lhls work ls on a managemenL level, buL as an example, Lhe use of documenLaLlon could be valuable Lo sLore lnformaLlon of developmenL declslons or guldlng maLerlal for Lhe demonsLraLlon of Lhe worklng proLoLype LhaL ls expecLed Lo be achleved aL Lhe end of each developmenL cycle, as Lhe pro[ecL managemenL framework chosen by Lhe company (Scrum) advocaLes. Also, Chow and Cao (2008) found ln Lhelr surveys done among Aglle professlonals LhaL Lhe rlghL amounL of documenLaLlon ls an lmporLanL facLor and should be Laken lnLo accounL. 1he paper does noL sLlpulaLe whaL LhaL amounL should be, whlch ls correcL, because Lhe level of documenLaLlon wlll always be varlable, dependlng on Lhe Lype of cllenL, Lype of pro[ecL and also dependlng on Lhe Leam and Lhelr experlence. uocumenLaLlon efforLs should be focused on Lasks llke ls posLulaLed by Llndvall eL al. (2004) and also 8oehm (2002) whlch defends LhaL process procedures LhaL lnvolve Lasks and mllesLone plans" should be documenLed. =<>< 75"(%."" ,'*/."" 0$%$1.0.%# 1he nexL focus of Lhls work ls an analysls of Lhe processes ln lLself and Lhe proposal of lmprovemenLs. AddlLlonally, Lhe deflnlLlon of meLrlcs LhaL wlll allow Lhe analysls of Lhe efflclency of sald lmprovemenLs wlll be soughL. 8uL Lo do LhaL, LheoreLlcal lnformaLlon ls needed Lo Lle up wlLh Lhe aforemenLloned analysls, speclflcally lnformaLlon abouL 8M Lechnlques. 1hese Lechnlques were chosen because, among all Lhe approaches LhaL could be lmplemenLed Lo lmprove Lhe pro[ecL managemenL buslness processes, 8M ls Lhe mosL comprehenslve, well-known, and wldely used pracLlce as 8ohloff (2009) posLulaLes and ls clLed by Skrln[ar and 1rkman (2013). 8M can have a loL of deflnlLlons, as Lhe one provlded by 8avesLeyn and 8aLenburg (2010) referenclng van uer AalsL, 1er PofsLede, and Weske (2003), ln whlch 8M ls a fleld of knowledge aL Lhe lnLersecLlon beLween 8uslness and lnformaLlon Lechnology, encompasslng meLhods, Lechnlques and Lools Lo analyze, lmprove, lnnovaLe, deslgn, enacL and conLrol buslness processes lnvolvlng cusLomers, humans, organlzaLlons, appllcaLlons, documenLs and oLher sources of lnformaLlon".
11 Powever, a deflnlLlon more flLLlng Lo Lhe scope of Lhls work ls LhaL 8M represenLs Lhe conLlnuous efforLs Lo analyze and lmprove Lhe company's processes llke Lhe ones of pro[ecL managemenL, developmenL, communlcaLlon, among oLhers (adapLed from 1rkman (2010)). 1he analysls and redeslgn of Lhe processes are done ln con[uncLure wlLh 8M crlLlcal success facLors and relaLed besL pracLlces found ln relevanL llLeraLure. AfLerwards, Lhe processes can be modeled Lo lmprove Lhem, respecLlng crlLlcal facLors of success. ConLrol meLrlcs are also deslgned Lo conLrol efflclency. =<><;< (CSQECELIPIJKL KR ITE 7,0 RFPCEUKFV ln Lhls work Lhe 8M framework referenced by Arlyachandra and lrollck (2008) ls used Lo help develop Lhe pro[ecL managemenL processes. 1hls framework ls composed by four core processes: "IFPIEWJXE - ln Lhls sLep, lL ls deLermlned whaL Lhe sLraLegy seL ouL by Lhe company ls. WlLh Lhls lnformaLlon lL can be posslble (alLhough lL ls sald by Arlyachandra and lrollck (2008) clLlng ollLano (2007) LhaL Lhls ls a challenglng Lask) Lo elaboraLe some meLrlcs Lo measure Lhe work belng done on Lhe processes. ,QPL - lL ls on Lhls sLep LhaL Lhe course of acLlon ls deflned and Lhe processes are deslgned as well as Lhe supporLlng acLlvlLles. 0KLJIKF - lL ls aL Lhls sLage LhaL Lhe lmporLance of Lhe 8M lmplemenLaLlon ls noLlced, slnce resulLs sLarL Lo appear aL Lhls polnL. 8ased on Lhe meLrlcs, measuremenL resulLs are avallable aL Lhe LM sysLem of Lhe company. WlLh Lhls lnformaLlon, Cs can plan addlLlonal acLlons, lf necessary. #PVE NKFFENIJOE PNIJKL - AL Lhls flnal sLep (before Lhe cycle sLarLs over), Lhe C deploys correcLlve acLlons when a problem arlses or when Lhe daLa avallable lndlcaLe LhaL problems could happen. An example relaLed wlLh Lhls work could be Lhe C sLarLs Lo noLlce LhaL Lhe Leam repeaLedly Lakes more Llme Lhan Lhe esLlmaLed, whlle compleLlng cerLaln Lasks. 1hls could lndlcaLe LhaL Lhe esLlmaLlons were under- budgeLed or LhaL Lhere ls a problem wlLh Lhe Leam and acLlon should be Laken.
12
lmage 2 - 8M framework lL ls worLh polnLlng ouL, LhaL Lhls looplng lLeraLlve framework works ln a slmllar fashlon as ln Lhe Scrum framework, Lhus lL makes sense puLLlng Lhese Lwo worklng Lechnlques LogeLher ln use, Lo achleve Lhe goals seL ouL for Lhls work. =<><=< /FJIJNPQ HBNNEHH RPNIKFH PLY FEQPIEY DEHI SFPNIJNE Whlle revlewlng relevanL llLeraLure, lL became apparenL LhaL many of sald llLeraLure presenLs a serles of common CSls Lo all of Lhem, buL sLlll Lhey are general purpose CSls ln Lhelr naLure, llke Lhe ones purposed by Arlyachandra and lrollck (2008). 8ecause of Lhls, Lhey could be used ln oLher flelds of research, llke ln Lhe case of lnformaLlon sysLems adopLlon research (ln facL Lhey are somewhaL used ln Lhe chapLer abouL lnformaLlon SysLems adopLlon). 8uL whlle belng general purposes ln Lhelr naLure, Lhey should neverLheless be used, as Lhey can form Lhe bulldlng blocks for Lhe 8M lnlLlaLlve. As such, Lhe mosL relevanL Lo Lhls work are presenLed ln Lhe followlng Lable.
/FJIJNPQ "BNNEHH +PNIKF 0EPLJLW Champlon CenLral flgure Lo supporL and promoLe Lhe 8M lnlLlaLlve lnslde Lhe company, glvlng lnformaLlon hlghllghLlng Lhe galns of 8M.
ManagemenL of 8eslsLance Applylng new work Lechnlques always leads Lo reslsLance lnslde Lhe company, ouL of fear or lgnorance. 8y havlng Lhe knowledge of who can have Lhls behavlor, addlLlonal supporL should be
13 /FJIJNPQ "BNNEHH +PNIKF 0EPLJLW glven Lo prevenL evenLual problems.
ManagemenL SupporL 1he lmplemenLaLlon of changes on how Lhe company works musL be supporLed by Lhe managemenL of Lhe company.
SufflclenL 8esources lor changes Lo occur, lL ls necessary Lhe lnlLlaLlve has enough moneLary resources, people, and Llme.
rocess Skllls lor Lhe success of Lhe 8M lnlLlaLlve, lL ls necessary LhaL people wlLh exLenslve knowledge abouL Lhe processes LhaL wlll be worked on are lnvolved ln Lhe process.
1echnlcal Skllls 8esldes Lhe lmporLance of Lhe knowledge abouL Lhe process, lL ls also lmporLanL Lo have people wlLh good knowledge of Lechnlcal skllls Lo help LranslaLe Lhe processes from a LacLlcal level Lo a more operaLlonal level.
user SupporL Slnce all or much of Lhe company wlll be affecLed by Lhe changes LhaL an lnlLlaLlve of Lhls naLure brlngs, Lhe supporL of Lhe users ls very lmporLanL for lLs success.
LffecLlve CommunlcaLlon Cnly wlLh an effecLlve communlcaLlon beLween all Lhe parLlclpaLlng agenLs of Lhe lnlLlaLlve ls lL posslble Lo LransmlL Lhe correcL lnformaLlon Lo achleve a successful lmplemenLaLlon of Lhe lnlLlaLlve.
1able 2 - CrlLlcal success facLors for buslness process managemenL. neverLheless, for Lhls work, whaL was deslred was Lo have a more concreLe and ln-depLh CSls LhaL could be used Lo help lmplemenL Lhe new deslgn of Lhe pro[ecL managemenL processes. Cne ldenLlfled CSl ls Lhe necesslLy Lo have Lhe performance of Lhe work belng done consLanLly measured (Skrln[ar & 1rkman, 2013). 1he purpose of Lhls CSl ls Lo help deLermlne Lhe allgnmenL level of Lhe process wlLh lLs goal, seL by Lhe managemenL.
14 Llkewlse, 8avesLeyn and 8aLenburg (2010) deLermlned ln Lhelr survey LhaL Laklng measuremenLs and conLrolllng Lhe work belng done was also an lmporLanL facLor, and Lhus lLs monlLorlng should be a conLlnuous efforL. Cne example of such conLrol happens aL Lhe end of each developmenL cycle ln a pro[ecL (a sprlnL). 1hls evenL ls called a SprlnL 8evlew, where Lhe daLa from Lhe work belng done ls revlewed and concluslons can be drawn from lL. Cne of Lhe ob[ecLlves of 8M ls Lo sLreamllne Lhe processes. 8elaLed Lo LhaL, one of Lhe CSl ldenLlfled by 1rkman (2010) ls Lhe need for organlzaLlonal and sLrucLural changes, buL also Lhe company's capablllLy Lo respond Lo LhaL change. 1hls crlLlcal facLor helps clean up" redundancles LhaL exlsL ln Lhe company's sLrucLure by glvlng a proper deflnlLlon of Lasks, Lhe compeLences expecLed from Lhe workers and where responslblllLy lles ln Lhe processes. 1hls ls Lrue when lmplemenLlng changes where 8M ls lnvolved, llke ln Lhe example of !arrar, Al-Mudlmlgh, and Zalrl (2000) where Lhe change of managemenL ls also ldenLlfled as a success vecLor, as well as Lhe re- englneerlng of Lhe buslness processes. Lven furLher, one of Lhe references of 8osemann and de 8ruln (2003) also sLaLes LhaL Lhls process of change musL rely noL only on sysLems and sLrucLural change, buL also on culLural change as well (Spanyl, 2003). 8eferenced ln many sources ls Lhe CSl of Lhe need of l1/lS supporL of Lhe 8M lnlLlaLlve. 1he usage of l1/lS should be an enabler and should be used Lo supporL Lhe lnlLlaLlve, and noL as Lhe maln drlver for change (1erzlovskl, llLzpaLrlck, & C'nelll, 2003). l1/lS can add lLs supporL Lo reduce Lhe problems ldenLlfled ln Lhe lnLervlews LhaL were relaLed wlLh Lhe lack of documenLaLlon and communlcaLlon. 8uL l1/lS supporL musL go beyond Lhe supporL for documenLaLlon, or else Lhe sysLem may become an lnformaLlon sllo. Lven more, Skrln[ar and 1rkman (2013) found LhaL Lhe use of a documenLaLlon sysLem has llLLle lmpacL on Lhe processes lnvolved. An analysls of Lhe exlsLlng sysLem (LM) wlll be done ln a laLer chapLer of Lhls work Lo lndlcaLe Lasks LhaL could be done Lo supporL Lhe 8M lnlLlaLlve. As reference Lo Lhese Lasks, some crlLlcal pracLlces LhaL were ldenLlfled by Skrln[ar and 1rkman (2013), llke Lhe auLomaLlon of several acLlvlLles wlLhln Lhe processes, Lo allow Lhe workers Lo have more Llme Lo work on whaL really maLLers and noL ln Lasks LhaL whlle necessary (Lhelr necesslLy musL also be ob[ecL of aLLenLlon ln Lhe re-deslgn), are of a repeLlLlve naLure and can be achleved by auLomaLlon. llnally, a lasL CSl can be referenced from Lhe work of 8avesLeyn and 8aLenburg (2010). 1he lnvolvemenL of Lhe employees ln Lhe re-deslgn of Lhe process resulLs ln
13 Lhelr empowermenL wlLhln Lhe process and Lhe company, whlch may conLrlbuLe Lo Lhe success of sald process. 1hls ls because by parLlclpaLlng ln Lhese acLlons, workers can see Lhe changes and Lhelr beneflLs and acL as champlons, one of Lhe mosL common CSl ln relevanL llLeraLure. 1hls also goes along wlLh whaL Scrum and Lhe Aglle ManlfesLo (lowler & PlghsmlLh, 2001) uphold, by havlng self-organlzlng Leams and allowlng buslness people and developers Lo work LogeLher dally LhroughouL Lhe pro[ecL. =<><>< 0EIFJNH RKF CEPHBFJLW SFKNEHHEH 1he lmplemenLaLlon of a 8M pro[ecL ls noL a guaranLee LhaL Lhe company wlll achleve beLLer resulLs. ln facL, several papers show LhaL 8M lnlLlaLlves usually fall, llke ln Arlyachandra and lrollck (2008). 1he work of Al-Masharl and Zalrl (1999) even sLaLes LhaL as many as 70 percenL do noL achleve Lhe dramaLlc resulLs Lhey seek". Cne reason glven for Lhls occurrence was Lhe lack of measuremenL Lo deLermlne lf Lhe processes are allgned wlLh Lhe company's sLraLegy. ln response Lo Lhls, measuremenLs of Lhe work belng done can be achleved Lhrough Lhe use of meLrlcs. 1he meLhod of meLrlcs for buslness processes ls slmllar Lo Lhe one deplcLed ln Lhe lmage 2 - 1he 8M framework. ln Lhls case, as suggesLed by kennerley and neely (2003), Lhere are Lhree phases Lo Lhe effecLlve lmplemenLaLlon of meLrlcs LhaL are also presenLed ln a cycle. 1hese are: 'ERQENIJKL - ln Lhls phase, an analysls ls done on Lhe exlsLlng meLrlcs (lf any) Lo deLermlne whlch are effecLlve or noL (and Lhus should be pursued or abandoned) and whaL changes should be lmplemenLed ln Lhe meLrlcs ldenLlfled wlLh such necesslLy, 0KYJRJNPIJKL - AL Lhls polnL, Lhe changes LhaL are requlred Lo be done ln Lhe meLrlcs should be lmplemenLed on Lhe lS, &ESQKZCELI - upon lmplemenLaLlon, Lhe lS should sLarL collecLlng daLa Lo feed Lhe meLrlcs requlred Lo monlLor Lhe processes. 1he work of kennerley and neely (2003) also presenLs a serles of facLors LhaL can help wlLh Lhe consLanL evoluLlon of meLrlcs. 1hese are Lhe ones, based on observaLlon, whlch would brlng addlLlonal value Lo Lhe meLrlcs deplcLed on Lhls work:
+PNIKF /PIEWKFZ 8egular process Lo revlew measures wlLh predeLermlned revlew daLes and allocaLed resources. rocess AvallablllLy of mechanlsms Lo Lransfer besL pracLlce. rocess
16 +PNIKF /PIEWKFZ lnLegraLlon of l1 and operaLlonal ob[ecLlves and resources. SysLems llexlble l1 sysLems. Lnabllng modlflcaLlon of daLa collecLlon, analysls and reporLlng Lools. SysLems AvallablllLy of dedlcaLed resources Lo faclllLaLe revlew and modlflcaLlon of measures. eople uevelopmenL of a communlLy of users of measures Lo Lransfer besL pracLlce (e-mall, user groups, benchmarklng). eople CulLure conduclve Lo measuremenL. Senlor managemenL drlvlng measuremenL. undersLandlng of Lhe beneflL of measuremenL. CulLure AccepLance of need for evoluLlon. CulLure LffecLlve communlcaLlon of measures and measuremenL lssues uslng accepLed medla. CulLure use of measures Lo prompL acLlons, reflecL on sLraLegy and processes, eLc. CulLure Cpen and honesL use of measures CulLure 1able 3 - lacLors LhaL help measurlng processes 1he caLegorles of Lhese facLors represenL Lhe followlng: ,EKSQE - 1hese facLors provlde Lhe necessary human resources Lo lmplemenL Lhe meLrlcs, ,FKNEHH - lacLors LhaL guaranLee LhaL a process exlsLs Lo manage Lhe meLrlcs, "ZHIECH - 1hey are Lhe facLors LhaL provlde a sLrucLure LhaL enables Lhe collecLlon of Lhe measuremenLs and ouLpuLs lLs resulLs, /BQIBFE - 1he exlsLence of a company culLure LhaL ls focused on Lhe gaLherlng of lnformaLlon Lo feed Lhe meLrlcs. Cf all Lhe facLors presenLed by kennerley and neely (2003), only Lhe culLural facLors were all adopLed for Lhls work. Whlle deLermlnlng Lhe necessary meLrlcs for Lhe processes can be a dlfflculL Lask, deallng wlLh culLural lssues ls also very challenglng. neverLheless, ln Lhe case of u8l, lLs culLure ls also whaL led noL Lo lnclude some facLors presenLed on Lhe referenced work, because u8l (Lhrough lLs workers) presenLs a sLrong and collaboraLlve efforL Lo consLanLly lmprove lLs meLhods of work.
17 When lmplemenLlng Lhls framework, a correcL proflllng of Lhe LargeL company should be obLalned ln order Lo conslder Lhe approprlaLe facLors from Lhe ones presenLed above: people, process, sysLem and culLure. AnoLher valuable polnL LhaL should be addressed when creaLlng Lhe meLrlcs ls deLermlnlng whlch purpose ls behlnd each Lype of meLrlcs. 8ehn (2003) presenLs a serles of posslble purposes ln hls paper, of Lhose, some can be used on Lhe processes LhaL wlll be ouLllned for Lhls framework. .OPQBPIE PLY NKLIFKQ - 8y evaluaLlng a process, a manager can have a beLLer grasp on how well (or noL) Lhe process ls belng used and lf lL meeLs Lhe goals LhaL should be seL when Lhe process ls lmplemenLed. lor Lhls purpose, 8ehn (2003) clLes a reporL done ln 1994, by Lhe naLlonal Academy of ubllc AdmlnlsLraLlon (nAA, 1994) LhaL sLaLes erformance measuremenL of program ouLpuLs and ouLcomes provldes lmporLanL, lf noL vlLal, lnformaLlon on currenL program sLaLus and how much progress ls belng made Loward lmporLanL program goals". Whlle Lhe prevlous reason Lo measure had lLs focus on Lhe process lLself, conLrol ls almed aL Lhe people LhaL are lnvolved ln Lhe process. AlLhough u8l, by adherlng Lo an aglle orlenLed framework (Scrum), ls looklng Lo empower Lhe Leam (lowler & PlghsmlLh, 2001), Lhere ls sLlll Lhe necesslLy Lo exerL conLrol Lo deLermlne lf Lhe resulLs are ln llne wlLh Lhe goals and sLraLegles seL by Lhe company, 0KIJOPIE PLY NEQEDFPIE - Measurlng a process can also be used Lo moLlvaLe a Leam. Long (buL noL LhaL long) goes Lhe day managers followed Lhe lrederlck 1aylor's process conLrol (8ehn, 2003), now shlfLlng ln favor of Lhe aglle empowermenL. 8y measurlng Lhe poslLlve resulLs of a process, Lhey can be used Lo properly moLlvaLe a Leam. lf Lhe resulLs are noL poslLlve, Lhey can be used Lo sLeer Lhe Leam Lowards Lhe goals deflned ln Lhe sprlnLs. 8y glvlng Lhls opporLunlLy Lo Lhe Leam, lL may lnduce Lhe [oy of reachlng Lhe goals on Lhelr own Lerms, lncreaslng Lhe moLlvaLlon Lo achleve more, whlch ls whaL 8lslng (2013) defended on her presenLaLlon. CelebraLlon of Lhe resulLs of measurlng a process ls, conLrary Lo Lhe prevlous polnL, LargeL Lo Lhe whole company. 8y celebraLlng Lhe reachlng of goals of a process LhaL were seL ln accordance wlLh Lhe sLraLeglc sLeer of a company ls someLhlng LhaL wlll lnduce a greaLer sense of Leam / company coheslon and may lead Lhe lndlvldual Lo lmprove Lhelr self-worLh by parLlclpaLlon ln such successful endeavor, whlch ln Lurn may lmprove Lhelr performance ouLpuL, 4EPFL PLY JCSFKOE - Alongslde wlLh Lhe evaluaLe purpose, Lhe need Lo learn whlle measurlng processes should be an lmperaLlve, slnce by dolng so, Lhe manager
18 can grasp Lhe causes Lo Lhose resulLs (8ehn, 2003). Speclflc meLrlcs should be lmplemenLed Lo obLaln such measures Lo perslsL wlLh Lhe conLlnuous lmprovemenL. 1hls leads Lo Lhe flnal purpose: lmprovemenL. 1hls purpose follows, Lo some exLenL, Lhe polnL presenLed on Lhe Aglle ManlfesL (lowler & PlghsmlLh, 2001) whlch sLaLes AL regular lnLervals, Lhe Leam reflecLs on how Lo become more effecLlve, Lhen Lunes and ad[usLs lLs behavlor accordlngly". 8y knowlng whaL happens on Lhe process, acLlons can be Laken Lo lmprove Lhe process lf necessary or Lo conLlnue ln Lhe same fashlon, lf Lhe lndlcaLors polnL Lo a saLlsfacLory measuremenL. =<@< (%+*'0$#(*% "6"#.0" $&*,#(*% 1he flnal LheoreLlcal focus of Lhls work ls on lnformaLlon sysLem (lS) adopLlon and lmplemenLaLlon pracLlces. 1hls lnformaLlon ls gaLhered from Lhe fleld of ManagemenL of lnformaLlon SysLems and lLs purpose ls Lo acL as Lhe basls of a serles of guldellnes Lo allow a beLLer adopLlon on Lhe lmplemenLaLlon of Lhe redeslgn processes by Lhe users of Lhe LM sysLem. 1he work of Sllver, Markus, and 8eaLh (1993) provldes a common academlc deflnlLlon of lnformaLlon SysLems (lS), ln whlch lS encompasses hardware, sofLware, daLa, people, and procedures. WlLh Lhe excepLlon of hardware, whlch does noL have much relevance ln Lhls work, Lhese componenLs of lS are of hlgh lmporLance for a company and for Lhe success of Lhe lmplemenLaLlon of Lhe processes dlscussed ln Lhls work. 1hese componenLs can be descrlbed ln Lhe followlng: ,FKNEYBFEH - 1he maln componenL, whlch ls Lhe basls for Lhe new lmplemenLaLlon of Lhls work - Lhe processes. 1hese musL be coded ln Lhe form of procedures ln Lhe LM sysLem, &PIP - uaLa ls whaL resulLs from Lhe acLlvlLles of Lhe procedures, produced by Lhe people and wlll be sLored ln Lhe LM sysLem, ,EKSQE - 1he workers of Lhe company (and more lf necessary) LhaL wlll lnLeracL wlLh Lhe processes and wlll produce Lhe daLa. Whlle Lhe procedures are of pro[ecL managemenL naLure, noL only Lhe C wlll lnLeracL wlLh Lhe sofLware, "KRIUPFE - 1he LM sysLem ln whlch Lhe procedures wlll be coded, Lhe daLa wlll be sLored and accesslble Lo Lhe people LhaL wlll lnLeracL ln Lhe fuLure processes. Whlle lL ls noL an exLenslve analysls for crlLlcal facLors for Lhe lmplemenLaLlon of an lS, Lhe work of Subramanlan, !lang, and kleln (2007), clLlng AlLer (1979) purposes a
19 serles of guldellnes for lmplemenLaLlon of an lS wlLh a naLure slmllar Lo whaL ls ln sLudy ln Lhls work, and as such, lL ls lmporLanL Lo Lake Lhem lnLo conslderaLlon. 1hese guldellnes are: 2EES JI HJCSQE. lmplemenL / ueploy only whlch ls really necessary and needed by Lhe users of Lhe lS, .[ENBIJOE KF IKS CPLPWECELI SPFIJNJSPIJKL NKCCJICELI. lL ls necessary for Lhe Lop managemenL Lo be lnvolved wlLh Lhe developmenL and supporL lLs lmplemenLaLlon and lLs use LhroughouL Lhe company. As can be seen before, managemenL commlLmenL ls a consLanL ln all of Lhe areas of sLudy ln Lhls work, #FPJLJLW. rovldlng Lralnlng Lo Lhe users can even permlL a furLher reflnemenL of Lhe lS and aL Lhe same Llme lnvolve Lhe users ln lLs deploymenL Lo achleve a beLLer accepLance wlLhln Lhe company, ,FKIKIZSJLW \ EOKQBIJKLPFZ YESQKZCELI. 1he lmplemenLaLlon / deploymenL of Lhe lS can and should sLarL slmple. 8uL, wlLh each lLeraLlon, new feaLures can be added or more users should be broughL ln. =<@<;< $YKSIJKL KR NTPLWEH Changes are many Llmes dlfflculL Lo lmplemenL. Some even say LhaL changes hurL on a psychologlcal level llke koch (2006), clLlng medlcal sources. 8eslsLlng change ls parL of who we are, movlng away from our comforL zone ls someLhlng LhaL we Lry Lo avold on a consclous and sub-consclous level. So Lhere ls a greaL necesslLy Lo pay aLLenLlon on how change ls managed wlLhln Lhe company Lo ensure Lhe success of Lhe lmplemenLaLlon of Lhe changes suggesLed on Lhls work. llrsL of all, and from an lS percepLlve, Lhe necesslLy of Lhese changes ls a by- producL of Lhe growLh LhaL Lhe sub[ecL company has been susLalnlng for some Llme and also ouL of Lhe necesslLy of lnnovaLlon (Swanson & Wang, 2003) on Lhelr processes, added by Lhe pressure of Lhe markeLs ln whlch Lhe company ls presenL Lo have servlces LhaL dellver hlgh value Lo Lhelr cllenLs. 1hus, Lhere was Lhe necesslLy Lo lmplemenL an LnLerprlse ro[ecL ManagemenL sysLem. Many of Lhe relevanL llLeraLure sLaLes LhaL ln Lhe process of Lhe adopLlon of such changes on Lhe lS level of Lhe company, Lhe changes should be supporLed by one or more champlons, who are ''Lechnologlcal leaders'' responslble for convlnclng Lhe resL of Lhe company of Lhe need for change. 1he champlons' role ls lmporLanL, when Lhey
20 supporL Lhe lnLroducLlon of new lnformaLlon Lechnology unequlvocally, flrmly and declslvely, so LhaL Lhe resL of Lhe company ls convlnced of Lhe need Lo adopL lL. As such, lL ls necessary Lo ldenLlfy Lhem wlLhln Lhe sLrucLure of Lhe company. 1hls ls sLaLed ln Lhe work of 8ruque and Moyano (2007), whlch sLudles Lhe deLermlnanLs for adopLlon of lnformaLlon Lechnologles by SMLs. Whlle Lhls work ls focused on famlly and cooperaLlve companles, a correspondence can be made Lo u8l, Lhe sub[ecL company of Lhls work, slnce as ln a famlly company, everyone knows one anoLher and many even have deallngs ouLslde Lhe company. neverLheless, llke ln Lhe work of 8ruque and Moyano (2007), lL can be observed on u8l LhaL Lhls envlronmenL ls sLarLlng Lo change and Lhe company ls becomlng more sLrucLured and (even) more professlonallzed. Llke ln Lhe example above, Lhese changes lead Lo a LransformaLlon on Lhe culLure LhaL should be Laken lnLo accounL when managlng change aL a company. CorporaLe culLure ls an ever changlng (and some aspecLs cannoL even be conLrolled) properLy of a company, ln whlch knowledge of more senlor workers ls passed Lo more [unlor and recenL workers ln Lhe company (!ackson, 2011). Pynes (2009) clLlng ueshpande and WebsLer !r (1989) descrlbes corporaLe culLure as Lhe paLLern of shared values and bellefs LhaL help lndlvlduals undersLand organlzaLlonal funcLlonlng and Lhus provldes Lhose norms for behavlor ln Lhe organlzaLlon". Whlle lmplemenLlng changes, Lhe aspecL of corporaLe culLure ls many Llmes overlooked and ls ofLen clLed as Lhe prlmary reason for Lhe fallure Lo lmplemenL corporaLe changes (Llnnenluecke & CrlfflLhs, 2010). lL has been observed wlLhln Lhe model company LhaL lLs culLure can be ldenLlfled followlng an lnLegraLlon perspecLlve, as a whole, ln whlch lL dlsplays an lnformaLlonal culLure, meanlng LhaL Lhe l1 and buslness goals are allgned (!ackson, 2011). lL ls also posslble, however, Lo ldenLlfy from Lhe perspecLlve of dlfferenLlaLlon, a culLure LhaL each of Lhe deparLmenLs (sysLems, web developmenL, appllcaLlons developmenL, eLc.) has, comprlslng dlfferenL perspecLlves on lnformaLlon sysLems and how Lo adopL Lhem. WlLh an unaLLended LransformaLlon caused by Lhe lmplemenLaLlon of changes, Lhls behavlor could even deepen, leadlng Lo a non-adopLlon of sald changes. 1hls concludes Lhe round-up of Lhe research done for Lhls work, whlch sLarLed wlLh Lhe analysls of Lhe pro[ecL managemenL processes of Lhe company, speclflcally Lhe problems LhaL plague Lhem and a comparlson wlLh proven crlLlcal success facLors ln relevanL llLeraLure. Slnce Lhe analysls ls on processes, lL makes sense LhaL as Lhe necesslLy of changes was ldenLlfled, a redeslgn should be made on sald processes,
21 uslng 8M for LhaL goal. 1o LhaL end, an analysls was also done for pracLlces on how Lo lmplemenL changes uslng 8M and employlng ldenLlfled CSls. 8ecause Lhese changes need Lo be supporLed by an lnformaLlon SysLem (lS), Lhe besL pracLlces on adopLlon changes on lS were also ob[ecL of sLudy for Lhls work.
22 >< +(%&(%1" WlLh Lhe LheoreLlcal background compleLed, Lhe flndlngs are now presenLed. llrsLly, Lhe sub[ecL company ls conLexLuallzed ln relevance wlLh Lhls work, regardlng Lhelr sLraLeglc goals, buL also regardlng Lhe chosen pro[ecL managemenL meLhodology. ><;< $7*5# #?. /*0,$%6 1he company focused on Lhls work ls a global consulLancy SML named u8l (hLLp://www.drl-global.com), headquarLered ln Llsbon, orLugal. u8l was esLabllshed ln 1999 as a resulL of an academlc spln off. As of !anuary 2013, lL counLs wlLh 40+ employees, placed ln 7 offlces ln Lurope and norLh Amerlca. lLs maln servlce vecLors are: 8uslness & Soclal lnLelllgence, C8M & Soclal C8M, Moblle & LmergenL Medla, Web & laLforms. 1hese servlces are malnly relaLed wlLh web developmenL, done prlmarlly wlLh open source Lechnology or servlces relaLed wlLh hosLlng and helpdesk supporL. 1he developmenL ls malnly done ln Lhe Llsbon offlces, alLhough Lhere are pro[ecLs developed by mulLl-counLry Leams, where communlcaLlon ls an lmporLanL Loplc. lLs currenL sLraLeglc plan alms (among oLhers LhaL fall ouLslde of Lhe scope of Lhls work), on a cusLomer level, Lo reach blgger cllenLs LhaL lead Lo more complex (and Lhus lengLhler) pro[ecLs. 1o supporL LhaL, on an lnLernal processes perspecLlve, Lhe company alms Lo lmprove Lhe operaLlng processes and also Lhe producL quallLy. Cn an organlzaLlon level, Lhe ma[orlLy of Leams ls focused ln web-relaLed developmenL. 8uL Lhere ls also a Leam of sysLem admlnlsLraLlon, a deslgn Leam, a flnanclal Leam and also a commerclal Leam. ln addlLlon Lo Lhe developmenL Leams and besldes Lhe pro[ecL managers LhaL acL as a C for one or more pro[ecLs, Lhere are a number of Leam members LhaL supporL Lhe developmenL of a pro[ecL, as speclallsLs. 1eam members llke Lhe C1C, Lhe ux Lead or a SofLware ArchlLecL are commonly lnvolved wlLh Lhe Leams Lo dellver a producL speclally crafLed for Lhe cllenL. 1here ls a Leam member wlLh Lhe poslLlon of ConLroller LhaL acLs as a hub of lnformaLlon beLween Lhe Leams of dlfferenL areas.
23 ><;<;< ,FKMENI CPLPWECELI CEITKYKQKWZ PI &'( lor much of lLs sLlll shorL llfe, u8l dld noL have an offlclal pro[ecL managemenL meLhodology. neverLheless, such lmporLance was polnLed ouL dally, as Lhe company grew and maLured. 1o achleve LhaL goal, for some Llme now, several Lop managemenL elemenLs aL u8l have been pushlng for a compleLe adopLlon of Scrum, an aglle sofLware developmenL framework. Scrum ls an enhancemenL of Lhe lLeraLlve and lncremenLal approach Lo dellverlng ob[ecL-orlenLed sofLware" (Schwaber, 1997). Some of lLs ldeals and meLhods were used ln a meeLlng of represenLaLlves of several meLhodologles, llke LxLreme rogrammlng, leaLure-urlven uevelopmenL, Scrum, eLc., produclng Lhe Aglle ManlfesLo, slgned by all parLlclpanLs ln sald meeLlng (lncludlng Schwaber), ln 2001. 1he manlfesLo defends Lhe followlng: lndlvlduals and lnLeracLlons over processes and Lools, Worklng sofLware over comprehenslve documenLaLlon, CusLomer collaboraLlon over conLracL negoLlaLlon, 8espondlng Lo change over followlng a plan. Scrum (whlch Lakes lLs name from a formaLlon ln rugby) posLulaLes LhaL Lhe sofLware developmenL should be done ln small Llmed lLeraLlve sLeps, wlLh Lhe lnvolvemenL of Lhe sLakeholder Lo lmprove Lhe quallLy of Lhe producL. SLakeholders can be Lhe cusLomer, end-users, eLc. 1hey are Lhe ones LhaL recelve added value wlLh Lhe developmenL of Lhe pro[ecL. Lach of Lhe developmenL lLeraLlons ls called a SprlnL and can Lake from one Lo four weeks, dependlng on Lhe sLlpulaLlon glven by Lhe ro[ecL Cwner (C). ln Scrum, Lhe ro[ecL Manager flgure does noL exlsL, and lL can be somewhaL replaced by Lhe C flgure. Slnce Lhe sprlnLs are falrly shorL, afLer evaluaLlng whaL wenL rlghL and wrong, changes can be lmplemenLed rapldly. Whlle analyzlng a pro[ecL, lLs requlremenLs are broken down lnLo sLorles by Lhe C. Lach sLory represenLs a funcLlonallLy of Lhe pro[ecL and Lhey are furLher decomposed lnLo lssues by Lhe developmenL Leam. A componenL of Lhe producL can be composed by several sLorles. WlLh a sorLlng order of Lhe lmporLance of sald lssues, Lhese seL of lssues ls whaL composes Lhe producL backlog. 1hls backlog ls where Lhe developmenL Leam, Lhe Scrum MasLer and Lhe C of Lhe pro[ecL come before Lhe sLarL of every SprlnL Lo bulld Lhe SprlnL backlog by chooslng Lhe flrsL sLorles from Lhe producL backlog LhaL flL Lhe lengLh of Lhe SprlnL chosen aL Lhe sLarL of Lhe pro[ecL by Lhe C, Lo answer Lhe goals for Lhe SprlnL, also seL by Lhe C.
24 1he lmplemenLaLlon of Lhls meLhodology ls Laken lnLo accounL and ls lncorporaLed ln Lhe modellng of Lhe processes descrlbed ln Lhe chapLer 3.3. 1hese processes (,FKNEYBFEH) are an lmporLanL componenL of Lhe lnformaLlon SysLem of Lhe company, referred ln Lhe LheoreLlcal background and Lhelr modellng wlll address Lhe problems found aL Lhe company. lrom Lhese processes, &PIP wlll resulL and ls Lo be sLored ln Lhe "KRIUPFE made especlally Lo manage pro[ecLs, Lhe LnLerprlse ro[ecL ManagemenL (LM) sysLem (boLh of Lhese componenLs are, as polnLed ouL ln Lhe LheoreLlcal background, parL of Lhe lS of Lhe company). 1he flnal componenL of an lnformaLlon SysLem ls Lhe ,EKSQE. AL Lhls company, and for Lhe case presenLed aL Lhls work, Lhe bulk ls Lhe developmenL Leams. Lach Leam can be comprlsed by Lwo or more people, dependlng on Lhe pro[ecL aL hand. 8esldes Lhe already dlscussed C flgure, Lhere ls anoLher elemenL presenL on Scrum Leams: Lhe Scrum MasLer. 1hls elemenL ls a person whose maln focus ls Lo overcome problemaLlc slLuaLlons, llke mlsslng lnformaLlon, mlsslng resources, eLc. 1he Scrum MasLer ls presenL aL all Lhe ually MeeLlngs, a meeLlng LhaL, llke Lhe name lmplles, ls done on a dally basls and where one or more Leams of a speclflc deparLmenL gaLher Lo dlscuss Lhe work done ln Lhe prevlous day, whlch problems occurred and whaL wlll be done ln Lhe presenL day. AL Lhese meeLlngs everyone ls encouraged Lo sLep forward wlLh soluLlons Lo problems LhaL Lhelr colleagues presenL. lf no soluLlon Lo Lhe problem ls found, lL ls up Lo Lhe Scrum MasLer Lo flnd resources Lo help solve Lhe problem. ><;<=< .LIEFSFJHE ,FKMENI 0PLPWECELI 1he LM sysLem should be a fundamenLal parL of Lhe lnformaLlon SysLem of any company and as such, ls one of Lhe focus polnLs of Lhls work. ln Lhls sysLem, all lnformaLlon of a pro[ecL can be sLored and consulLed, by uslng Lhe dlfferenL componenLs. 1he sulLe used aL u8l ls from ALlasslan and Lhe maln componenLs Lhls work wlll concenLraLe lLs flndlngs and dlscusslons are: ! -('$ - A pro[ecL managemenL appllcaLlon, LhaL helps Lo Lrack, plan and analyze Lhe daLa from a pro[ecL. 1hls appllcaLlon has several plug-lns lnsLalled, one of whlch ls Lhe #ECSK plug-ln LhaL ls used by Lhe developmenL Leam Lo log Lhe Llme spenL on each Lask of a pro[ecL. AnoLher plug-ln ls Lhe 1FEEL?KSSEF, LhaL ls a vlsuallzaLlon plug-ln Lo dlsplay LhaL Lasks of Lhe pro[ecL ln a Scrum board, wlLh Lhe several sLages of Lhe developmenL, whlch Lyplcally ls 1o-uo, ln rogress and uone.
23 AnoLher plug-ln used ls Lhe )ESTZF, used for sLorlng lnformaLlon abouL Lhe LesLs LhaL should be done for Lhe developmenL belng made, ! /KLRQBELNE - A wlkl llke appllcaLlon LhaL ls focused on Leam collaboraLlon. ln Lhls appllcaLlon, Cs and Lhe developmenL Leam allke can sLore lnformaLlon (LhaL should be ln a common and sLrucLured way) abouL Lhe pro[ecL. 1hls appllcaLlon also allows conLrolllng / vlewlng several calendars, where much lnformaLlon abouL Lhe Leams (among oLher lnformaLlon) can be dlsplayed Lo help every user plan lLs work. Some LM sulLes were analyzed and Lhls was Lhe one chosen Lo supporL Lhe efforL belng made aL u8l Lo lmprove Lhe quallLy of lLs work. lL ls sLlll a work ln progress, wlLh good accepLance from lLs users, buL lL requlres some guldellnes abouL lLs use, whlch are presenLed ln Lhls work. When worklng wlLh Lhe LM sysLem, users should log noL only Lhe Llme spenL on Lhelr work on lssues of Lhe pro[ecL, buL Lhe Llme used on common Lasks LhaL exlsL ln every pro[ecL.
%PCE &EHNFJSIJKL $BYJELNE Daily Meeting
Lach Llme a worker goes Lo a dally meeLlng, he / she should lnpuL Lhe Llme spenL ln LhaL meeLlng. lf a worker ls on Lwo pro[ecLs aL Lhe same Llme, Lhe Llme should be logged ln Lhe pro[ecL where Lhe mosL Llme wlll be spenL ln LhaL day. uevelopmenL 1eam and Scrum MasLer. Cccaslonally for C. SprlnL 8evlew AL Lhe end of a SprlnL Lhere are several meeLlngs LhaL Lake place. 1he Llme spenL on Lhe sprlnL revlew, where Lhe developed work ls demoed Lo Lhe C, should be sLored here. uevelopmenL 1eam, Scrum MasLer and C. SprlnL 8eLrospecLlve AL Lhe end of a SprlnL Lhere are several meeLlngs LhaL Lake place. 1he Llme spenL on Lhe sprlnL reLrospecLlve, where everyone lnvolved ln Lhe pro[ecL glves an lnpuL on whaL wenL rlghL or wrong wlLh Lhe sprlnL, should be sLored here. uevelopmenL 1eam, Scrum MasLer and C. SprlnL lannlng AL Lhe end of a SprlnL Lhere are several meeLlngs LhaL Lake place. 1he Llme spenL on Lhe plannlng for Lhe nexL sprlnL should be sLored here. uevelopmenL 1eam, Scrum MasLer and C. 8acklog Croomlng 1he developmenL Leam should know, aL leasL, whaL Lhey wlll do for Lhe nexL monLh or so. knowlng Lhls allows looklng aL Lhe producL backlog and dolng some groomlng, Lo see lf any lnformaLlon ls mlsslng. 1he Llme spenL dolng Lhls Lask should be lnpuLLed here. uevelopmenL 1eam. ro[ecL lannlng When someone ls lnvolved ln Lhe lnlLlal phase of deflnlng aspecLs of Lhe pro[ecL Lhey should lnserL Lhe Llme spenL ln Lhls lssue. uevelopmenL 1eam, Scrum MasLer and C.
26 %PCE &EHNFJSIJKL $BYJELNE Analysls 1he Llme spenL on analyzlng some new Lechnology or developmenL process should be logged here. uevelopmenL 1eam 1able 4 - Common pro[ecL Lasks ><=< (%#.'9(.3" $%$46"(" As sLaLed above, one of Lhe meLhods Lo gaLher lnformaLlon for Lhls work conslsLed on a seL of lnLervlews. 1hese were made aL Lhe sLarLlng polnL of Lhls work (hase lll) Lo each of Lhe Lhree avallable Cs aL Lhe Llme (Lhe fourLh was unavallable). upon conducLlng an analysls on Lhe lnLervlews, lL became clear LhaL Lhe maln problems ln managlng pro[ecLs aL u8l are: communlcaLlon, documenLaLlon, requlremenLs' assessmenL and Lhe general lack of sLandardlzaLlon of meLrlcs and oLher sLaLus and conLrol sLrucLure Lo evaluaLe pro[ecLs and Lhelr success. As polnLed ouL ln Lhe LheoreLlcal background, problems of communlcaLlon, documenLaLlon (and aL some exLenL requlremenLs' assessmenL) are relaLed mosL of Lhe Llme, slnce much of Lhe communlcaLlon belng done ls noL sLored ln a cenLral reposlLory and documenLaLlon ls noL sLrucLured ln order Lo supporL communlcaLlon. lor compleLe LranscrlpLlons of Lhe lnLervlews, please consulL Lhe chapLer 7.1 of Lhe Appendlx. ><=<;< /KCCBLJNPIJKL] YKNBCELIPIJKL PLY FE^BJFECELIHA PHHEHHCELI SFKDQECH ro[ecL managemenL communlcaLlon ls a problem LhaL has mulLlple vecLors: wlLh Lhe flnanclal deparLmenL, beLween Lhe producL owners and wlLh Lhe cllenL. naLurally Lhe Cs also communlcaLe wlLh Lhe developmenL Leam, buL ln Lhe lnLervlews Lhls vecLor was noL referenced as problemaLlc. 1he communlcaLlon beLween Lhe C and Lhe flnanclal deparLmenL ls malnly done ln person and / or Lhrough emall. 1he flnanclal deparLmenL mosLly sLarLs Lhls communlcaLlon by lnqulrlng abouL Lhe sLaLe of a pro[ecL, because Lhey have Lhe lnformaLlon LhaL ln LhaL perlod Lhey were supposed Lo blll Lhe cllenL for Lhe pro[ecL. Whlle Lhls form of communlcaLlon usually works mosL of Lhe Llme and was only marglnally referenced ln Lhe lnLervlews, lL ls noL as effecLlve as lL can be, slnce Lhls slLuaLlon leaves Lhe flnanclal deparLmenL dependenL on Lhe lnlLlal lnformaLlon referrlng Lo when a cllenL should be bllled. lf for some reason LhaL lnformaLlon geLs ouLdaLed, Lhere ls noL a process Lo dlsLrlbuLe LhaL lnformaLlon. 1he case of communlcaLlon problems among producL owners ls due Lo Lhe lack of a sLrucLured procedure of communlcaLlon beLween Lhem. Cne lssue LhaL was noL polnLed ln Lhe lnLervlews buL was commenLed afLerwards was LhaL Lhe lnformaLlon of
27 Lhe allocaLlon of Lhe Leams ls noL easlly accesslble Lo every C, alLhough Lhere ls some progress ln LhaL fleld, wlLh regular Cs meeLlngs, where Lhe lnformaLlon abouL Lhe sLaLus of pro[ecLs ls shared. 1he lnformaLlon ls noL avallable on Lhe LM sysLem of Lhe company, as lL resldes Loday on a shared onllne documenL. 1hese problems can lead, for example, Lo Lhe loss of some of Lhe members of Lhe developmenL Leam (or Lhe whole Leam) by Lhe roducL Cwner Lo anoLher pro[ecL LhaL has a greaLer percelved lmporLance, alLhough Lhere ls noL a dlsLlncL deflnlLlon of whaL Lhe lmporLance of a pro[ecL ls. llnally, Lhe problems of communlcaLlon wlLh Lhe cllenL are a broader lssue as Lhey also lnvolve documenLaLlon and requlremenLs' assessmenL lssues. Whlle Lhe maln problem ls when lnformaLlon ls needed from Lhe cllenL durlng Lhe developmenL of Lhe pro[ecL and sald lnformaLlon longs Lo arrlve (someLhlng LhaL eludes Lhe purposes of Lhls work), Lhere are also problems wlLh Lhe lnlLlal communlcaLlon wlLh Lhe cllenL, when Lhe pro[ecL ls ln Lhe requlremenLs' assessmenL sLage. Slnce Lhere ls noL a sLrucLured paLh on how Lo proceed aL Lhls sLage, Lhe lnformaLlon ls gaLhered and sLored ln dlfferenL manner by each C, dependlng on Lhelr experlence. 1hls also leads Lo poor esLlmaLlons, as Lhe pro[ecL ob[ecLlves are noL clearly known. Also, ln Lhe lnLervlews, lL was posslble Lo deLermlne LhaL Lhere ls noL a meLhod Lo classlfy a pro[ecL regardlng lLs lmporLance for Lhe company and also Lo caLegorlze each of Lhe requlremenLs ln Lerm of Lhelr lmporLance, Lo deLermlne Lhe crlLlcal success facLors for Lhe pro[ecL. 8esolvlng Lhese lssues can also help reduce Lhe necesslLy of lnformaLlon from Lhe cllenL ln Lhe laLer sLages of Lhe pro[ecL, whlch causes walLlng Llmes and hlnders Lhe normal flow of Lhe pro[ecL. ><=<=< 4PNV KR NKLIFKQ KL SFKMENI HIPIBH PLY RBFITEF PLPQZHJH SFKDQECH AnoLher ma[or lssue ralsed on Lhe lnLervlews was LhaL Lhere are noL esLabllshed sLrucLures Lo conLrol Lhe sLaLus of Lhe pro[ecL yeL. 1hls means LhaL from Lhe sLarL, Lhere ls lnformaLlon LhaL ls mlsslng, namely whlch are Lhe crlLlcal funcLlonallLles LhaL should be lmplemenLed for Lhe pro[ecL Lo be consldered a success. Llkewlse, Lhere ls noL a sLandard way for Lhe C Lo deLermlne lf Lhe pro[ecL ls on Lrack and on budgeL. A slLuaLlon LhaL ls also Lrue regardlng Lhe developmenL Leam, whlch has a large lmporLance ln Lhe organlzaLlon of work ln Lhe aglle meLhodologles. lL ls deslrable LhaL Lhey have Lhls lnformaLlon Lo glve Lhelr lnpuL relaLlng Lhe developmenL LhaL should be done. As for when Lhe pro[ecL ls concluded, Lhere ls noL a
28 predeLermlned seL of rules Lo declare lf Lhe pro[ecL was a success or noL, slnce Lhere are no meLrlcs Lo help deLermlne lL, allke Lhere ls no meLrlc Lo deLermlne lf Lhe esLlmaLlons made for Lhe Lasks were ln facL Lhe Llme spenL on Lhem. ><>< 0*&.4(%1 #?. ,'*/.""." WlLh Lhe lnformaLlon obLalned from Lhe lnLervlews, unscrlpLed Lalks wlLh peers and Lhrough work experlence, lL ls posslble Lo ldenLlfy Lhe flve buslness processes relaLed Lo pro[ecL managemenL LhaL represenL Lhe workflow slnce Lhe beglnnlng of Lhe pro[ecL Lo lLs end. 1he end may represenL Lhe acLual end of Lhe lmplemenLaLlon of Lhe producL or Lhe end of a phase of Lhe pro[ecL lf Lhe pro[ecL has Lhe necesslLy Lo be dlvlded lnLo phases. Lach of Lhese processes wlll lnclude, whenever posslble, meLrlcs Lo measure Lhe quallLy of Lhe work belng done. WlLh Lhls daLa lL wlll be posslble Lo deLermlne lf Lhe meLhodology employed ls worklng as expecLed or lf lL ls necessary Lo lnLroduce changes onLo lL. 1hls analysls should be done yearly. 1hls Llme perlod ls besL sulLed slnce a shorLer Llme frame would noL generaLe enough daLa Lo allow an accuraLe percepLlon of Lhe effecLlveness of Lhe process, and a longer Llme frame, whlle valld, could presenL a rlsk, by perpeLuaLlng lnefflclency. 1o lncrease Lhe changes of a successful lmplemenLaLlon of Lhe 8M framework on Lhe processes of pro[ecL managemenL, Lhe usage of Lhe crlLlcal facLors dlscovered on Lhe LheoreLlcal background ls plvoLal. As such, Lhe champlon of Lhls pro[ecL ls long ldenLlfled (lL was Lhe flrsL sub[ecL belng lnLervlewed), buL from day one, all Lhe managemenL supporLs Lhe lnlLlaLlve. Moreover, wlLh a correcL usage of Lhe lS supporLlng Lhose changes wlll come a good dlsLrlbuLlon of Lhe global knowledge abouL Lhe processes as well as Lhe Lechnlcal skllls requlred for Lhese Lasks. 8egardlng Lhe sLeps polnLed ouL ln Lhe LheoreLlcal background for modellng a process uslng a 8M framework, Lhe sLages of SLraLegy and lan are conducLed ln Lhe scope of Lhls work. llrsL by allgnlng Lhe modellng wlLh Lhe sLraLeglc goals of Lhe company, Lhen by reflnlng Lhe processes and employlng a sLrucLured documenLaLlon Lo supporL Lhe lmprovemenL of quallLy of Lhe work, by provldlng Lhe lnformaLlon when needed wlLh a common sLrucLure across pro[ecLs. llnally, Lhe lan phase wlll focus on Lhe acLual modellng of Lhe processes. Powever, flrsLly lL ls necessary Lo model Lhe processes, as Lhe processes dlscussed on Lhls chapLer had noL been prevlously modeled, slnce Lhere was noL a formal lmplemenLaLlon of Lhem. 8ecause of Lhls, each C would lmplemenL Lhe pro[ecL
29 ln Lhelr own way, maklng Lhe modellng of Lhe processes AS lS lmpracLlcable for Lhls work. 8ecause of Lhls, only Lhe 1C 8L modellng wlll be presenLed. All Lhe processes wlll be presenLed as a graphlcal represenLaLlon uslng Lhe 8uslness rocess Model and noLaLlon (8Mn). 1he usage of Lhe 8Mn ls based on Lhe easlness of lmplemenLaLlon and Lhe ample seL of arLlfacLs LhaL can be used for modellng, demonsLraLed on relaLed sub[ecLs on Lhe masLers' course. Also, Lhe analysls of Wohed, AalsL, uumas, PofsLede, and 8ussell (2006) whlch evaluaLes Cn Lhe SulLablllLy of 8Mn for 8uslness rocess Modellng", susLalned Lhe declslon Lo use Lhls noLaLlon. 1hese nexL chapLers descrlbe each of Lhe processes, by dlsplaylng Lhe 1C 8L model, a compleLe descrlpLlon of lL, Lhe usage of Lhe LM sysLem on how Lo obLaln Lhe lnformaLlon gaLhered ln Lhe processes and lf Lhere are any, a compleLe descrlpLlon of Lhe meLrlcs used on Lhe processes. ln Lhe scope of Lhls work, each pro[ecL should have Lhe followlng processes' workflow:
lmage 3 - CompleLe process workflow for a pro[ecL ><><;< &JHNKOEFZ SFKNEHH 1he flrsL process ls called &JHNKOEFZ, and lL ls sLarLed when Lhe pro[ecL ls awarded Lo Lhe company. Slnce Lhe lnlLlal conLacL ls Lyplcally made wlLh Lhe commerclal deparLmenL (alLhough, dependlng on Lhe Lype of cllenL and pro[ecL, some managemenL sLaff can also be lnvolved early on), many aspecLs of Lhe pro[ecL have Lo be dlscovered. A represenLaLlon ln 8Mn noLaLlon ls presenLed below:
30
lmage 4 - ulscovery process 5"5"#"#" 9/%*200 82,+):0 lor Lhls process, Lhe focus ls Lo deLermlne Lhe pro[ecL needs, whlch range from Lhe buslness' needs, Lhe cllenL's needs and Lhe flnal user's needs. 1he pro[ecL's goals and relaLed consLralnLs LhaL may exlsL are also gaLhered aL Lhls sLage. All of Lhls lnformaLlon ls dlscovered ln collaboraLlon wlLh Lhe pro[ecL's sLakeholders, buL also wlLh Lhe developmenL Leam, as Lhe Leam can add a valuable lnslghL based on Lhelr experlence galned ln prevlous pro[ecLs. So Lhe flrsL Lask (#1) wlll be Lhe ldenLlflcaLlon of sald sLakeholders, slnce Lhey are Lhe ones LhaL should have all Lhe requlred lnformaLlon lndlcaLed above, Lhus Lhls Lask should be done as early as posslble. ln Lhe nexL Lask (#2) Lhe pro[ecL raLlng wlll be deLermlned. 1hls raLlng ls an average of several lndlcaLors Lo deflne Lhe level of lmporLance of Lhe pro[ecL for Lhe company. 1hls can lead Lo declde whlch Leam ls asslgned Lo Lhe pro[ecL, dependlng of lLs complexlLy, lf Lhe pro[ecL ls of hlgh prlorlLy for Lhe company's lnLeresLs or Lhe order on whlch Lhe pro[ecLs wlll be developed. 8elow are Lhe lndlcaLors ln a Lable formaL:
(LYJNPIKF &EHNFJSIJKL value of Lhe pro[ecL
lnformaLlon relaLed wlLh Lhe overall prlce or Llme equlvalenL agreed wlLh Lhe cllenL. lmporLance for Lhe company
ueLermlne lf Lhe compleLlon of Lhls pro[ecL wlll brlng fuLure pro[ecLs wlLh Lhe cllenL or lf Lhe pro[ecL ls of hlgh vlslblllLy.
ComplexlLy of Lhe pro[ecL Plgh-level evaluaLlon of Lhe complexlLy ln order Lo
31 (LYJNPIKF &EHNFJSIJKL
compleLe Lhe pro[ecL. 1able 3 - ro[ecL raLlng lndlcaLors Lach of Lhe lndlcaLors should Lake a numerlc value from 1 Lo 10 and Lhelr average wlll lndlcaLe Lhe pro[ecL's lmporLance for Lhe company. 1hese lndlcaLors should be revlsed ln Lhe yearly evaluaLlon of Lhe processes and more can be added or removed lf Lhelr pracLlcablllLy ls noL proven. WlLh Lhe raLlng of Lhe pro[ecL lL wlll be easler Lo deLermlne (on Lask #3) whlch Leam should Lake on Lhe pro[ecL or even declde Lo use several Leams, based on Lhe knowledge of Lhelr developmenL capablllLy Lo Lackle Lhe complexlLy of Lhe pro[ecL, buL also lf LhaL Leam lf avallable or noL ln Lhe expecLed perlod of developmenL. 1o deLermlne when a Leam ls avallable, Lhe C wlll have a Lool on Lhe LM sysLem. 1he Confluence appllcaLlon ln Lhe LM has a serles of calendars where lL ls posslble Lo obLaln lnformaLlon of varlous Lypes. lor Lhls case, Lhere ls Lhe 1eam AsslgnmenL Calendar deplcLlng LhaL lnformaLlon. 1hls lnformaLlon ls Lyplcally collecLed aL a meeLlng LhaL occurs weekly, where all Lhe Cs convene Lo reporL on Lhelr pro[ecLs. 1he company's ConLroller gaLhers LhaL lnformaLlon and updaLes lL on Confluence. lL ls lmporLanL Lo deLermlne whlch Leam wlll be asslgned Lo Lhe pro[ecL as early as posslble Lo lnLroduce Lhem Lo Lhe pro[ecL, buL also Lo have Lhelr lnpuL based on Lhelr LaclL knowledge (galned ln prevlous pro[ecLs) aL Lhls crlLlcal phase of Lhe pro[ecL, where Lhe requlremenLs are belng ldenLlfled. 1he nexL Lask (#4) ls Lhe plvoLal one and ls where all of Lhe lnlLlal lnformaLlon wlll be gaLhered. uurlng Lhls collecLlon, Lhe lnformaLlon can be sLrucLurally sLored ln Confluence for reference and communlcaLlon wlLh Lhe resL of Lhe Leam (Lhe sLrucLure ls presenLed ln Lhe nexL chapLer). 1hls Lask ls lmporLanL Lo deLermlne Lhe pro[ecL needs, goals (such as Lhe lmplemenLaLlon of Lhe mosL lmporLanL componenLs for Lhe sLakeholders - CSls), and funcLlonallLles (also known as funcLlonal analysls). 1hese Loplcs are dlscussed ln close collaboraLlon wlLh Lhe Leam, as polnLed ouL as a crlLlcal success facLor ln Lhe LheoreLlcal background. 1he lnformaLlon gaLhered ls also a safeguard, because Lhe sLakeholder musL acknowledge and accepL lL before movlng Lo Lhe nexL Lask of Lhe process. AfLerwards, on Lhe lasL Lask of Lhls process (#3), Lhe lnformaLlon archlLecLure wlll be deLermlned. 1he lnformaLlon archlLecLure ls where lL ls deLermlned how Lhe daLa ls
32 organlzed and how lL ls processed. Also, Lhe sLakeholders musL formally accepL Lhls lnformaLlon Lo close Lhls process. AL each of Lhe process sLeps, a serles of lnformaLlon ls gaLhered and should be sLored ln Lhe company's LM. WlLh Lhls documenLaLlon, Lhe problems ldenLlfled ln Lhe lnLervlews should be quenched, such as Lhe problems ln geLLlng Lhe correcL requlremenLs from Lhe cllenL and pro[ecL's documenLaLlon. 1he developmenL Leam wlll have access Lo Lhls documenLaLlon, and a valldaLlon of lnformaLlon musL be obLalned from Lhe sLakeholders by means of elecLronlc communlcaLlon. 5"5"#"!" ;0+62 %. ,72 <9= All of Lhe lnformaLlon gaLhered ln Lhls process should be lnserLed onLo Confluence for fuLure reference. 8egardlng Lhls process, Lhe followlng dlagram deplcLs how Lhe lnformaLlon should be lnserLed. Lach elemenL ln Lhe dlagram represenLs a page ln Confluence:
lmage 3 - lnformaLlon sLrucLure of Lhe ulscovery process 1he flrsL page, ro[ecL lnformaLlon, works [usL as a placeholder for all oLher pages and wlll be used ln oLher processes. ! 1ELEFPQ (LRKFCPIJKL - A general descrlpLlon of Lhe pro[ecL. Also presenL ln Lhls page wlll be Lhe lnformaLlon of Lhe pro[ecL raLlng (from Lask #2), ! "IPVETKQYEFH PLY ,EFHKLLEQ - 1hls page sLores Lhe lndlcaLlon of Lhe sLakeholders and Lhelr conLacLs for reference, as well as Lhe personnel of u8l LhaL wlll be lnvolved ln Lhe pro[ecL, savlng Lhe lnformaLlon gaLhered ln Lasks #1 and #3,
33 ! *DMENIJOEH PLY 'EHIFPJLIH - All of Lhe lnformaLlon gaLhered on Lask #4 should be sLored aL Lhls page, ! (LRKFCPIJKL PFNTJIENIBFE - 1he lnformaLlon gaLhered from Lhls Loplc on Lask #3 ls sLored ln Lhls page. AfLer Lhls lnformaLlon ls dlscovered", Lhe pro[ecL can move on Lo Lhe nexL process. ><><=< &ERJLJIJKL SFKNEHH 1he nexL (second) process ls &ERJLJIJKL. AL Lhls sLage, whaL ls lmporLanL ls Lo analyze Lhe lnformaLlon gaLhered from Lhe prevlous process, break Lhe lnformaLlon down lnLo manageable blLs and caLegorlze lL, maklng lL posslble Lo obLaln more accuraLe esLlmaLlons of Lhe work Lo be done. 1he maln agenLs lnvolved ln Lhls process, besldes Lhe C, are Lhe sLakeholders and Lhe developmenL Leam.
lmage 6 - ueflnlLlon process 5"5"!"#" 9/%*200 82,+):0 ln Lhls process, Lhe maln Lask ls Lhe deflnlLlon of Lhe scope of Lhe pro[ecL, buL also Lhe scope of Lhe producL. 1he dlfference of Lhese Lwo scopes ls descrlbed ln Lhe M8Ck Culde (ro[ecL ManagemenL lnsLlLuLe lnc, 2008): ! 1he SFKMENI HNKSE relaLes wlLh Lhe goals of Lhe pro[ecL, meanlng LhaL Lhe pro[ecL should resulL ln an answer (Lyplcally a producL) Lo Lhe needs of Lhe cllenL, ! 1he SFKYBNI HNKSE ls relaLed wlLh Lhe producL LhaL ls produced ln Lhe pro[ecL and represenLs Lhe feaLures LhaL lL musL dellver.
34 Whlle developlng Lhe pro[ecL, Lhe lack of a clear deflnlLlon of scopes has been ldenLlfled as Lhe cause of delay ln some pro[ecLs. As such, Lhe flrsL Lask of Lhls process (#1) ls where Lhe deflnlLlon of Lhe flrsL scope wlll be addressed. 1he C can consulL wlLh varlous elemenLs of Lhe developmenL Leam Lo asslsL ln Lhls Lask, lf necessary. AfLerwards, Lhe sLakeholders should approve Lhls scope. lf Lhe sLakeholders laLer requlre changes of scope ln Lhe pro[ecL, Lhe lnformaLlon gaLhered here wlll help Lhe C Lo medlaLe any confllcL LhaL may arlse from Lhls requesL. lL ls noL LhaL change cannoL be done. ln facL, by uslng an aglle meLhodology, Lhe company embraces change. 1he polnL here ls LhaL by deflnlng a scope, a commlLmenL ls done by Lhe C ln whlch LhaL pro[ecL wlLh sald scope wlll be done ln x Llme. 8y addlng changes, Lhe sLakeholders musL drop some oLher funcLlonallLles LhaL are noL as necessary as Lhe changes proposed by Lhe sLakeholders. AnoLher opLlon ls LhaL Lhe changes are lncorporaLed ln Lhe scope by renegoLlaLlng furLher paymenL and Llme. nexL ln llne (#2) wlll be Lhe deflnlLlon of Lhe producL scope and Lhe caLegorlzaLlon of lLs componenLs ln accordance wlLh Lhelr lmporLance, ln order Lo obLaln Lhe crlLlcal success funcLlonallLles for Lhe pro[ecL. 1hls may be seen as a breaklng-down of Lhe goals of Lhe pro[ecL LhaL were prevlously deflned ln Lhe &JHNKOEFZ process, wlLh furLher reflnemenL. Agaln a close collaboraLlon wlLh elemenLs of Lhe developmenL Leam and Lhe sLakeholders ls lmporLanL Lo obLaln Lhe maxlmum amounL of lnformaLlon. AfLer all of Lhe lnformaLlon ls collecLed and sLored ln Lhe LM sysLem, Lhe sLakeholders musL formally accepL Lhese requlremenLs Lo advance Lhe process. AcLlvlLy #3 ls where LacLlcal and operaLlonal opLlons are dlscussed and declded. 1ogeLher wlLh Lhe developmenL Leam and oLher ln-house experLs llke Lhe C1C or Lhe Senlor SofLware ArchlLecL, lL ls dlscussed ln deLall whaL ls Lhe besL way Lo lmplemenL Lhe pro[ecL, whlch Lechnologles should be used, among oLher declslons relevanL Lo Lhe pro[ecL, for lnsLance lf Lhe pro[ecL should be done ln whole, or ln phases Lo dellver addlLlonal value fasLer Lo Lhe flnal cllenL. lf Lhls happens, Lhe nexL processes repeaL as many Llmes as Lhere are phases ln Lhe pro[ecL. 8ased on Lhe lnformaLlon gaLhered of Lhe scope of Lhe pro[ecL / producL and also wlLh Lhe pro[ecL's goals gaLhered on Lhe &JHNKOEFZ process, lL ls posslble Lo obLaln a llsL of LesLs LhaL should be done Lo deLermlne lf Lhe producL ls ready for dellvery (accepLance LesLs). 1he flnal acLlvlLy (#4) ls Lhe furLher breakdown of Lhe requlremenLs lnLo sLorles LhaL form Lhe producL backlog, whlch ls an arLlfacL of Scrum. lor each of Lhe sLorles creaLed, Lhe C musL lndlcaLe an esLlmaLe for lLs cosL of developmenL, Lhe Crlglnal
33 1lme LsLlmaLe (C1L). 1o reach Lhls esLlmaLe Lhe C can consulL wlLh Lhe developmenL Leam. AddlLlonally Lhe C musL deflne whaL Lhe accepLance LesLs are for each of Lhe sLorles. AL Lhls polnL, Lhe C should have enough lnformaLlon Lo declde Lhe lengLh of Lhe SprlnLs, and alLhough Scrum does noL posLulaLe Lhe change of sald lengLh, as lL may cause a mlsrepresenLaLlon of Lhe meLrlcs, Lhe C musL have room Lo change lL when Lhe pro[ecL sLaLus requlres lL, slnce Scrum also upholds Lhe embraclng of change. 5"5"!"!" ;0+62 %. ,72 <9= unllke Lhe prevlous process, Lhe lnformaLlon gaLhered ln each of Lhe Lasks of Lhls process wlll noL only be sLored ln Confluence, buL also ln !l8A because Lhe flnal Lask (#4) of Lhe process ouLpuLs sLorles. 1o creaLe sLorles on Lhe LM, Lhe C should creaLe lssues of Lhe Lype SLory" wlLh Lhe followlng lnformaLlon:
%PCE &EHNFJSIJKL uescrlpLlon 1he general descrlpLlon of Lhe sLory. WhaL ls expecLed from Lhe sLory and oLher relevanL lnformaLlon. Where Lo uemo 1he locaLlon on where Lo demonsLraLe Lhe funcLlonallLy of Lhe sLory. Pow Lo uemo Pow Lo demonsLraLe Lhe funcLlonallLy of Lhe sLory Lo Lhe C. AccepLance 1esLs 1he deflnlLlon of Lhe accepLance by Lhe C, whlch generally corresponds Lo Lhe how Lo demo lnformaLlon. AddlLlonally, lndlcaLlon of LesLs necessary Lo achleve Lhe sLaLus of accepLed sLory should be lnserLed Lhrough Lhe use of Lhe Zephyr plugln. ueflnlLlon of uone 1he general deflnlLlon Lo deLermlne lf Lhe sLory ls done. Crlglnal 1lme LsLlmaLe 1he esLlmaLe done by Lhe C. 1hls ls Lhe Llme LhaL, glven all Lhe lnpuLs provlded by Lhe developmenL Leam (lf any), Lhe C expecLs Lo spend on Lhls sLory. Sub-Lasks 1hls daLa ls lnLroduced by Lhe developmenL Leam, wlLh all of Lhe necessary Lasks (acLlons) Lo be performed Lo compleLe Lhe sLory. All exlsLlng sub-Lasks musL have an esLlmaLlon done by Lhe developmenL Leam, and are enLlLled 8emalnlng 1lme LsLlmaLe (81L). CreaLlng sub- Lasks ls noL always necessary as Lhe sLory may be aLomlc ln naLure, meanlng LhaL Lhe sLory descrlbes a unlque acLlon Lo be Laken. ln Lhls case Lhe developmenL Leam lnserLs Lhelr esLlmaLlon (81L) ln Lhe sLory lLself. 1able 6 - lnformaLlon needed for a sLory
36 1he sLorles musL have all Lhe lnformaLlon needed for Lhe developmenL Leam Lo analyze and Lo creaLe sub-Lasks Lo develop Lhe sLory. lL wlll be on Lhese lssues LhaL Lhe Leam wlll work on and log Lhelr work Llme (also ln !l8A). 8uL as Lhls Lask ls Lhe responslblllLy of Lhe developmenL Leam, lL ls noL deplcLed on Lhls process' model. 1hls acLlvlLy ls called backlog groomlng. 1he sLory can only be consldered ready Lo be lncluded ln a sprlnL backlog afLer Lhe Leam lnserLs Lhe lnformaLlon descrlbed above and marks Lhe sLory as 8eady". When analyzlng Lhe pro[ecL aL Lhe end of a sprlnL (done ln Lhe &EOEQKSCELI process) or even aL Lhe end of Lhe pro[ecL (done aL Lhe &EQJOEFZ process), Lhe C can look aL Lhe dlscrepancles obLalned (Lhrough Lhe analysls of meLrlcs) beLween Lhe esLlmaLlon glven Lo Lhe sLory and Lhe sum of Lhe lssues Lhe developmenL Leam gave. 1haL lnformaLlon can help Lhe C make more accuraLe esLlmaLlons on oLher pro[ecLs. 1he followlng dlagram dlsplays how Lhe lnformaLlon LhaL Lhls process ouLpuLs ls sLored on Confluence:
lmage 7 - lnformaLlon sLrucLure of Lhe ueflnlLlon process 1he flrsL page, ro[ecL lnformaLlon, ls Lhe same deplcLed on Lhe prevlous process LhaL also holds Lhe general lnformaLlon relaLed Lo Lhe pro[ecL. ! ,FKMENI HNKSE - Llke polnLed ouL ln Lhe prevlous chapLer, Lhls lnformaLlon ls gaLhered from Lhe sLakeholders ln Lask #1 and should be formally approved by Lhem. Any changes LhaL are done afLerwards, should be lnserLed here along wlLh Lhe raLlonale of LhaL change, whaL has been declded and Lhe formal approval of Lhe sLakeholders, ! ,FKYBNI HNKSE - 1he lnformaLlon of Lhe scope of Lhe pro[ecL wlll be lodged aL Lhls page. Llke Lhe prevlous scope, all changes ln Lhls scope
37 should also be sLored ln Lhls page, wlLh Lhe [usLlflcaLlon of Lhe change and Lhe formal agreemenL of Lhe sLakeholders, ! #ENTLJNPQ (LRKFCPIJKL - 1hls page ls on Lhe same level as Lhe page ro[ecL lnformaLlon and lL gaLhers all Lhe lnformaLlon LhaL resulLs from Lask #3. 1he lnformaLlon ls, as Lhe name lmplles, of a more Lechnlcal naLure conLrary Lo Lhe prevlous pages. 1he lnformaLlon ranges from access daLa Lo appllcaLlons and servlces used on Lhe pro[ecL. Slnce much of Lhls lnformaLlon varles from pro[ecL Lo pro[ecL, lL ls lmposslble Lo easlly caLegorlze Lhe lnformaLlon Lo dlfferenLlaLe ln dlverse pages. Also, Lhe organlzaLlon of Lhe subpages ls done by Lhe developmenL Leam and noL by Lhe C of Lhe pro[ecL, whlch ls Lhe maln reclplenL of Lhe flndlngs of Lhls work. WlLh all Lhe lnformaLlon worked on and accesslble by all members of Lhe Leam, Lhe pro[ecL can evolve Lo Lhe nexL process. ><><>< &EOEQKSCELI SFKNEHH 1he &EOEQKSCELI process, Lhe nexL process ln llne, ls where mosL of Lhe developmenL of Lhe producL ls effecLlvely done. 1hls process ls hlghly lLeraLlve, meanlng LhaL lL wlll be processed as many Llmes as Lhere are sprlnLs for Lhe currenL phase of Lhe pro[ecL or for Lhe enLlre developmenL sLage. 1he maln developmenL sLage ls where Lhe producL ls belng developmenL based on Lhe lnformaLlon gaLhered on Lhe prevlous processes. neverLheless, new developmenL lLeraLlons could happen ln Lhe &EQJOEFZ process, lf small correcLlons are proven Lo be necessary and lf Lhey can be done. lor Lhese correcLlons (and for Lhe dellvery of Lhe producL lLself), anoLher seL of sprlnLs wlll Lake place.
38
lmage 8 - uevelopmenL process ln Lerms of pro[ecL managemenL and followlng Lhe acLlons proposed by Lhe Scrum framework, Lhe pro[ecL owner has, wlLh Lhls process, Lhe maln mlsslon of monlLorlng and conLrolllng Lhe developmenL of Lhe producL. 5"5"5"#" 9/%*200 82,+):0 1he lnlLlal Lask (#1) ln Lhe process ls Lhe deflnlLlon of Lhe sprlnL by Lhe C ln close collaboraLlon wlLh Lhe Leam. Llke prevlously dlscussed, lf necesslLy arlses, Lhe duraLlon of Lhe SprlnL can change. 1he reasons could range from Lhe avallablllLy of Lhe Leam, Lhe complexlLy of Lhe developmenL of Lhe upcomlng sprlnLs or Lhe sLaLus of developmenL requlres more Llme Lo produce a presenLable proLoLype. SLlll ln Lhls Lask, and wlLh Lhe lnformaLlon of Lhe duraLlon of Lhe sprlnL, Lhe Leam proposes a sprlnL backlog, based on Lhe sorLlng order Lhe C gave Lo Lhe producL backlog. As for Lhe C, he should seL ouL Lhe goals for Lhe sprlnL, based on Lhe pro[ecL plannlng. So, Lo compleLe Lhls Lask, an undersLandlng beLween Lhe C and Lhe Leam musL be reached on whaL Lhe C accepLs as Lhe resulL of Lhe sprlnL and how far Lhe Leam commlLs Lo develop ln sald sprlnL. WlLh all Lhls lnformaLlon deflned, Lhe sprlnL ls sLarLed by Lhe C aL Lhe Lask #2, by lndlcaLlng lL on Lhe !l8A appllcaLlon. AfLerwards, Lask #3 occurs durlng Lhe lengLh of Lhe sprlnL, ln whlch Lhe C wlll monlLor Lhe Llme spenL on Lhe developmenL belng made by Lhe Leam (agalnsL whaL was prevlously esLlmaLed). 1hls monlLorlng ls achleved by uslng Lhe LM sysLem. lor
39 each pro[ecL, !l8A provldes a vlsual ald, Lhe burn-down charL. 1hls Lool wlll be furLher deLalled ln chapLer 3.3.3.3. AnoLher Lool Lo conLrol Lhe sLaLe of Lhe pro[ecL can be Lhe aLLendance of Lhe dally meeLlngs. Cver Lhese meeLlngs, every member of a Leam or even of an area of Lhe company (Lhe web developmenL deparLmenL, for lnsLance) parLlclpaLes ln a dally basls, lL has a duraLlon of no more Lhan flfLeen mlnuLes. ln Lhese meeLlngs, ln whlch Lhe C can be presenL buL [usL ln a waLcher capaclLy, Lhe Leam members dlscuss whlch Lasks were done ln Lhe prevlous day, Lhe dlfflculLles LhaL appeared and oLher lssues LhaL are LhoughL relevanL. Lvery oLher co-worker can glve Lhelr lnpuL Lo solve a problem. 1hls way, Lhe C can have a good grasp of Lhe sLaLus and problems of Lhe pro[ecL and can dlscuss Lhem wlLh Lhe Leam afLer Lhe meeLlng, lf necessary. lf a problem (delay) ls deLecLed, Lhe C musL deLermlne lLs cause wlLh Lhe Leam (#4). lf Lhe problems cannoL be quenched wlLhln Lhe sprlnL, some lLems of Lhe sprlnL backlog may have Lo be posLponed for a fuLure sprlnL. lf Lhls happens, lL may even lead Lo some funcLlonallLles wlLh leasL lmporLance Lo be puL aslde ln order Lo conclude Lhe mosL lmporLanL funcLlonallLles of Lhe pro[ecL. 1haL declslon musL be done ln collaboraLlon wlLh Lhe sLakeholders (#3), and LhaL lnformaLlon ls sLored ln Confluence for fuLure reference. AL Lhe end of every SprlnL, Lhe C gaLhers wlLh Lhe Leam and Lhe Scrum MasLer (aL Lask #6) Lo revlew Lhe pro[ecL's progress and Lhe Leam demonsLraLes Lhe new funcLlonallLles Lo Lhe C. 8y dolng Lhe meeLlng, Lhe C has anoLher Lool Lo conLrol Lhe progress of Lhe pro[ecL and make managemenL declslons upon Lhe ouLcome of Lhe meeLlng. 1hese resulLs should be sLored ln Confluence. lL may be posslble LhaL noL all of Lhe sLorles of Lhe sprlnL are dealL wlLh accordlng wlLh Lhe ob[ecLlves of each sLory, for lnsLance lmplemenLlng someLhlng on an esLlmaLed Llme, or even wlLh Lhe goals of Lhe sprlnL, resulLlng ln a relmplemenLaLlon of Lhe sLory. 1hls may lead Lo more Llme belng consumed so oLher sLorles may have Lo be dropped. ln Lhls slLuaLlon lL ls necessary Lo check Lhe accepLance of Lhe sLakeholders (#7). ln Lhe poslLlve case, changes are made ln Lhe producL backlog (#8), sorLed accordlng Lo Lhe lmporLance Lhe C glves Lo Lhe sLory. ln Lhe negaLlve case, changes are done neverLheless, buL no oLher sLorles are dropped, slnce lL was Lhe llablllLy of Lhe company Lo reach Lhe seL goal ln Lhe esLlmaLed Llme. lL ls worLh polnLlng ouL LhaL Lhls slLuaLlon should noL occur wlLh Lhe lmplemenLaLlon of Lhese new processes, as Lhe daLa LhaL supporLs Lhe esLlmaLes ls more compleLe.
40 5"5"5"!" =2,/)*0 1he lnformaLlon gaLhered ln Lhe LM can be showed ln meLrlcs LhaL can be of Lwo caLegorles: JLRKFCPIJOE, a slmpler meLrlc LhaL lndlcaLes a value LhaL can be compared wlLh oLher meLrlcs Lo reach a concluslon, and NKCSPFPIJOE, a more compleLe meLrlc LhaL already compares Lwo or more values Lo help Lhe C reach some concluslon. lor Lhls process, Lhe followlng meLrlcs are avallable: "EQENIEY 3KFV - 1hls lnformaLlve meLrlc presenLs Lhe sum of Lhe esLlmaLlons made by Lhe C (C1L) for Lhe lssues selecLed for a sprlnL, /KCCJIIEY 3KFV - 1hls lnformaLlve meLrlc ls a sum of Lhe esLlmaLlons Lhe Leam gave (81L) Lo Lhe sub-Lasks of each of Lhe sLorles selecLed for a sprlnL. 1hls may dlffer from Lhe esLlmaLlon glven by Lhe C Lo Lhe sLory Lhese sub-Lasks belong Lo, slnce Lhese were glven wlLh more flne deLall. 1he C can also examlne Lhe dlscrepancles (lf any) and acqulre lmporLanL knowledge Lo apply ln fuLure pro[ecLs, $YKSIEY 3KFV - 1hls lnformaLlve meLrlc ouLpuLs Lhe sum of Lhe esLlmaLes of sLorles LhaL were lnserLed by Lhe developmenL Leam lnLo Lhe sprlnL backlog (afLer obLalnlng clearance by Lhe C) afLer Lhe sprlnL sLarLed, meanlng LhaL some work was noL properly deflned and addlLlonal lssues were used Lo compleLe Lhe goals of Lhe sprlnL. 8uL Lhls ls noL Lhe only case, Lhe C can also lnserL lssues on Lhe sprlnL backlog, afLer dlscusslng lL wlLh Lhe developmenL Leam. lL may happen for varlous reasons LhaL resulL ln Lhe necesslLy Lo do addlLlonal work, alLhough Lhls wlll lead Lo Lhe dropplng of some oLher lssues, #KIPQ #JCE "SELI - ln Lhls lnformaLlve meLrlc Lhe C can have a compleLe cumulaLlve sum of Lhe work done by Lhe developmenL Leam ln Lhe sprlnL. ln a perfecL world lL would be Lhe sum of Lhe /KCCJIIEY 3KFV wlLh Lhe $YKSIEY 3KFV meLrlcs, buL LhaL ls very rarely Lhe case. 1hls meLrlc wlll asslsL noL only Lhe C, buL also Lhe developmenL Leam Lo help deLermlne whaL wenL wrong wlLh Lhelr esLlmaLlons, #JCE .HIJCPIE OH< #JCE "SELI - WlLh Lhls comparaLlve meLrlc, Lhe C can assess Lhe dlfference beLween Lhe sum of esLlmaLes glven by Lhe Leam and Lhe sum of whaL was acLually spenL ln Lhe developmenL. 1hls conLrol meLrlc allows Lhe C Lo deLermlne whlch lssues are Lyplcally under- budgeLed (or over-budgeLed) and acL accordlngly, by dlscusslng Lhls wlLh
41 Lhe Leam. 1he C also obLalns lmporLanL lnformaLlon LhaL can be used when glvlng Lhe Crlglnal 1lme LsLlmaLe Lo fuLure pro[ecLs LhaL have slmllar lssues, Whlle mosL of Lhese meLrlcs can be used for conLrolllng purposes Lo ascerLaln Lhe sLaLus of Lhe sprlnL, Lhey serve mosLly as a learnlng Lool, slnce Lhey are malnly obLalned aL Lhe end of Lhe sprlnL, when concluslons can be drawn. lor acLual and real-Llme" conLrol Lhe C has Lhe dallles and speclally Lhe burn-down charL, descrlbed ln Lhe nexL chapLer. WlLh Lhe resulLs obLalned from Lhese meLrlcs, Lhe C can moLlvaLe Lhe developmenL Leam Lo perfecL Lhelr work or Lo celebraLe whaL was achleved. 5"5"5"5" ;0+62 %. ,72 <9= 8egardlng Lhe monlLorlng of Lhe sprlnL's progress, Lhe C has Lhe burn-down charL ln !l8A whlch ls a Lool LhaL dlsplays Lhe worklng hours belng logged by Lhe developmenL Leam on Lhe sLorles of each sprlnL.
lmage 9 - 8urn-down charL lf Lhe lndlcaLlve red llne goes up Lhe grey dlagonal, lL means LhaL Lhe work ls geLLlng delayed and Lhe goals of Lhe sprlnL may be ln [eopardy. Pere Lhe C has Lwo opLlons: elLher go dlrecLly Lo Lhe Leam and Lalk abouL Lhe delay, or aLLend Lhe dally meeLlng Lo Lry Lo undersLand Lhe reason of Lhe delay. AfLerwards, lf necessary, Lhe C can acL upon LhaL lnformaLlon, as descrlbed on Lhe process model. 1he lnformaLlon collecLed ln Lhls process wlll also be sLored ln Confluence and has Lhe followlng sLrucLure:
42
lmage 10 - lnformaLlon sLrucLure of Lhe uevelopmenL process 1he flrsL page, ro[ecL managemenL holds Lhe pro[ecL lnformaLlon regardlng Lhe sprlnL. ! "SFJLI (LRKFCPIJKL - Cn Lhls page, Lhe C should lnserL ln a Labular form Lhe general lnformaLlon abouL each sprlnL, gaLhered along Lhls process - lLs goals, sLarL and end daLe and oLher relevanL lnformaLlon. 1he flnanclal deparLmenL can use Lhls lnformaLlon Lo blll based on Lhe Lerms of Lhe pro[ecL's conLracL, ! "SFJLI 'EOJEU - 1hls page sLores Lhe lnformaLlon abouL all Lhe sprlnL revlews LhaL are made aL Lhe end of Lhe sprlnL. AL Lhe sprlnL revlew, Lhe Leam members dlscuss how Lhe sprlnL wenL and polnL ouL lssues (lf any) LhaL could affecL fuLure sprlnLs or even Lhe flnal producL, LhaL need Lo be addressed and by whom should Lhey be addressed, ! "SFJLI 'EIFKHSENIJOE - 1he lnformaLlon on Lhls page ls gaLhered on Lhe sprlnL reLrospecLlve, whlch Lakes place afLer Lhe revlew, ln whlch Lhe Leam members reflecL on evenLs LhaL happened ln Lhe sprlnL, how Lo lmprove Lhe developmenL of Lhe pro[ecL and whaL was learn ln Lhe sprlnL, for lnsLance whaL was correcLly or wrongly execuLed, oLher ldeas Lo LesL and even pralse someone LhaL dld a really greaL [ob, ! "SFJLI 0EIFJNH - 1hls page dlsplays all Lhe lnformaLlon Lhe LM sysLem collecLed ln Lhe sprlnL. 1hls lnformaLlon ls presenLed by Lhe Scrum MasLer or by Lhe C who curaLes Lhe daLa lnLo perspecLlve wlLh oLher sprlnLs Lo
43 allow a crlLlcal ouLlook of Lhe Leam's performance Lo glve room for lmprovemenLs, ! ,FKMENI PLY ,FKYBNI "NKSE - 1hese pages are sLored under Lhe umbrella page ,FKMENI JLRKFCPIJKL creaLed on prevlous processes. Lach Llme Lhere ls a change ln Lhe scope, pro[ecL or producL, Lhe lnformaLlon should be lnserLed on Lhls page, wlLh Lhe reply glven by Lhe sLakeholders accepLlng Lhe scope change. AlLhough a pro[ecL scope change could happen, LhaL would mean LhaL Lhe requlremenL assessmenL falled conslderably and may lead Lo a ma[or redeslgn of Lhe pro[ecL. AfLer Lhe sprlnL ls done, Lhe C can declde lf Lhe currenL sLage of Lhe pro[ecL ls slgnlflcanL Lo demo Lo Lhe sLakeholders. lf so, anoLher process ls seL ln moLlon, Lhe &ECK process. ><><@< &ECK SFKNEHH When a sprlnL ls flnlshed, Lhe C needs Lo deLermlne lf a demo wlLh Lhe pro[ecL's sLakeholders ls necessary, meanlng Lhere are funcLlonallLles or lnformaLlon LhaL Lhe sLakeholder can vlew and glve addlLlonal lnpuLs Lo lmprove Lhe producL.
lmage 11 - uemo process 5"5">"#" 9/%*200 82,+):0 up fronL, on Lhe #1 Lask Lhe necesslLy or noL of Lhe demo ls esLabllshed. 1he C should look lnLo Lhe goals of Lhe sprlnL, as well as Lhe goals and resulLs of Lhe prevlous
44 sprlnLs (lf any) Lo help deLermlne LhaL necesslLy. ln case Lhere ls someLhlng consLrucLlve Lo show Lo Lhe sLakeholders, Lhe C, LogeLher wlLh elemenLs of Lhe developmenL Leam, prepares Lhe demo (#2). 1hls demo ls an exLenslon of Lhe demo Lhe Leam had done for Lhe C, whlch here as Lhe role of presenLer. 8y dolng Lhe demo, Lhe C wlll galn a hollsLlc vlew of Lhe pro[ecL: lLs sLaLus, any dlfflculLles LhaL may noL have been prevlously dlscussed and deLermlne Lhe necesslLy of change. 1hese lnslghLs occur [usL aL Lhls sLage, because ln Lhe sprlnL demo, Lhe C only had a feaLure orlenLed vlew of Lhe pro[ecL. 1he demo of Lhe producL musL (lf lL ls Lechnlcally posslble or feaslble) be prepared Lo be presenLed ln Lhe same server used aL Lhe end of Lhe sprlnL demo. ldeally Lhls should be an exLernal server because Lhe demo wlLh Lhe sLakeholder usually Lakes place ouLslde Lhe company's offlce, or by means of remoLe sesslon. 1hls way, lL should be posslble for Lhe sLakeholders Lo lnLeracL wlLh Lhe producL belng demoed. Any lnformaLlon LhaL ls gaLhered aL Lhls Lask should be sLored ln Confluence. AL Lhe nexL Lask (#3), Lhe demo Lakes place wlLh Lhe sLakeholders. LlemenLs of Lhe developmenL Leam can be called ln Lo help wlLh Lhe demo. WlLh lL, Lhe sLakeholders have a beLLer grasp on Lhe progress of Lhelr producL, and aL Lhls polnL, may requlre some changes, lmprovemenLs or new feaLures. lf LhaL ls Lhe case, Lhe C musL analyze wheLher Lhe changes are wlLhln Lhe scope of Lhe pro[ecL (Lask #4) elaboraLed ln Lhe &ERJLJIJKL process or noL. lf Lhey fall wlLhln Lhe scope of Lhe pro[ecL, Lhe C should lnpuL Lhose changes onLo Lhe pro[ecL's backlog aL Lask #7, make Lhe necessary ad[usLmenLs Lo Lhe sorLlng order of sald backlog and lnform Lhe developmenL Leam of Lhose changes, Lo allow Lhem Lo analyze Lhe changes and add Lhelr esLlmaLlons (81L). 8uL lf Lhe changes are beyond Lhe scope of Lhe pro[ecL, lL ls necessary for Lhe C Lo declde (#3) along wlLh Lhe sLakeholders whaL acLlon should be Laken. Cne opLlon wlll be Lhe lnserLlon of changes on Lhe pro[ecL's backlog resulLlng on Lhe drop of some oLher sLory lnsLead, allowlng Lhe process Lo proceed Lo Lask #7. lf noL, Lhe commerclal deparLmenL (#6) musL lssue an addendum Lo Lhe cosL of Lhe pro[ecL Lo accommodaLe Lhe addlLlonal change. Cnly lf approved, are Lhe changes added Lo Lhe producL backlog on Lhe Lask #7. 5"5">"!" =2,/)*0 %BCDEF KR "NKSE /TPLWEH JL ITE ,FKMENI - AfLer Lhe demo, lL ls valld LhaL sLakeholders deLermlne Lhe necesslLy of change. 1he lnformaLlon of Lhese
43 changes ls recorded ln Confluence wlLh a correcL conLexLuallzaLlon, maklng Lhls lnformaLlve meLrlc useful Lo help lmprove Lhe requlremenLs' assessmenL phase ln fuLure pro[ecLs. 1he lnformaLlon gaLhered for Lhls meLrlc musL be lnpuLLed manually. 1he meLrlc descrlbed above can be used for conLrolllng purposes for Lhe pro[ecL aL hand, buL also for learnlng and lmprovemenL, slnce Lhe C can ascerLaln from Lhe changes ln Lhe scope whaL were hls / her mlsLakes ln gaLherlng LhaL lnformaLlon ln earller processes and lmprove LhaL ln fuLure pro[ecLs.
5"5">"5" ;0+62 %. ,72 <9= ln Lhls process Lhe lnformaLlon collecLlon ls mosLly done on Confluence. lf some changes are approved, Lhey are added Lo Lhe pro[ecL's backlog on !l8A. 1he sLrucLure of Lhe lnformaLlon LhaL resulLs from Lhls process should be ln Lhe followlng manner:
lmage 12 - lnformaLlon sLrucLure of Lhe uemo process 1he flrsL page, ro[ecL managemenL, holds Lhe pro[ecL lnformaLlon regardlng Lhe demo lnformaLlon. ! &ECK (LRKFCPIJKL - Cn Lhls page, Lhe C should lnserL ln a Labular form Lhe general lnformaLlon abouL each demo LhaL ls performed for Lhe sLakeholder. 1he lnformaLlon sLored can be Lhe daLe of Lhe demo, access daLa Lo Lhe plaLform, eLc.,
46 ! ,FKMENI 0EIFJNH - 1hls page dlsplays Lhe pro[ecL meLrlcs lnformaLlon for Lhls process. 1hls page also holds Lhe lnformaLlon of Lhe meLrlcs descrlbed ln Lhe prevlous chapLer, 1he Lhlrd page, ro[ecL lnformaLlon ls Lhe same holder page referenced on Lhe &JHNKOEFZ and &ERJLJIJKL processes. lor Lhls process, Lhe lmporLanL pages are Lhe ,FKMENI HNKSE and ,FKYBNI HNKSE pages (see chapLer 3.3.2.1 for Lhe dlfferences beLween Lhem), LhaL have been deflned ln Lhe &ERJLJIJKL process. 1hese are only relevanL lf Lhere are changes approved by Lhe sLakeholders and Lhe C regardlng Lhe budgeL. lf such ls Lhe case, Lhe lnformaLlon regardlng Lhe change should be lnserLed ln Lhe approprlaLe page (dependlng lf Lhe change ls on ro[ecL or roducL scope). When Lhls process ls compleLed, and no more developmenL sprlnLs exlsL (LhaL lead Lo Lhe compleLlon of Lhe &EOEQKSCELI process), Lhe pro[ecL comes Lo Lhe lasL process, Lhe &EQJOEFZ process. ><><_< &EQJOEFZ SFKNEHH 1he lasL process, called &EQJOEFZ ls Lhe flnal process of Lhe LoLallLy of Lhe pro[ecL or of a currenL phase lf Lhe pro[ecL has been dlvlded ln phases. lL occurs when Lhe producL ls ready Lo be dellvered (deployed). ln Lhls process lL also occurs: Lhe elaboraLlon of Lechnlcal documenLaLlon and Lhe Lralnlng sesslon (lf such componenLs were sold), Lhe pro[ecL (or phase) revlew Lo wrap-up Lhe developmenL ln whlch an analysls ls done on pro[ecL evoluLlon and ouLcome and whaL could be done Lo lmprove Lhe efflclency of Lhe processes (elLher of pro[ecL managemenL or developmenL ln naLure). 1he process ls deplcLed below:
47 lmage 13 - uellvery process
48 5"5"?"#" 9/%*200 82,+):0 ln Lhls flnal process, flrsLly lL ls necessary Lo perform a flnal check (Lask #1) of Lhe producL on Lhe pre-producLlon server used for Lhe demos, uslng Lhe accepLance LesLs deflned ln Lhe &ERJLJIJKL process. 1hese LesLs are creaLed ln a !l8A componenL, Lhe Zephyr, whlch can sLore Lhe lnformaLlon of Lhe LesL sLaLus and provldes a dashboard wlLh Lhe relevanL lnformaLlon of Lhe LesL cycle. lf any of Lhe LesLs falled on Lask #2, Lhe [usLlflcaLlons are collecLed on Lhe LM sysLem. ln Lhe case a hoLflx (qulck flx done by Lhe developmenL Leam) can be lmplemenLed wlLhouL Lhe necesslLy of a new sprlnL, lL ls asked of Lhe developmenL Leam (#3) Lo do so. CLherwlse Lhe process ends and Lhe pro[ecL reLurns Lo Lhe &EOEQKSCELI process, slnce anoLher sprlnL / oLher sprlnLs of developmenL musL be done. Cn Lhese Lasks, Lhe developmenL Leam and sLakeholders are lnvolved Lo valldaLe Lhe resulLs of Lhe LesLs and posslble errors. lf Lhe LesLs have all been successfully passed, Lhe nexL Lask (#4) ls Lo deLermlne lf lL ls necessary Lo wrlLe documenLaLlon for Lhe producL (lf LhaL dellverable was sold). ln poslLlve case, sald documenLaLlon ls elaboraLed ln Lask #3 wlLh Lhe collaboraLlon of Lhe developmenL Leam, as Lhey possess much of Lhe knowledge of Lhe lnner worklngs of Lhe producL. lf no documenLaLlon ls necessary, Lhe flow passes Lo Lask #6 where lL ls deLermlned lf one or more Lralnlng sesslons were sold. ln poslLlve case, Lhey are prepared (#7) ln con[uncLlon wlLh Lhe developmenL Leam. 1he Lralnlng sesslon(s) (#8) can lnvolve one or more sLakeholders, as well as elemenLs of Lhe developmenL Leam, whlch can even be Lasked Lo presenL Lhe Lralnlng sesslon Lhemselves. AfLerwards Lhe sLakeholders evaluaLe (#9) Lhe sesslon Lo raLe Lhe quallLy of Lhe Lralnlng, ln order Lo lmprove lL. AfLer Lhe Lralnlng sesslon, or lf Lhe sesslon was noL necessary, Lhe acLual dellvery (#10) of Lhe producL on Lhe producLlon server Lakes place. AfLerwards, an addlLlonal LesLlng sesslon ls employed (#11), uslng Lhe same accepLance LesLs, Lhls Llme ln Lhe producLlon server. lf Lhe LesLs are noL meL wlLh success, Lhe causes musL be deLermlned along wlLh Lhe developmenL Leam (#12), an analysls musL be done abouL Lhe posslblllLy Lo develop and deploy a hoLflx. lf Lhls ls noL posslble, Lhen Lhe process ends and Lhose problems musL be quenched ln a reLurn Lo Lhe &EOEQKSCELI process. lf a hoLflx ls posslble, Lhen lLs deploymenL (#10) ls requesLed Lo Lhe developmenL Leam (#13) and Lhe process goes Lhrough Lo Lask #11 of Lhe flnal accepLance LesLs.
49 When all of Lhe flnal LesLs are meL wlLh success, Lhe C should lnform (#14) Lhe flnanclal and commerclal deparLmenL of Lhe dellvery of Lhe producL so Lhey are able Lo conLlnue Lhelr own processes. 1he flnal lssue (#13) represenLs a flnal pro[ecL / phase reLrospecLlve lnvolvlng Lhe developmenL Leam and Lhe Scrum MasLer, Lo reflecL on Lhe progresslon and resulLs of Lhe pro[ecL / phase. 1hls way lL may be posslble Lo Lransform Lhe LaclL knowledge galned on Lhe pro[ecL / phase Lo a more expllclL knowledge ln Lhe form of lmprovemenLs ln Lhe processes, wheLher Lhey are of managemenL or developmenL naLure. As deplcLed by Lhe dlagram, documenLaLlon ls produced on every acLlvlLy Lo quench Lhe problem ldenLlfled ln Lhe lnLervlews: lack of documenLaLlon and communlcaLlon. Whlle Lhe aglle manlfesLo upholds LhaL documenLaLlon should noL be a focus, lL ls sLlll much necessary Lo guaranLee Lhe safekeeplng of Lhe lnformaLlon. 5"5"?"!" =2,/)*0 %BCDEF KR IJCEH ITE PNNESIPLNE IEHIH RPJQEY - WlLh Lhls lnformaLlve meLrlc Lhe C as well as Lhe developmenL Leam can have a beLLer grasp on how many Llmes Lhe pre-deploymenL accepLance LesLs falled. 1hls maLLers Lo lmprove noL only Lhe developmenL meLhods (of Lhelr lnLernal LesLs), buL also Lhe accepLance LesLs Lhe C does aL Lhe end of each sprlnL Lo ensure Lhe success of Lhe dellvery of Lhe producL, 0KHI IZSJNPQ NPBHEH KR PNNESIPLNE IEHIH RPJQBFE - 1hls lnformaLlve meLrlc wlll gaLher all Lhe reasons for falllng ln Lhe lnlLlal LesLs, maklng lL posslble Lo deLermlne whlch are Lhe mosL Lyplcal causes Lo prevenL Lhem from happenlng ln fuLure pro[ecLs, %BCDEF KR IJCEH ITE RJLPQ PNNESIPLNE IEHIH RPJQEY - AfLer Lhe dellvery, anoLher baLch of LesLs musL be done by Lhe sLakeholders. 8y uslng Lhls lnformaLlve meLrlc, lL ls posslble for all Lhe parLles lnvolved Lo deLermlne Lhe quallLy of Lhe producL. 1he measuremenLs for Lhls meLrlc musL be lnserLed manually on Confluence, as Lhe LesLs are performed ouLslde of Lhe LM sysLem, 0KHI IZSJNPQ NPBHEH KR RJLPQ PNNESIPLNE IEHIH RPJQBFE - As ln Lhe prevlous meLrlc, Lhls lnformaLlve meLrlc can be used Lo unsure LhaL repeaLlng causes LhaL occur afLer Lhe producL ls dellvered can be quenched and
30 prevenLed ln Lhe fuLure. Slmllar Lo Lhe prevlous meLrlc, Lhe measuremenLs musL be lnserLed manually on Confluence, 'PIJK KR IJCEH ITE PNNESIPLNE IEHIH RPJQEY OEFHBH ITE IJCEH ITE RJLPQ PNNESIPLNE IEHIH RPJQEY - 1hls comparaLlve meLrlc wlll allow deLermlnlng lf Lhe flnal LesLs are havlng a superlor number of fallures Lhan Lhe lnlLlal LesLs. WlLh Lhls lnformaLlon Lhe developmenL Leam and Lhe C can ascerLaln Lhe cause and Lry Lo mlLlgaLe Lhe problem. Cne cause could be LhaL a Lechnlcal requlremenL was noL gaLhered ln Lhe approprlaLe process and, wlLhouL LhaL lnformaLlon, Lhe error occurred. 8ecause Lhe measuremenLs of Lhe flnal LesLs are lnserLed manually, Lhe measuremenLs for Lhls meLrlc musL also be lnpuLLed ln Lhe same manner on Confluence, "IPIJHIJNPQ PLPQZHJH KR ITE EOPQBPIJKL KR ITE IFPJLJLW HEHHJKL - WlLh Lhls lnformaLlon Lhe C can have a beLLer undersLandlng of Lhe effecLlveness of Lhe Lralnlng sesslon and can use Lhls lnformaLlon for fuLure lmprovemenLs. Llke ln uevelopmenL process, Lhese meLrlcs are Lransversal ln Lhelr naLure, slnce lL ls posslble Lo use Lhem Lo conLrol Lhe level of errors belng done ln Lhe process, buL also Lhey can be used for moLlvaLlon purposes based on Lhe resulLs obLalned. 1hese meLrlcs are moreover capable of permlLLlng lmprovemenLs on fuLure pro[ecLs. 5"5"?"5" ;0+62 %. ,72 <9= 1he lnformaLlon collecLed on Lhls process wlll be sLored ln Lhe LM sysLem ln Lhe followlng sLrucLure:
31
lmage 14 - lnformaLlon sLrucLure of Lhe uellvery process 1he page 1echnlcal lnformaLlon, also menLloned on Lhe &ERJLJIJKL process, holds Lhe gaLhered lnformaLlon on Lhls process. 1he followlng pages should be lnserLed ln Confluence. ! $NNESIPLNE IEHIH FEHBQIH - 1hls page should be used Lo sLore all Lhe resulLs of Lhe accepLance LesLs done ln Lhls process, Lhe lnlLlal and Lhe flnal ones, ln a Labular form. ln case of fallure ln one or more LesLs, Lhe causes should be polnLed ouL. ! #ENTLJNPQ YKNBCELIPIJKL - 1hls page only exlsLs lf lL ls deLermlned LhaL Lhe Lechnlcal documenLaLlon ls a dellverable LhaL was agreed Lo be produced. AfLer Lhe documenLaLlon ls concluded on Lhls page, Confluence allows exporLlng seLs of pages as a ul Lo dellver Lo Lhe flnal cllenL, ! #FPJLJLW YKNBCELIPIJKL - 1hls page wlll exlsL only lf Lhe Lralnlng was an opLlon sold ln Lhe pro[ecL and should hold all Lhe lnformaLlon and maLerlals regardlng Lhe Lralnlng, ! &EQJOEFZ JLRKFCPIJKL - All Lhe relevanL lnformaLlon abouL Lhe deploymenL of Lhe producL on Lhe producLlon envlronmenL should be sLored on Lhls page. 1he ro[ecL managemenL page, descrlbed ln Lhe prevlous processes, holds Lhe ,FKMENI 0EIFJNH page LhaL sLores Lhe lnformaLlon abouL some meLrlcs. Slnce Lhe lnlLlal LesLs are done uslng a plugln of !l8A, Zephyr, Lhe lnformaLlon ls dlsplayed ln Lhe Zephyr dashboard.
32 Also, lL ls worLh polnLlng ouL ln case Lhe accepLance LesLs ouLpuLLed errors, glven LhaL Lhey were Lhe resulLs of an error of developmenL, a new sLory should be creaLed on !l8A wlLh Lhe relevanL lnformaLlon Lo correcL Lhem. WlLh Lhe concluslon of Lhls process and Lhe dellvery of Lhe producL, Lhe pro[ecL or a phase of Lhe pro[ecL can be consldered compleLe from Lhe company's polnL of vlew. ><@< *7#$(%(%1 #?. 0.$"5'.0.%#" When sLarLlng a pro[ecL, Lhe C should bulld Lhe sLrucLure descrlbed ln Lhe prevlous chapLers ln Lhe LM. neverLheless, Lhe pages relaLed wlLh meLrlcs lnformaLlon can only be bullL aL Lhe Llme Lhe acLual developmenL sLarLs, when Lhe measuremenLs begln and Lhe meLrlcs Lake meanlng. Lach of Lhe meLrlcs ls presenLed on Lhe Confluence appllcaLlon Lhrough Lhe use of a speclal macro, enLlLled -('$ (HHBEH. WlLh Lhose Lables ln place, meLrlcs wlll be avallable as Lhe daLa ls lnpuLLed. lL ls apparenL LhaL Lhe LM sysLem, Lhrough !l8A and Confluence provldes good resources Lo obLaln Lhe measuremenLs of some meLrlcs, buL noL all, especlally Lhe ones from Lhe &EQJOEFZ process LhaL do noL rely on !l8A lssues, creaLlng Lhe necesslLy of manual lnpuL for Lhe measuremenLs. 1hls lssue wlll be furLher analyzed ln Lhe LlmlLaLlons sub-chapLer of Lhe ulscusslon. Also lmporLanL Lo polnL ouL ls LhaL, whlle Lhe lmplemenLaLlon of Lhe processes descrlbed above could be done somewhaL easlly, slnce Lhe sysLem LhaL wlll supporL lL ls already ln use, obLalnlng Lhe daLa (measuremenLs) necessary Lo fuel Lhe meLrlcs Lo reLrleve some useful lnformaLlon ls more dlfflculL. lL ls necessary Lo change Lhe worklng hablLs of Lhe developmenL Leams for Lhem Lo sLarL lnpuLLlng valld worklng Llmes. 1o do Lhls, verbal lncenLlve musL be presenL aL all Llmes by Lhe already ldenLlfled Lechnologlcal leader - one of Lhe Managlng arLners. 8uL Lhls role of champlon can also be played by Lhe C of a pro[ecL or even Lhe Scrum MasLer asslgned Lo Lhe Leam, slnce Lhey are ln consLanL conLacL wlLh Lhe Leam and can encourage Lhe accepLance of Lhese changes. Cn a concluslon noLe, by obLalnlng Lhese meLrlcs, Lhe C has a good sLarLlng polnL LhaL enables hlm / her Lo deLermlne lf Lhe pro[ecL ls a success or noL. 1haL deLermlnaLlon should be done bearlng ln mlnd Lhe sLraLeglc goals of Lhe company, buL malnly Lhe goals of Lhe pro[ecLs, whlch vary from pro[ecL Lo pro[ecL and should be deLermlned aL Lhe beglnnlng of Lhe pro[ecL.
33 @< &("/5""(*% upon compleLlng Lhe flndlngs of Lhls work, lL ls necessary Lo deLermlne Lhe lmpllcaLlons lL has on a LheoreLlcal level buL also on a more pracLlcal scope. lor Lhe LheoreLlcal angle, Lhe llnkages beLween Lhe flelds of pro[ecL managemenL, buslness process managemenL and managemenL of lnformaLlon sysLems are done. As for Lhe pracLlcal dlscusslon, conslderaLlons are done regardlng Lhe flve processes LhaL were modeled, Lhelr use ln Lhe pro[ecLs and Lhe valldlLy proven by Lhe use of Lhe meLrlcs. Concludlng Lhls chapLer, some recommendaLlons, some llmlLaLlons of Lhls work and posslble follow up works are presenLed. @<;< #?.*'.#(/$4 (0,4(/$#(*%" lrom a LheoreLlcal perspecLlve, Lhe lmpllcaLlons can be dlvlded ln Lhe Lhree flelds of Lhls work. lor Lhe fleld of pro[ecL managemenL, Lhls work presenLs several crlLlcal success facLors LhaL musL be observed as a means Lo resolve Lhe lssues polnLed ouL: NKCCBLJNPIJKL, FE^BJFECELIHA PHHEHHCELI and Lhe FJWTI PCKBLI KR YKNBCELIPIJKL. Clven Lhe aglle meLhodologles depend on Lhe LaclL knowledge Lo be passed on face-Lo-face communlcaLlon, many of Lhe relevanL lnformaLlon regardlng a pro[ecL can be losL. ln order Lo obLaln a coherenL communlcaLlon, Lhls framework presenLs a seL of rules on how Lo sLrucLurally sLore Lhe lnformaLlon, whlle keeplng ln mlnd anoLher crlLlcal facLor, Lhe rlghL amounL of documenLaLlon. 1hese lnformaLlon sLrucLures are meanL Lo supporL Lhe C of a pro[ecL Lo make pro[ecL relaLed declslons, buL also Lhe developmenL Leam. SLlll regardlng Lhe crlLlcal facLors, Lhe requlremenLs' assessmenL, Lyplcally done by Lhe C of a pro[ecL, ls presenLed ln an aglle way, a paradlgm shlfL lf lL may, by lnvolvlng Lhe developmenL Leam rlghL aL Lhe beglnnlng of Lhe pro[ecL, lnsLead of [usL on Lhe developmenL sLage. WlLh Lhls earller lnvolvemenL, Lhe C can use Lhelr LaclL knowledge Lo obLaln a more accuraLe assessmenL on whaL ls requlred of Lhe pro[ecL and lLs cosL. A 8M framework was used as a sLarLlng polnL on Lhe modellng of Lhe processes presenLed on Lhe +JLYJLWH chapLer. 1hls hlghly lLeraLlvely framework emulaLes, Lo a cerLaln polnL, Lhe mechanlsms upheld by Scrum. 1he 8M framework revolves around Lhe noLlon LhaL every process has room Lo evolve afLer an analysls of lLs resulLs. AfLer a new plan ls drafLed, an lLeraLlon of Lhe process can be lmplemenLed.
34 8uL Lhe analysls of Lhe resulLs of Lhe processes ls only posslble lf Lhere are meLrlcs feed by measuremenLs Lo ascerLaln Lhe valldlLy of Lhe work belng done. lor LhaL purpose, a serles of meLrlcs was devlsed, as well as Lhe meLhod on how Lo obLaln lLs daLa. 8uL glven LhaL Lhe meLrlcs are only usable lf daLa ls lnserLed (on Lhe LM sysLem), a culLural change musL occur ln Lhe reclplenL company Lo encourage Lhe lnLroducLlon of work-logs by Lhe developmenL Leams. Slnce lnLroduclng Lhls Lype of lnformaLlon wlll glve a greaL deal of scruLlny Lo Lhe dally acLlvlLles of Lhe developmenL Leams, some elemenLs of Lhe Leams may feel susplclous of Lhe purpose of such acLlvlLles and mlghL noL lnserL Lhe lnformaLlon accuraLely. 1o counLeracL Lhe reslsLance Lo change, Lhls work also recommends Lhe employmenL of one or more champlons. 1hese roles should be asslgned Lo people LhaL are percelved noL only as Lechnologlcal orlenLed people, buL also leaders LhaL can moLlvaLe oLhers Lo undersLand LhaL Lhe necesslLy of change ls also beneflclal for everybody, belng a sLep Lo lmprove Lhelr work and Lhemselves. @<=< ,'$/#(/$4 (0,4(/$#(*% 1he framework presenLed on Lhls work can acL as a sLepplng sLone for a company worklng on sofLware developmenL Lo lmprove Lhe efflclency of Lhelr pro[ecL managemenL processes. Whlle Lhls framework was bullL uslng Scrum as a reference, lL ls [usL LhaL, a reference and any oLher aglle meLhodologles (such as exLreme programlng among oLhers) could be used wlLh llLLle modlflcaLlon Lo Lhe processes. Also lmporLanL Lo polnL ouL ls LhaL Lhe lnformaLlon gaLhered Lo help bulld Lhls framework was based on a company whlch had problems LhaL were deemed common on mosL companles. Powever, before embarklng on Lhe lmplemenLaLlon of Lhls framework, an analysls of Lhe lssues occurrlng on Lhe reclplenL company musL be underLaken ln order Lo adapL Lhe framework and lLs processes. Llkewlse, Lhls work presenLs Lhe usage of a Lype of speclflc sofLware, Lhe LM sysLem by ALlasslan. 8uL Lhls sysLem should only be vlewed as a reference, slnce Lhere are several oLher sulLes LhaL can accompllsh Lhe same resulL. ln Lhe lnLervlews done wlLh Lhe Cs, aL phase lll, several lssues regardlng Lhe managemenL of pro[ecLs were ldenLlfled: CommunlcaLlon among all Lhe parLlclpanLs of a pro[ecL (ranglng from Lhe cllenL Lo Lhe developmenL Leam), documenLaLlon scaLLered and unorganlzed, requlremenLs' assessmenL problems and Lhe dlfflculLy Lo ldenLlfy Lhe sLaLus of a pro[ecL.
33 1o address such problems, flve processes for pro[ecL managemenL were modeled: ulscovery, ueflnlLlon, uevelopmenL, uemo and uellvery, already descrlbed ln deLall ln Lhe +JLYJLWH chapLer. lor each process, lL was devlsed a sLrucLure on how Lo sLore Lhe daLa ouLpuLLed from every process, for easy access of Lhe documenLaLlon from all Lhe parLlclpanLs LhaL requlre lL. Pavlng a common sLrucLure for all pro[ecLs wlll allow a more efflclenL flow of lnformaLlon, slnce everybody wlll know where Lo look for lL. Also, maklng Lhe lnformaLlon easlly avallable on an LM wlll allow a beLLer communlcaLlon, slnce Lhe LM should have Lhe necessary mechanlsms Lo alerL oLhers abouL Lhe avallablllLy of relevanL lnformaLlon. lL was ldenLlfled LhaL Lhe LM does noL always respond Lo every necesslLy of a company, as such, some adapLaLlon musL be done Lo elLher Lhe LM or Lhe company. 1hls declslon musL be based on Lhe dlfflculLy of changlng elLher of Lhe componenLs of Lhe company's lS (,FKNEHHEH and ,EKSQE vs. "KRIUPFE). 1he same concluslon could be drawn regardlng Lhe usage of Scrum. 8elng a framework, a company could use only some of lLs componenLs and bend" some premlses lf lL beneflLs Lhem. AddlLlonally, meLrlcs were also ldenLlfled and descrlbed Lo allow Lhe Cs Lo have a compleLe plcLure of Lhe sLaLus of Lhe pro[ecL. neverLheless, slnce Lhese meLrlcs are LargeLed for measurlng Lhe developmenL belng made, lL ls noL posslble Lo have Lhem ln all Lhe processes, speclflcally Lhe flrsL Lwo: ulscovery and ueflnlLlon. 1hls ls because aL Lhese sLages, no acLual developmenL ls done, buL only analysls and esLabllshmenL of goals whlch are noL sulLable Lo be measured. 8uL Lhe lnexlsLence of meLrlcs does noL mean LhaL Lhese processes do noL answer Lhe lssues ldenLlfled. As polnLed ouL earller, Lhe developmenL Leam should be broughL ln Lo Lhe pro[ecL as soon as posslble Lo help Lhe C obLaln a proper requlremenLs' assessmenL, based on Lhelr experlence (LaclL knowledge). WlLh Lhese guldellnes a company has addlLlonal Lools Lo counLer Lhe lack of resources such as Llme, knowledge and money, whlch ls a common slLuaLlon among SMLs and Lhus can permlL Lhem Lo sLrlve ln Lhelr markeL. @<>< 4(0(#$#(*%" $%& './*00.%&$#(*%" Whlle researchlng for Lhe Loplcs of Lhls work, lL became apparenL Lhe lack of a Lhorough research on Lhe lmpacL aglle meLhodologles could have ln SMLs and how Lhey are belng used by Lhose companles, a Loplc LhaL could have helped ln Lhe maklng of Lhls work. 8elng SMLs Lhe bulk of Lhe companles operaLlng worldwlde,
36 lmplemenLlng Lhls sLudy ln Lhe fuLure could brlng a greaL lnslghL on Lhe currenL sLaLe of Lhe arL ln sofLware developmenL ln SMLs. ln Lerms of Lhe premlse of Lhls work, Lo sLandardlze Lhe pro[ecL managemenL processes, Lhe scope could be expanded ln an efforL Lo lmplemenL a sLandardlzaLlon of oLher processes, as Lhe acLual developmenL. neverLheless, Lhls ls a greaLer efforL (buL noL lmposslble) slnce Lhe lmplemenLaLlon of sofLware has many varlables and each Lechnology has parLlcular aspecLs Lo lL, requlrlng a greaL deal of analysls before embarklng on Lhe modellng of Lhese processes. 1aklng an example from Lhe sub[ecL company of Lhls work, Lhe cholce of Lhe l1 sysLem LhaL wlll supporL Lhese changes ls of greaL lmporLance, Lo check lf Lhe lnformaLlon needed Lo manage Lhe pro[ecLs ls easlly obLalned from Lhe sysLem. lL became apparenL LhaL Lo obLaln and use Lhe meLrlcs Lo assess Lhe valldlLy of Lhe work, Lhe u8l's LM sysLem alone was noL enough, slnce some of Lhe meLrlcs are only obLalned Lhrough manual labor. 1he daLa ls all Lhere, buL Lhere are noL auLomaLlsms avallable Lo dellver Lhem how Lhey are lnLended. lurLhermore, Lhls ls noL a case where Lhe meLrlcs and relaLed processes can be changed as percelved on Lhe LM sysLem, slnce Lhe sysLem does noL provlde by defaulL many meLrlcs Lo evaluaLe Lhe work LhaL ls belng done (whlch lead Lo Lhe developmenL of meLrlcs descrlbed ln Lhls work). uslng Lhe sysLem as ls wlll force Lhe managers Lo conLlnue Lo use excel worksheeLs, a lengLhy and manual operaLlon LhaL lnvolves collecLlng all Lhe requlred lnformaLlon leadlng Lo much Llme belng consumed and sacrlflclng avallablllLy of sald lnformaLlon. A soluLlon for Lhls problem, on u8l, was Lhe developmenL of anoLher layer deployed on Lop of Lhe LM. 1hls new layer can communlcaLe wlLh Lhe LM sysLem Lhough lLs Al Lo auLomaLlcally gaLher all Lhe requlred lnformaLlon Lo feed Lhe meLrlcs and Lhus provlde valuable and Llmely lnformaLlon Lo Lhe pro[ecL owners and oLher managers. 1hls soluLlon ls besL sulLed for Lhe company slnce sald Al was already used Lo lnLegraLe Lhe LM sysLem on oLher Lools used by Lhe company. AnoLher soluLlon could be Lo creaLe a plug-ln Lo Lhe LM sysLem, buL slnce Lhls sysLem lf wrlLLen ln !AvA and Lhe maln programmlng language used on Lhe company ls P, lL was noL a vlable opLlon. 1hese shorLcomlngs furLher hlghllghL Lhe necesslLy Lo perform a careful analysls of whlch sysLem Lo lmplemenL Lo supporL Lhe changes on Lhe company. lf Lhere ls already an LM sysLem ln use and lssues of Lhe same naLure as Lhe ones presenLed ln Lhls work arlse, lL wlll be necessary Lo deLermlne Lhe course of acLlon: can Lhe LM be adapLed or should lL be subsLlLuLed by a new one. lL ls almosL cerLaln LhaL no LM (or
37 slmllar sysLem) wlll flL perfecLly wlLh Lhe processes of a company, so an analysls musL be conducLed comparlng Lhe cosL of changlng Lhe exlsLlng LM and Lhe cosL of adapLlng a new one. ln Lhe case of u8l, a greaL amounL of Llme was already spenL wlLh Lhls sysLem, lLs users are used Lo lL and maklng furLher cusLomlzaLlons ls a far more cosL effecLlve soluLlon for Lhe company. 1hls ls a speclflc soluLlon, speclally lndlcaLed for Lhe sub[ecL company, u8l. Whlle Lhls work presenLs several meLrlcs Lo help quench Lhe problems ldenLlfled ln Lhe lnLervlews and respond Lo Lhe crlLlcal facLors polnLed ouL ln Lhe LheoreLlcal background, more meLrlcs could be lmplemenLed. Cne speclflc meLrlc could be creaLed Lo help Lhe C deLermlne whaL Lhe raLe of compleLlon of a glven componenL ls. 8uL slnce Lhe LM cannoL make calculaLlons of a speclflc group of sLorles, Lhls prevenLs Lhe collecLlon of measuremenLs Lo feed Lhls meLrlc, renderlng lL useless. 1hls slLuaLlon can be resolved wlLh addlLlonal developmenL as already dlscussed. Armed wlLh Lhls framework, Lhe managemenL should draw lLs own concluslons and lmplemenL Lhe changes lL conslders allgned wlLh Lhe company's goals and sLraLegy. Cne opLlon ls Lo embark on a plloL LesLlng wlLh one C and lLs pro[ecLs Lo valldaLe Lhe lmplemenLaLlon of Lhe guldellnes ln close collaboraLlon wlLh Lhe Lop managemenL. AfLerwards, wlLh lessons learned Lhe meLhod can be expanded Lo all Cs and Lhelr pro[ecLs. llnally, lL ls worLh polnLlng ouL one Loplc LhaL was proposed ln a reference for Lhls work, auLhored by 8ruque and Moyano (2007). 1hls work proposes Lhe lmplemenLaLlon of a quallLy sysLem Lo furLher help Lhe lnformaLlon Lechnology adopLlon, Lo work as a faclllLaLor ln Lhe adopLlon of change buL also Lo provlde oLher goals. lor Lhe u8l example, lL could represenL anoLher Lool Lo achleve a long-Lerm goal LhaL ls Lo have Lhe company's pro[ecL managemenL processes cerLlfled ln an lSC sLandard. @<@< +5#5'. 3*'2 AfLer Lhe concluslon of Lhls work, Lhe nexL sLep could be Lhe analysls of Lhe resulLs Lhe changes broughL Lo Lhe sub[ecL company, u8l. WlLh Lhls analysls, a reflnemenL could be done noL only on Lhe processes Lo furLher lmprove Lhem, buL on Lhe lnformaLlon sLrucLure and Lhe meLrlcs belng used. 1he analysls could also answer lf Lhe new developmenLs solved Lhe llmlLaLlons presenL ln Lhe prevlous chapLer.
38 AnoLher analysls LhaL could be done ls Lhe role aglle meLhodologles have on larger companles, wheLher Lhey are belng used aL all and lf Lhe processes proposed by Lhls work could be used on such companles.
39 _< /*%/45"(*%" CollaboraLlon! 1hls work seL ouL wlLh a dlfferenL ldea, efflclency. 8uL whaL was learned from Lhe analysls of all Lhe references collecLed ls LhaL for any efforL of Lhls Lype Lo succeed, ln facL Lo achleve efflclency, collaboraLlon ls paramounL. Cne of Lhe Loplcs dlscussed on Lhe LheoreLlcal background revolves around Lhls, Lo geL Lhe lnpuL of everyone, Lo have champlons Lo lnvoke LhaL ldea among Lhe workers. Cne example ls wlLh Lhe case of 8M, ln many of Lhe references lL ls sLaLed LhaL 8M cannoL guaranLee success on lLself. Changes are necessary, Lo Lransform Lhe mlndseL of Lhe company, buL noL only Lhe workers, Lhe Lop managemenL as well. CommunlcaLlon ls also a key facLor ln all Lhe flelds dlscussed ln Lhls work. WlLhouL lL, Lhe lnformaLlon does noL flow and goals are noL reached ln lLs fullesL. Moreover, lnformaLlon ls of greaL lmporLance for every company and havlng lL prompLly avallable conLrlbuLes Lo lmprove communlcaLlon, buL also permlLs a beLLer sense of collaboraLlon beLween everyone. Cf course a company does noL sLrlfe on collaboraLlon alone, and Lhrough Lhe use of Lhe guldellnes dlscussed on Lhls work, lL has an essenLlal Lool ln Lhe efforL Lo lncrease lLs producLlvlLy and reduce unnecessary cosLs, maxlmlzlng lLs revenue. !usL by reallzlng LhaL change ls necessary, Lhe LransformaLlon ls already happenlng. lL may be a long Lerm pro[ecL, slnce changlng menLallLles and processes ls a Llme consumlng endeavor, buL lL ls a necessary sLep for Lhe paLh Lo achleve Lhe so much referenced efflclency.
60 `< 7(74(*1'$,?6 Al-Masharl, M., & Zalrl, M. (1999). 88 lmplemenLaLlon process: an analysls of key success and fallure facLors. !"#$%&## ()*+&## ,-%-.&/&%0 1*")%-23 4(1), 87-112. 8eLrleved from hLLp://www.bus.lasLaLe.edu/nllakanL/MlS338/8eadlngs/8820 success20and20fallure20facLors.pdf Arlyachandra, 1. 8., & lrollck, M. n. (2008). CrlLlcal success facLors ln buslness performance managemenL-sLrlvlng for success. 5%6*)/-0$*% #7#0&/# /-%-.&/&%03 84(2), 113-120. dol: 10.1080/10380330801941304 8ehn, 8. u. (2003). Why measure performance? ulfferenL purposes requlre dlfferenL measures. ("92$+ -:/$%$#0)-0$*% )&;$&<3 =>(3), 386-606. dol: 10.1111/1340- 6210.00322 8oehm, 8. (2002). CeL ready for aglle meLhods, wlLh care. ?*/@"0&)3 >4(1), 64-69. 8eLrleved from hLLp://academlc.regls.edu/dmmlller/courses/mscc630/r1064.pdf 8ruque, S., & Moyano, !. (2007). CrganlsaLlonal deLermlnanLs of lnformaLlon Lechnology adopLlon and lmplemenLaLlon ln SMLs: 1he case of famlly and cooperaLlve flrms. A&+B%*;-0$*%3 8C(3), 241-233. dol: 10.1016/[.LechnovaLlon.2006.12.003 Chow, 1., & Cao, u. 8. (2008). A survey sLudy of crlLlcal success facLors ln aglle sofLware pro[ecLs. 1*")%-2 *6 D7#0&/# -%: D*60<-)&3 EF(6), 961-971. dol: 10.1016/[.[ss.2007.08.020 lowler, M., & PlghsmlLh, !. (2001). 1he aglle manlfesLo. D*60<-)& G&;&2*@/&%03 H(8), 28-33. 8eLrleved from hLLp://www.pmp-pro[ecLs.org/Aglle-ManlfesLo.pdf Pynes, n. (2009). CorporaLe culLure, sLraLeglc orlenLaLlon, and buslness performance: new approaches Lo modellng complex relaLlonshlps. A&+B%*2*.$+-2 I*)&+-#0$%. -%: D*+$-2 ?B-%.&3 C=(3), 644-631. dol: 10.1016/[.Lechfore.2008.08.014 lnL. (2010). Mlcro, equenas e Medlas Lmpresas em orLugal - 2008. 8eLrleved from hLLp://www.lne.pL/ngL_server/aLLachflleu.[sp?look_parenL8oul=91762214&aLL_ dlsplay=n&aLL_download=y !ackson, S. (2011). CrganlzaLlonal culLure and lnformaLlon sysLems adopLlon: a Lhree- perspecLlve approach. 5%6*)/-0$*% -%: J).-%$K-0$*%3 8F(2), 37-83. dol: 10.1016/[.lnfoandorg.2011.03.003 !arrar, . l., Al-Mudlmlgh, A., & Zalrl, M. (2000). LM( $/@2&/&%0-0$*% +)$0$+-2 #"++&## 6-+0*)#N0B& )*2& -%: $/@-+0 *6 9"#$%&## @)*+&## /-%-.&/&%0O aper presenLed aL Lhe lLLL lnLernaLlonal Conference on ManagemenL of lnnovaLlon and 1echnology
61 kennerley, M., & neely, A. (2003). Measurlng performance ln a changlng buslness envlronmenL. 5%0&)%-0$*%-2 1*")%-2 *6 J@&)-0$*%# P ()*:"+0$*% ,-%-.&/&%03 8>(2), 213-229. dol: 10.1108/01443370310438463 koch, C. (2006). Change ManagemenL - undersLandlng Lhe Sclence of Change. ?5JO+*/. 8eLrleved from hLLp://www.clo.com/arLlcle/24973/Change_ManagemenL_ undersLandlng_Lhe_Sclence_of_Change Llm, M., CrlfflLhs, C., & Sambrook, S. (2010). J).-%$K-0$*%-2 #0)"+0")& 6*) 0B& 0<&%07N 6$)#0 +&%0")7O aper presenLed aL Lhe annual meeLlng of 1he lnsLlLuLe for CperaLlons 8esearch and 1he ManagemenL Sclences. Llndvall, M., MuLhlg, u., uagnlno, A., Wallln, C., SLupperlch, M., klefer, u., kahkonen, 1. (2004). Aglle sofLware developmenL ln large organlzaLlons. ?*/@"0&)3 >C(12), 26- 34. 8eLrleved from hLLp://cs.LxsLaLe.edu/[s236/201203/cs3398/aglle_large_ organlzaLlons.pdf Llnnenluecke, M. k., & CrlfflLhs, A. (2010). CorporaLe susLalnablllLy and organlzaLlonal culLure. 1*")%-2 *6 Q*)2: !"#$%&##3 R4(4), 337-366. dol: 10.1016/[.[wb.2009.08.006 Mlsra, S. C., kumar, v., & kumar, u. (2009). ldenLlfylng some lmporLanL success facLors ln adopLlng aglle sofLware developmenL pracLlces. 1*")%-2 *6 D7#0&/# -%: D*60<-)&3 E8(11), 1869-1890. dol: 10.1016/[.[ss.2009.03.032 Moe, n. 8., Aurum, A., & uyb, 1. (2012). Challenges of shared declslon-maklng: A mulLlple case sLudy of aglle sofLware developmenL. 5%6*)/-0$*% -%: D*60<-)& A&+B%*2*.73 4R(8), 833-863. dol: 10.1016/[.lnfsof.2011.11.006 ro[ecL ManagemenL lnsLlLuLe lnc. (2008). S T"$:& 0* 0B& ()*U&+0 ,-%-.&/&%0 !*:7 *6 V%*<2&:.& W(,!JV T"$:&X N I*")0B L:$0$*% (vol. 1). newLown Square, A: ro[ecL ManagemenL lnsLlLuLe, lnc. 8avesLeyn, ., & 8aLenburg, 8. (2010). Surveylng Lhe crlLlcal success facLors of 8M- sysLems lmplemenLaLlon. !"#$%&## ()*+&## ,-%-.&/&%0 1*")%-23 F=(3), 492-307. dol: 10.1108/1463713101104967 8lslng, L. (2013). ?-))*0# -%: D0$+Y#Z QB-0 5%+&%0$;&# M&-227 Q*)Y[ owerolnL slldes. 8osemann, M., & de 8ruln, 1. (2003). A*<-):# - !"#$%&## ()*+&## ,-%-.&/&%0 ,-0")$07 ,*:&2O aper presenLed aL Lhe LClS 2003 roceedlngs. Schwaber, k. (1997). SC8uM uevelopmenL rocess. ln !. SuLherland, C. Casanave, !. Mlller, . aLel & C. Pollowell (Lds.), !"#$%&## J9U&+0 G&#$.% -%: 5/@2&/&%0-0$*% (pp. 117-134): Sprlnger London.
62 Sheffleld, !., & LemeLayer, !. (2013). lacLors assoclaLed wlLh Lhe sofLware developmenL aglllLy of successful pro[ecLs. ArLlcle. 5%0&)%-0$*%-2 1*")%-2 *6 ()*U&+0 ,-%-.&/&%03 >F(3), 439-472. dol: 10.1016/[.l[proman.2012.09.011 Sllver, M. S., Markus, M. L., & 8eaLh, C. M. (1993). 1he lnformaLlon Lechnology lnLeracLlon model: a foundaLlon for Lhe M8A core course. ,5D \"-)0&)27, 361- 390. Skrln[ar, 8., & 1rkman, . (2013). lncreaslng process orlenLaLlon wlLh buslness process managemenL: CrlLlcal pracLlces'. 5%0&)%-0$*%-2 1*")%-2 *6 5%6*)/-0$*% ,-%-.&/&%03 >>(1), 48-60. dol: 10.1016/[.l[lnfomgL.2012.03.011 Subramanlan, C. P., !lang, !. !., & kleln, C. (2007). SofLware quallLy and lS pro[ecL performance lmprovemenLs from sofLware developmenL process maLurlLy and lS lmplemenLaLlon sLraLegles. 1*")%-2 *6 D7#0&/# -%: D*60<-)&3 E](4), 616-627. dol: 10.1016/[.[ss.2006.06.014 Swanson, L. 8., & Wang, . (2003). knowlng why and how Lo lnnovaLe wlLh packaged buslness sofLware. 1*")%-2 *6 5%6*)/-0$*% A&+B%*2*.73 8](1), 20-31. 8eLrleved from hLLp://LerpconnecL.umd.edu/pwang/Swanson26Wang2003.pdf 1erzlovskl, M., llLzpaLrlck, ., & C'nelll, . (2003). Successful predlcLors of buslness process reenglneerlng (88) ln flnanclal servlces. 5%0&)%-0$*%-2 1*")%-2 *6 ()*:"+0$*% L+*%*/$+#3 ER(1), 33-30. dol: 10.1016/S0923-3273(02)00378-x 1rkman, . (2010). 1he crlLlcal success facLors of buslness process managemenL. 5%0&)%-0$*%-2 1*")%-2 *6 5%6*)/-0$*% ,-%-.&/&%03 >](2), 123-134. dol: 10.1016/[.l[lnfomgL.2009.07.003 Wohed, ., AalsL, W. M. ., uumas, M., PofsLede, A. P. M., & 8ussell, n. (2006). Cn Lhe SulLablllLy of 8Mn for 8uslness rocess Modelllng. ln S. uusLdar, !. lladelro & A. SheLh (Lds.), !"#$%&## ()*+&## ,-%-.&/&%0 (vol. 4102, pp. 161-176): Sprlnger 8erlln Peldelberg. Wymenga, ., Spanlkova, v., 8arker, A., konlngs, !., & CanLon, L. (2012). Lu SMLs ln 2012: aL Lhe crossroads. 8eLrleved from hLLp://ec.europa.eu/enLerprlse/ pollcles/sme/facLs-flgures-analysls/performance-revlew/flles/supporLlng- documenLs/2012/annual-reporL_en.pdf
63 a< $,,.%&(8 a<;< (%#.'9(.3 #'$%"/'(,#" a<;<;< (LIEFOJEU ; ! lnLervlewer: Mlguel erelra ! lnLervlewed: Cludlo Calvo ! uaLe: 1hursday, 8Lh of november of 2012
0,: When a pro[ecL ls sLarLlng, alLhough Lhese days, Lhe processes are consLanLly ln muLaLlon, buL normally when a pro[ecL ls dellvered Lo you, from whlch areas of Lhe company dld lL come? ls lL from Lhe commerclal area? //: 1haL's rlghL, from Lhe commerclal area. 0,: 8uL how ls Lhe lnformaLlon passed, ls lL a meeLlng wlLh Lhe commerclal area where Lhe lnformaLlon ls passed, or Lhere ls dlrecLly a meeLlng wlLh Lhe cllenL Lo deLermlne Lhe requlslLes or boLh or none of Lhls? //: Ck, so normally Lhere ls Lhe followlng. 1he pro[ecLs can arrlve Lo us by Lwo means, or because Lhe commerclal area Lells us LhaL Lhere ls a necesslLy Lo do a plloL pro[ecL, LhaL wlll enLer as a pro[ecL LhaL we musL do ln order Lo demonsLraLe Lo Lhe cllenL (as a proof of concepL), or because we've done a commerclal proposlLlon and we musL lmplemenL a pro[ecL based on LhaL commerclal proposlLlon. 0,: Llke ln a 8equlremenL or someLhlng? //: 1haL's rlghL. 1he cllenL dellvered a llsL of requlslLes ln a proposlLlon. We do a commerclal proposlLlon. Cbvlously we have aL LhaL momenL a serles of meeLlngs Lo geL a beLLer grasp of whaL lL ls lnLended and Lo glve a rough esLlmaLe ln order Lo have an ldea of Lhe cosL of Lhe pro[ecL for Lhe cllenL. And lf Lhe cllenL accepLs, Lhe commerclal proposlLlon ls won and lL Lransforms ln a proper pro[ecL. 1hese are Lhe Lwo ways ln whlch pro[ecLs came Lo us. AfLerwards, whaL happens? When Lhe pro[ecL ls won, normally Lhere ls Lhe phase of pro[ecL managlng, and LhaL phase usually encompasses flve sLages, Lhe one descrlbed by Ml (ro[ecL ManagemenL lnsLlLuLe), whlch are Lhe beglnnlng of Lhe pro[ecL, plannlng of Lhe pro[ecL, execuLlon of Lhe pro[ecL, conLrol of Lhe pro[ecL and closlng of Lhe pro[ecL. And we are golng Lo manage Lhe pro[ecLs ln Lhls manner. We are already dolng Lhls, alLhough noL clearly deflned ln Lhls manner, as a process, buL we are ln Lhe sLage of deflned Lhem clearly Lo undersLand Lhem. WhaL Lasks l have Lo do ln Lhe beglnnlng, ln Lhe plannlng, ln Lhe execuLlon, ln Lhe conLrol and
64 ln Lhe closlng of Lhe pro[ecL. 8aslcally lL ls Lo deLermlne a seL of acLlvlLles Lo be done, who ls responslble by dolng Lhem Lo lmplemenL Lhe pro[ecL. normally ln Lhe beglnnlng of Lhe pro[ecL whaL lL ls done? normally a meeLlng wlLh Lhe cllenL ls conducLed where Lhe scope of Lhe pro[ecL LhaL has been ad[udlcaLed ls presenLed, lLs resLrlcLlons ln Lerms of Llme, a hlgh-level plannlng, whlch ls a plannlng LhaL was already glven Lo Lhe cllenL ln Lhe commerclal proposlLlon, LhaL sLaLes LhaL Lhe pro[ecL wlll Lake Lwo Lo Lhree monLhs (for example) and lL wlll have Lhls sequence of acLlvlLles, lL can be done by sprlnLs, lL wlll be done ln 4 sprlnLs (for example) and Lhe deslgnaLlon of Lhe Leam LhaL wlll work on Lhe pro[ecL and Lhe ldenLlflcaLlon of Lhe sLakeholders on Lhe slde of Lhe cllenL LhaL we wlll need Lo lnLeracL. lrom Lhere we wlll go Lo Lhe sLage of analysls, of dlscovery, rlghL? Cf course, we have an ldea of whaL lL ls, buL we have meeLlngs, lnLervlews and conversaLlons Lo undersLand wlLh more deLall Lhe problem ok? ln order Lo afLerwards bulld Lhe backlog. A backlog glves us a broader scope of Lhe pro[ecL. ln Lhe proposlLlon sLage Lhe cllenL lndlcaLes LhaL he wanLs Lhls and LhaL, and Lhem when we Lalk wlLh Lhe persons lL's noL exacLly LhaL, lL ls a blL Lo Lhe slde and we need Lo reglsLer all LhaL Lo undersLand LhaL ok, from all LhaL we gaLhered ls Lhe beglnnlng (sLage of Lhe pro[ecL). 0,: And aL LhaL sLage can be Lhe slLuaLlon (Lhe cllenL says) LhaL lsn'L exacLly whaL we wanL? //: es. ln Lhls phase of beglnnlng we have a klck-off meeLlng and we have a hlgh-level plannlng and we have an ldenLlflcaLlon of sLakeholders. 1hen ln Lhe followlng sLage, whlch ls a plannlng sLage, we are golng Lo dlscover / analyze furLher, and ln Lhese reunlons we Lry Lo grasp Lhe acLual scope of whaL we have Lo do and we creaLe Lhe backlog, and whlle dolng so we can arrlve Lo a concluslon LhaL whaL Lhe cllenL says he acLually wanLs ln Lhe lnLervlews ls blgger or dlfferenL from Lhe flrsL proposal. lor lnsLance ln Lhe proposal we conLemplaLed four eplcs: we need modules A, 8, C, and lnLegraLlon wlLh sysLem Z. 8uL Lhen we are golng Lo reflne Lhe assessmenL: ln module A whlch are Lhe feaLures, Lhe vlews, Lhe valldaLlon and oLher requlremenLs, and we can flnd we sald Len, Len, flfLeen and LwenLy days, for a LoLal of flfLy flve days ln Lhe proposal buL ln LhaL reflned assessmenL we conclude LhaL lL ls golng Lo acLually Lake sevenLy days, and aL Lhls Llme Lhe assessmenL for an esLlmaLe ls [usL made by Lhe producL owner, because when Lhe Leam geLs lnvolved Lhe esLlmaLes can wlden or shorLen. 8uL lL ls essenLlal aL Lhls sLage LhaL Lhe producL owner acLs accordlngly Lo Lhe dlfference whlch was deLecLed. Pe needs Lo say whlch backlog lLems aren'L golng Lo be done Lo comply wlLh Lhe budgeL or suggesL and addlLlonal proposal Lo comprlse Lhe
63 addlLlonal lLems. 1hls would be ldeal, buL Lhese Lasks arenL done lmmedlaLely, lL Lakes Llme and Lhere are several Lhlngs LhaL we can'L assess aL Lhls sLage, [usL afLer a couple of monLhs 0,: SomeLlmes [usL afLer a couple of sprlnLs, or even aL Lhe Lhlrd . //: 1haL's when we reallze Lhere are some Lasks ouL of Lhe lnlLlal scope, evldenLly as Llme goes by, lL geLs harder Lo negoLlaLe wlLh Lhe cllenL, lf ln Lhe lnlLlal proposal Lhere are sevenLy days he may or may noL accepL lL. When he buys flfLy flve days and lL's golng Lo Lake sevenLy ln Lhe flrsL reunlons Lhere's sLlll a margln Lo negoLlaLe or abandon Lhe pro[ecL, buL afLer some Llme he ls cerLaln LhaL Lhe flfLy flve days comprlse all hls requlremenLs, ln Lhe long run he won'L accepL Lhe sevenLy days. 1hls mlghL happen, buL lL needs Lo be mlLlgaLed. AL Lhe plannlng sLage, afLer Lhese reunlons we can have a more concreLe plannlng, well lf Lhere's all Lhls backlog, we sLarL managlng Lhe Llme and cosL of Lhe pro[ecL. 0,: AL Lhls sLage lL's noL Lhe responslblllLy of Lhe commerclal area. //: 1he commerclal area afLer Lhe momenL of Lhe creaLlon of Lhe pro[ecL, dlsappears: lL's Lhe pro[ecL area: Lhe pro[ecL owner and lLs Leam 0,: 1he commerclal area ls no longer lnvolved wlLh Lhe pro[ecL. //: no, noL dlrecLly wlLh Lhe acLlvlLles, obvlously Lhey mlghL geL called Lo reunlons Lo dlscuss dlfferences ln scope, Lhe can also be called Lo some reunlons ln order Lo dlplomaLlcally help Lhe cllenL when Lhe relaLlonshlp geLs Lense. 0,: 1hey were Lhe flrsL conLacL. //: 1he proposal was Lhelrs, and Lhe cllenLs LrusL Lhem so Lhey can mlLlgaLe problems. 1hey can also be called [usL Lo malnLaln a courLeous relaLlonshlp, and ln Lhe end Lhey can be called Lo access wheLher Lhe cllenL llked Lhe Leam or noL, Lo evaluaLe how Lhe pro[ecL was developed, ln order Lo geL more poslLlve feedback for Lhe Leam Lo galn LrusL for fuLure pro[ecLs. 0,: AL whaL sLages does Lhe flnanclal area geL lnvolved? //: When Lhe commerclal proposal ls done usually Lhere are predeLermlned sLages of bllllng. 1hese bllllng sLages are seL ln Lhe proposal lL mlghL be 30 ln Lhe ad[udlcaLlon, 30 when A ls made or dellvered, and 40 ln Lhe flnal dellvery. lL means Lhe flnanclal area wlll emlL lnvolces of Lhe percenLage values aL predeLermlned daLes, obvlously Lhe flrsL bllllng ls made as soon as Lhe pro[ecL ls ad[udlcaLed, and Lhe lnformaLlon ls passed along Lo Lhe flnanclal area 0,: and how ls Lhls lnformaLlon passed on? lL ls by e-mall or verbally? lsn'L Lhere an auLomaLlc response aL Lhe momenL, Lo deLermlne ln whlch sLage Lhe pro[ecL ls?
66 //: no, lL's always communlcaLlon, Lhe Leam, ln Lhls case Lhe pro[ecL owner, who ls responslble for Lhe pro[ecL managemenL know we've achleved a deLermlned sLage and he communlcaLes LhaL Lo Lhe conLroller or dlrecLly Lo Lhe flnanclal area so as Lo blll aL a cerLaln sLage because we achleved a mllesLone or dellvered someLhlng. SomeLlmes lL ls Lhe flnanclal area LhaL seLs ouL and makes enqulres abouL Lhe sLaLus of Lhe pro[ecL, so Lhey can lssue a blll Lo Lhe cllenL. 0,: So, do Lhey have access Lo deLermlne Lhe sLaLe of Lhe pro[ecL or ls Lhe LlmeLable seL lnlLlally Lhelr guldellne? //: 1hey make a predlcLlon because ln Lhe beglnnlng of Lhe pro[ecL Lhey Lry Lo plan when Lhe dlfferenL sLeps of bllllng wlll occur. When Lhe pro[ecL ls ad[udlcaLed, for lnsLance ln !anuary Lhey ask for a predlcLlon on Lhe followlng sLages, and we glve Lhem an esLlmaLe based on Lhe number of days seL Lo reach Lhe mllesLones and Lhey plan Lo blll accordlngly, for lnsLance Lwo monLhs Lo compleLe Lhe flrsL Lwo Lasks so Lhey plan Lo blll ln March, and ln Lhe end of Lhe pro[ecL esLlmaLed ln four monLhs, so ln May we can blll Lhls. 1hen Lhey conflrm aL Lhe dlfferenL sLages, and we have Lo explaln Lo Lhe conLroller lf Lhe Llmes are belng meL or noL. We should be able Lo explaln every week Lo Lhe conLroller Lhe movlng forward of Lhe pro[ecL aL Lhe esLlmaLe Llme or lf lL's slldlng so Lhe bllllng can be done accordlngly. 0,: ls lL expecLed or deslrable Lhe exlsLence of a Lool ln Lhe new plaLform Lo allow a beLLer communlcaLlon, so Lhe flnanclal area can percelve Lhe flowlng of Lhe pro[ecL? //: We aren'L focuslng ln Lhe Lool as much as ln Lhe process Lo communlcaLe. WheLher or noL we're golng Lo have a Lool LhaL solves Lhls problem l don'L know yeL. 1hls could be done llke Lhls: lnslde Sugar C8M we have Lhe opporLunlLles module where we creaLe our commerclal proposals, Lhere's Lhe pro[ecL module, whlch generaLes a pro[ecL when Lhe commerclal proposal ls won. And lnslde Lhe pro[ecLs, we could have, leLs lmaglne, anoLher module LhaL corresponds Lo Lhe bllllng sLages where we would deLermlne Lhls flrsL bllllng sLage corresponds Lo 30 of Lhe LoLal, Lhls ls 30 and Lhls ls 40 and l could relaLe lL Lo Lhe Llmeframe, Lhus Lhe flnanclal area would only have Lo llsL Lhe bllllng sLages whlch were expecLed Lo occur aL a glven Llme. 1he flnanclal area now has Lhls ln Lhelr flnanclal maps and LhaL ls Lhe reason Lhey don'L ask us abouL lL. lL would be lmporLanL Lo have Lhls, because Lhen Lhe pro[ecL manager would be able Lo look aL hls/her pro[ecLs and creaLe pressure accordlngly, and focus on Lhose pro[ecLs Lo overcome Lhe seLbacks whlch won'L allow Lhe bllllng. 8esldes, when Lhe cllenL asks abouL Lhe pro[ecL, Lhe pro[ecL manager could lnform Lhe sLages of Lhe
67 process and Lhe correspondlng bllllng. Many Llmes Lhe cllenL asks Lhe pro[ecL manager how much ls Lhere Lo blll or whaL was already bllled? 0,: 1hey don'L have LhaL lnformaLlon, because Lhere are several sLakeholders lnvolved. //: 1he producL owner / pro[ecL manager should be lnformed. 1here are oLher posslble scenarlos, lmaglne LhaL we are behlnd schedule because Lhe cllenL lsn'L answerlng, Lhen we can negoLlaLe Lhe bllllng accordlngly, bllllng lL 30, 13, 13, 20, 20, 10 ln order Lo achleve a more deLalled pro[ecL and Lo negoLlaLe oLher mllesLones Lo prevenL noL bllllng 40 because Lhe pro[ecL lsnL reachlng concluslon. lmaglne LhaL because of Lhe cllenL we are Lwo years walLlng for Lhe bllllng of 40. 0,: lndeflnlLely. //: 8uL Lhls ls sLlll [usL an ldea, lL would be lnLeresLlng Lo have Lhls lnformaLlon, Lhere would be much more .. 0,: 1hls plaLform wlll be lnLegraLed wlLh Sugar. //: Sugar has Lhe posslblllLy LhaL Lhe commerclal pro[ecL and Lhe pro[ecL managemenL from Lhe vlew of. why do we puL Lhe pro[ecL here? We puL Lhe pro[ecL here Lo deLermlne LhaL Lhe pro[ecL has a budgeL of 18k, buL lL was sold by 20k, Laklng lnLo accounL LhaL ln Lerms of developmenL cosL lL Lakes 18k, buL Lhere was Lhe commerclal ablllLy Lo sell lL above, alLhough LhaL doesn'L glve us Lhe rlghL Lo Lake 20k, we have Lo do Lhe developmenL wlLh Lhe esLlmaLed 18k, buL as lL was sold by 20k Lhere ls a margln from Lhe commerclal area, buL Lo manage Lhe pro[ecL lL's 18k. 1hen, Lhere should be posslble Lo creaLe prlorlLles ln order Lo deLermlne Lhe mosL lmporLanL pro[ecLs from Lhe many ongolng pro[ecLs, and make Lhem surface and focus on Lhem as Lhey could be essenLlal. lor lnsLance Lhey were low prlorlLlzed lnlLlally because you could end Lhem early, buL because of Lhls Lhey can brlng more pro[ecLs or Lake anoLher pro[ecL wlLh oLher cosLumers or Lhey are ldenLlcal Lo oLhers made. 1hus we have Lo deLermlne Lhls degree of lmporLance, so lL's here on Sugar where we can deflne Lhls degree of lmporLance, ln a commerclal or markeLlng perspecLlve. 0,: 1he Llme plannlng wlll no longer be done here on Sugar llke we use Lo, lL wlll be used [usL for flnanclal analysls... //: LxacLly. 8udgeL and prlorlLy would be Lhe maln lLems buL anoLher lmporLanL feaLure would be Lo deLermlne Lhe sLage of Lhe pro[ecL wheLher lL ls acLlve or noL, for lnsLance Lhe pro[ecL we have Loday on Lhe board (MC) wheLher lL's ln an analyLlcal sLage or ln developmenL
68 0,: 1he sLage wlll be updaLed ln Sugar. //: lL wlll all be managed and updaLed here so LhaL ln Lhe fuLure we can have a dlglLal board whlch shows Lhe pro[ecLs and Lhelr sLages and LhaL a log (a MC log) can be assoclaLed Lo Lhese pro[ecLs. lf someone from Lhe commerclal deparLmenL wanLs Lo know how Lhe pro[ecL ls golng along, he goes Lo Lhe board and looks aL Lhe log. Pe vlews Lhe lnformaLlon LhaL ls collecLed on Lhe MC meeLlngs and he Lhen know Lhe sLage of Lhe pro[ecL. Pe can see lndlcaLors LhaL lndlcaLe lf Lhe pro[ecL ls over-budgeLed, under-budgeLed, whaL ls Lhe percenLage of Lhe concluslon of Lhe pro[ecL, LhaL should be feed (Lhe lnformaLlon) Lhrough Lhe slze of Lhe backlog, lnformaLlon LhaL ls on anoLher plaLform, uslng !l8A and Creenhopper. 0,: ou would have Lhe concluslon.. //: ou have Lhe enLlre backlog and Lhe Llmeframes, you can collecL Lhe lnformaLlon abouL Lhe amounL of Llme remalnlng Lo conclude. 1hen l know Lhe scope of Lhe pro[ecL and l can prompLly undersLand Lhe percenLage of concluslon ln Lhe pro[ecL. And Lhls can be feed auLomaLlcally lf l have lL updaLed, lf Lhere ls a Llme change lL would generaLe an alarm. lL would be good Lo have delLas lf from one week Lo anoLher we change Lhe percenLage, Lhe backlog lncreases, naLurally. uld we negoLlaLe Lhls? 1he pro[ecL manager needs Lo negoLlaLe for more days or noL. So we wanLed Lo have all Lhese lndlcaLors LhaL allow managemenL Lo undersLand how Lhe pro[ecLs are golng, Lhe delLas on a weekly basls lnformlng lf Lhere are more or less Lhlngs. knowlng LhaL Lhe arLlfacLs of Lhe Lasks Lo perform, Lhe Llmes whlch are reglsLered agalnsL Lhe Lasks LhaL exlsL, and Lhe amounL of Lhlngs Lo do (scope, Llme remalnlng and Llme elapsed).1hese are all managed here. 0,: 1hls answer was very comprlslng and you answered a loL of quesLlons, we have also covered Lhe dlfferenL sLages of Lhe pro[ecL. 1he ldea of Lhe sLages ls Lhey are always Lhe same for all Lhe dlfferenL pro[ecLs. //: es, Lhe sLages are Lhe same because any pro[ecL a beglnnlng and an end, any pro[ecL needed plannlng, blg or small lL needs plannlng. All pro[ecLs have an execuLlon and all pro[ecLs need Lo be conLrolled. As far as pro[ecL managemenL goes Lhls ls Lrue for all, buL how much documenLaLlon, how much lnformaLlon you need Lo develop Lo have and show Lo Lhe cllenL l Lhlnk we should devlse a maLrlx. lL's whaL we are Lrylng Lo do nexL, LhaL ls Lo have a maLrlx LhaL Lells us lf Lhe pro[ecL has less Lhan 30 days, low prlorlLy and wlLh a new Lechnology. Lhen whaL applles Lo Lhls pro[ecL ls Lhls seL of acLlvlLles. 0,: And Lhe prlorlLles?
69 //: l don'L know lf Lhese are golng Lo be Lhe crlLerla, Lhere wlll be a maLrlx LhaL shows several hypoLheses, lf Lhe pro[ecL ls of Lhls Lype, Lhen we need Lo do Lhls, lL sLlll has all Lhe sLages seL earller, buL lL sLaLes here you should have a klck off meeLlng and a Leam presenLaLlon, here you don'L, you need very rlgorous plannlng and you need Lo do such and such. 1he closure of Lhese pro[ecLs, for lnsLance blg emblemaLlc pro[ecLs, always has Lhe presence of Lhe commerclal, Lhe company's CLC and whaLever, lL needs some proLocol, lf lL's a small pro[ecL lL doesn'L: Lhe Leam flnlshes Lhe pro[ecL, sends a mall Lo Lhe cllenL upon compleLlon, Lhey recelve and e-mall sLaLlng lL's all ok and Lhe pro[ecL ls closed, for lnsLance lL can have some furLher cosLumer survey Lo Lhe cllenL Lo geL feedback on Lhe pro[ecL, Lhe quallLy, Lhe Leam, Lhe lnLeracLlons. 0,: 1here ls no ldea on Lhe meLrlcs and clear dlvlslons of Lhe maLrlx lLems? //: lL's belng developed, whaL Ana aula ls dolng ls, as she knows Lhe meLrlcs LhaL should be analyzed Cne Llme value" and oLher lndlcaLors LhaL normally are lmporLanL Lo pro[ecL managemenL on Lhe Ml meLhodology and Lhe ob[ecLlve ls LhaL she Lells us whlch are Lhe lndlcaLors you should have, lf Lhls lndlcaLor ls above Lhls value Lhen you should worry abouL Lhls pro[ecL because lL's noL ln conLrol, Lhese are lndlcaLors of execuLlon and conLrol of Lhe pro[ecL. 1hese lndlcaLors are, lf Lhe pro[ecL as such a Lype you should flnlsh llke Lhls, lf Lhere ls anoLher, Lhen ln a dlfferenL way, because Lhere ls a loL of documenLaLlon you aren'L golng Lo make for a flve day pro[ecL and dellver lL. 0,: lL's abouL lLs value. //: lL's noL [usLlflable, only a blg pro[ecL. All Lhe pro[ecLs LhaL surpass a cerLaln value or have a cerLaln degree of lmporLance we have Lo dellver a checkllsL of documenLaLlon. 1hese dellverables are a servlce we provlde and LhaL servlce ls pald for LhaL's why Lhey are more expenslve, we need Lo dellver Lhem Lo [usLlfy Lhe value. AnoLher Lhlng LhaL ls made, we have an esLlmaLed backlog, Lhe people execuLlng Lhe Lasks, all Lhe Leam ls called upon Lo do Lhe plannlng, who creaLes Lhls backlog ls Lhe producL owner, saylng my cllenL wanLs Lhls and l'm wllllng Lo pay Lhls amounL of Llme 0,: So each Lask ls lnlLlally planned. //: es, he has already planned, Lhen comes Lhe Leam and ln Lhe sprlnL plannlng ls golng Lo deLall, maklng a new plannlng, dolng a breakdown, much more Lhorough and we can reach Lhe concluslon abouL Llme esLlmaLes Lo each Lask and Lhen, based on Lhese esLlmaLes, lf Lhere are dlfferences he needs Lo renegoLlaLe wlLh Lhe cllenL and he has Lo manage all Lhese dlfferences. Whlle execuLlng Lhe Leams logs Lhe Llme and says how much Llme sLlll remalns, meanlng Lhe execuLlon and conLrol are always LogeLher.
70 When we perform a Lask we are execuLlng conLrol, because we are saylng whaL's sLlll mlsslng, Lhen everyday Lhere ls evldence. 0,: 1hls conLrol ls made by Lhe Leam lLself. //: lL's Lhe Leam LhaL says whaL's mlsslng. 1he Leam has Lhe noLlon LhaL Lhls should Lake flfLeen days and Loday l spenL one day, buL Lhere are sLlll flfLeen remalnlng, buL afLer one day's work, Lhere are only Len days lefL, because we advanced and Lhe complexlLy dlmlnlshed. Lvery day conLrol lndlcaLors are avallable LhaL say lL's reduclng or rlslng. lL's ln Lhe burn-down (charLs) LhaL we see lf we are desLroylng backlog or fllllng backlog. So every day Lhere ls conLrol and execuLlon, we plan ln Lhe beglnnlng of each sprlnL, every day we execuLe, every day we have meeLlngs Lo mlLlgaLe rlsks: obsLacles, whaL were we worklng on yesLerday, and whaL we are worklng on Loday. Lvery day we have conLrol, because as we say whaL we have already spenL ln an acLlvlLy, we are also saylng whaLs sLlll mlsslng Lhese lndlcaLors are evldenced. So execuLlon and conLrol are hand ln hand ln each sprlnL, wlLh Lhls aL Lhe end of each sprlnL, Lhe producL owner ls able Lo communlcaLe Lo Lhe cllenL, my veloclLy on Lhls sprlnL was such, so lL's predlcLable LhaL Lhe pro[ecL lnsLead of Lwo weeks ls golng Lo Lake Lhree. lmaglne LhaL you sald sprlnL one - Len days, Lwo people wlLh weekly sprlnLs are Len days per person, Lhls ls whaL Lhe Leam commlLLed Lo do, buL Lhen durlng execuLlon Lhe Leam could only do elghL days, Lhls permlLs Lo do Lhls: lf l could only make elghL, whaL am l golng Lo do? 1hen l assume l can only do elghL ln each. and Lhen l know, ln Lhe end, l need elghL more. 0,: An addlLlonal sprlnL. //: llve sprlnLs and l'm golng Lo communlcaLe wlLh Lhe cllenL, based on Lhe lnformaLlon people reporLed whaL Lhey could do, whaL Lhey weren'L able Lo do, oLhers, LhaL weren'L on Lhe scope of LhaL sprlnL, lnlLlally l commlLLed Lo Lhls, buL only Lhls ls done 0,: And as lL's communlcaLed, lL's renegoLlaLed. //: As lL's communlcaLed and Lhen all abouL managlng Lhe expecLaLlons. 1he cllenL can say no, lL's nonnegoLlable, so we need Lo Lell Lhe Leam Lo make up for Lhe losL Llme, Lhe cllenL ls plssed, annoyed. 1he ldea ls Lhls, Lhe producL owner dlscusses wlLh Lhe cllenL and Lhere's no problem, we go along aL Lhe same pace, buL Lhe Leam knows, we had planned Lo do ln 10+10+10+10, buL we are loslng. nexL Lhe Leam, lmaglnlng Lhe cllenL agrees, and Lhe Leam ls commlLLed Lo elghL and does Lwelve, Lhe average ls Len so ln Lhe nexL sprlnL we say Len and we are golng Lo be on Llme agaln. lL can be done on a dally basls. 1he followlng day, someone ls slck, we can only do four
71 whaL's Lhe average on Lhls? 24/3 - elghL agaln. Always wlLh Lhls, we look aL Lhe veloclLy charL and say all we can accompllsh we spllL by Lhe number of sprlnLs we can obLaln an average veloclLy, and aL each sprlnL closure we can say lf Lhe veloclLy was such, Lhen lL Lakes x Llme. lmaglne Lhere ls a sprlnL when noLhlng ls done, we sLarLed Lhe sprlnL buL Lhere was no LracLlon lL geLs ero, when Lhe sprlnL ends and we are golng Lo be penallzed because lf we haven'L done anyLhlng we had commlLLed Lo do, Lhen lL ls golng Lo Lake a much longer Llme. 1hen Lhe ldea around Lhls conLrol ls Lo glve us lndlcLors we can look aL and Lell cllenL lf lL's golng Lo be delayed or noL. 1he producL owner needs Lo undersLand how long wlll lL Lake and negoLlaLe wlLh Lhe Leam and lf Lhere ls no margln renegoLlaLe wlLh Lhe cllenL, glve hlm an esLlmaLe on how long lL wlll Lake. LxecuLlon and conLrol (repeaLs 3 Llmes). AL Lhe end of Lhe pro[ecL, lL's Lhe closure, Lhe ceremonles of closure: deploys Lo producLlon, closure of all, flnlshlng up Lhe documenLaLlon and dellvery of lL, run Lhe check llsL Lo see lf all we sald we'd do ls done, have a flnal meeLlng wlLh Lhe cllenL's Leam or send an e-mall lf Lhe Lype of pro[ecL ls such, walL for feedback, make a cosLumer's survey and Lhen lL's closed. AnoLher Lhlng l forgoL was, whlle managlng Lhls, lnslde Lhese appllcaLlons we can also do resource plannlng. We'll manage Lo have Leams whlch are sLable durlng a Llme perlod and we can say Lhls Leam ls dolng Lwo, Lhree or four or one pro[ecL, dependlng on Lhe dlmenslon, ln Lhe near fuLure and so we can plan Lhls Leam, Lhls week ls dolng Lhls pro[ecL, Lhe nexL week, Lhey'll be dolng anoLher, Lhen Lwo weeks, Lhls, Lhe followlng Lwo, LhaL. So you can plan, and based on LhaL plannlng, undersLand LhaL lf more pro[ecLs are avallable buL Lhere's no one Lo hand Lhem Lo, you need Lo speak Lo human resources. 0,: 1hey need Lo geL more conLracLs. //: 1here are more pro[ecLs, we are mlsslng resources, Lhe Leams are smaller, aL Lhe speed Lhey perform and aL Lhe speed Lhe pro[ecLs are comlng ln, Lhey are golng Lo have pro[ecLs on Lhe backlog, Lhey are only golng Lo be able Lo deal wlLh Lhree monLhs from now, because Lhey need Lo close everyLhlng aL hand. 0,: 1hen lL's also dlfflculL Lo manage LhaL wlLh Lhe cllenL. //: 1he pro[ecL comes ln, Lhere ls a flrsL meeLlng Lo presenL Lhe Leam responslble for Lhe pro[ecL, Lo collecL and have an analyLlcal sLage, because Lhen Lhe Leam can conLlnue Lo develop, lf Lhere ls someone responslble Lo do Lhls flrsL parL of analysls, lL manages Lo work, because Lhls sLage needs Lo be done Lo bulld Lhe backlog Lo glve Lo Lhe Leams and work on lL Lo break lL down lnLo sprlnLs, buL lL can sLarL golng forward wlLh someLhlng, maybe flfLeen days or a monLh, buL noL Lhree monLhs and Lhen lL
72 needs Lo Lell Lhe cllenL LhaL, based on Lhe plannlng of our resources, we wlll only be able Lo sLarL Lhls pro[ecL Lhree monLhs from now. We are very sorry buL. lf we have deadllnes, we someLlmes need Lo alLer prlorlLlzlng. We have Lhe order of pro[ecLs, we can delay a pro[ecL, renegoLlaLe wlLh LhaL oLher cllenL, Lo flL Lhls one ln Lhe mlddle. 0,: 1here ls all LhaL managemenL. //: up Llll now we have lncluded commerclal deparLmenL, flnanclal, human resources and . 0,: Lveryone lnvolved. now, for anoLher parL: Lhe dlfflculLles encounLered ln Lhe pro[ecLs, ls Lhere any ldea of Lhe naLure of Lhe problems. Are Lhey all of Lhe same Lype? //: 1here are several and from all sources. lor lnsLance, Lhere are problems relaLed Lo Lhe cllenL, for example Lhe cllenL does noL glve lnformaLlon, or whlch ls compllcaLed, whose ob[ecLlve ls noL Lo conLrlbuLe Lo Lhe success of Lhe pro[ecL, lnslde Lhe companles we are exLraneous Lo Lhls, buL have Lo deal wlLh lL, relaLlonshlp problems beLween cllenL and company, problems abouL lnLernal lnefflclencles: noL belng able Lo have esLlmaLed and ready backlogs, poor analysls and esLlmaLes, loss of Leam (leavlng Lhe company, becomlng lll, produclng below of Lhe expecLed, because Lhey're noL moLlvaLed, Lhere ls people and moLlvaLlon managemenL. 0,: 1hls people managemenL ls always done by Lhe ro[ecL manager or ls Lhere any oLher way? //: noL necessarlly, who worrles abouL Lhe Leam wellbelng ls Lhe scrum masLer, Lhe role of Lhe scrum masLer, hls ob[ecLlve ls ln flrsL place guaranLee LhaL Lhe Leam has sufflclenL lnformaLlon and all lL needs Lo develop Lhelr acLlvlLles, lf l have a problem, whlchever lL may be lL ls Lo be reporLed Lo Lhe scrum masLer. 8ecause whaL Lhe scrum masLer wanLs ls . l feel you can glve some more, whaL ls mlsslng? l'm mlsslng a machlne, lL's Loo old lL's delaylng me. Ck, LhaL's a problem, we can make lL beLLer, leL me see whaL l can do. 1he Scrum MasLer ls Lhere Lo help Lhe Leam Lo overcome any problem LhaL may arlse, Lo guaranLee Lhe Leam dellvers, ln Lhe maxlmum performance and Luned ln. 1hus keeplng levels of speed sLable (noL lrregular) and hlgh. 8eLLer Lhan hlgh, ln a flrsL sLage, sLable. redlcLablllLy, ln order Lo be able Lo know we always keep Lhls veloclLy, l llke Lo know ChrlsLmas comes on uecember 23 Lh , noL be ln doubL when ChrlsLmas ls, Lhls Llme. 1haL occurs wlLh movable holydays, when ls LasLer? We never know. And Lhls Lype of predlcLablllLy ls good. ou can counL on LhaL. eople llke Lo know whaL Lhey can
73 counL on, aL Lhe end of Lhe monLh recelvlng a pay, someLlmes noL aL Lhe end of Lhe monLh, aL Lhe mlddle, someLlmes Lwo monLhs afLer. redlcLablllLy ls only LhaL, lf we are comforLable wlLh lL. 1hen lL can be more veloclLy, leL's work on LhaL, as Lhe Leams become more experlenced, we are golng Lo have componenLs we can reuse ln oLher pro[ecLs, our speed lncreases. CommunlcaLlon problems, noL knowlng LhaL Lhere ls someLhlng golng wrong wlLh a pro[ecL unLll lL blows on our hands. So lndlcaLors, lnformaLlon on dlfferences of delLas on a dally basls, whlle people [usLlfy Llme losses and l'm sLlll mlsslng Lhls and lL's more Lhan whaL l LhoughL lL was. lf Lhls doesn'L flux lnformaLlon LhaL Lells you, aLLenLlon we have a problem, Lhen you can'L manage efflclenLly, because you don'L know, you Lhlnk everyLhlng ls alrlghL. 1here ls a need Lo geL people responslble Lo concenLraLe all Lhls and unblock, Lhrough communlcaLlon, Lhe facL LhaL S (a Leam member) has enLered Lo perform Lhls conLrol and Lo be scrum masLer aL Lhe same Llme ls Lo Lry Lo undersLand whaL Lhe Leam needs, whaL Lhey wanL or need ln Lhe pro[ecL or oLher Lhlngs, Lhen l'm golng Lo pass LhaL on, Lo Lhe proper responslble, aL Lhls Llme he may noL be able Lo be on LhaL meeLlng buL l am Lhe keeper of LhaL lnformaLlon, and LhaL's my funcLlon. And lf we can solve lL, very well, we are all glad, everyLhlng ls flowlng. CuaranLee speed. CuaranLee Lhe conLrol and LhaL Lhls lnformaLlon ls belng properly recorded, because Lhls lnformaLlon glves us Lhese lndlcaLors and Lhey allow us Lo acL. lL's her LhaL needs Lo Lell us, we have Lo blll Lhls, we need Lhls or LhaL, we need more resources, because of Lhe pro[ecLs LhaL are enLerlng and we have Lhe Leam compleLely allocaLed for Lhe followlng Lwo monLhs. ConLrolllng. 0,: all Lhe varlables //: lL's noL only on Lhe pro[ecL, buL globally all Lhe acLlvlLy of Lhe company. 0,: And speclflcally on Lhe esLlmaLes, ls Lhere any analysls Lo Lhe sprlnL plannlng and revlews over Lhe esLlmaLed versus Lhe real Llme spenL? //: Lvery day. lL's Lhe conLrol, monlLorlng and conLrol. Lvery day we can undersLand lf Lhere are delLas comparlng Lhe Llme we LhoughL would Lake and Lhe Llme LhaL we are execuLlng and whaL's mlsslng. 1here ls a Lable here aL !l8A. 0,: ln lan Lab? //: es, l can'L Lell you Lhe name exacLly buL Lhls reporL ls really nlce, because lL lnforms you LhaL Lhe sprlnL lnlLlaLed wlLh all Lhese lLems on Lhe backlog and Lhese Llmes lncluded, remalnlng esLlmaLe, Lhls one sLarLs baldly because Lhere are many aL zero, all should have a value, and Lhen every day as long as Lhere ls a.
74 for lnsLance on Lhe followlng day of Lhe sprlnL sLarL, Lhe sprlnL sLarLed aL 12.48 (lL was lLs creaLlon), aL 3pm Lhere was also an lLem belng added, lL doesn'L have Llme assoclaLed, buL lL should have, because lf Lhere ls Lhls lLem sLlll lncluded, whaL's mlsslng, and LhaL would make a dlfference, a delLa here aL Lhe esLlmaLed Llme. 0,: 1hen you can conflgure Lhe sendlng of warnlngs? //: 1he ob[ecLlve ls Lo look here and be able Lo undersLand by Lhe charLs. Lhe pro[ecL owner looks aL Lhls and undersLands we are worklng and Lhe backlog ls belng desLroyed, lf Lhe backlog lncreases, he says damn! We needed a day Lo undersLand Lhere was sLlll a loL mlsslng. 1hen lL wasn'L a good plannlng. 0,: 1haL one dropped rlghL afLer LhaL. //: 1hen he sald false alarm, buL anyways, lL ls lnformaLlon LhaL needs Lo be properly conLrolled. Scrum ls exLremely demandlng on LhaL polnL. Lvery day you say l spenL Lhls, Lhls ls sLlll remalnlng, every day you geL lnformaLlon, every mlnuLe you log on Llme and saylng whaL's mlsslng you're exerLlng conLrol. 8ecause you're permanenLly saylng, yes, we are beLLer now because Lhe Llme spenL advanced Lhe pro[ecL or no l've spenL Lhls Llme buL Lhere are sLlll so much more Lo do. Lvery Llme you log on Llme, you do Lhls managemenL. lL's very demandlng. 0,: 8uL lL's noL evaluaLed lf Lhere ls a cerLaln Lype of acLlvlLy whlch ls usually underesLlmaLed. //: noL aL Lhls Llme, buL Lhese are meLrlcs LhaL may be lnferred ln Lhe fuLure. 8ecause Lhen we can make componenL analysls and conclude, normally Lhese componenLs lL sLreLches, we esLlmaLed such and lL slldes. AfLer Lhe pro[ecL, we can ask: whaL were Lhe Lhlngs LhaL essenLlally slld? lL was Lhls user sLory, Lhls and Lhls. 1hen we look and ask lf Lhls user sLory ls always Lhe same, why dld lL sllde? A new componenL, we had never done anyLhlng llke Lhls, and Lhere was a splke Lo lnvesLlgaLe, we grasped someLhlng LhaL we had never undersLood, we learnL and we hope ln Lhe fuLure we can do Lhls. SLarL Lo know, even whlch person, because each one esLlmaLes ln a dlfferenL way, undersLands Lhe Llme he Lhlnks he's golng Lo Lake, buL only afLer a whlle does Lhe person sLarL Lo undersLand exacLly how long he Lakes Lo do Lhlngs normally, and sLarLs belng much more asserLlve. lor example, Lwo days Lo do Lhls, Lomorrow you are golng Lo do Lhe same and say Lwo days Lo do lL, and ln facL you only Lake one day afLer some 4 or 3 lLeraLlons you can say one day Lo do Lhls. 8ecause you have analyzed and undersLood, LhaL lL always Lakes a day. As lL becomes all Lhe same, you can be much more asserLlve on your esLlmaLes, Lhls ls broughL by experlence, Llme and Lhe way Lhe Leam works, Lhe worklng process of Lhe Leam,
73 several Lhlngs. We can ldenLlfy LhaL people may need help or Lralnlng, because Lhey're worklng on componenLs Lhey shouldn'L, because Lhey don'L have enough experlence on LhaL Lechnology, we could have oLher people worklng on Lhose componenLs because Lhey are more able, and we can schedule LhaL componenLs for Lwo weeks from now, when we have Lhe mosL experL person Lo do Lhls avallable and lf lL ls done aL Lhls Llme, lL ls golng Lo be done by someone who wlll Lake a loL longer and lL lsn'L worLh lL. And Lhen you can negoLlaLe wlLh Lhe cllenL saylng, our experL ls only avallable Lwo weeks from now, so we're dolng Lhls ln Lwo weeks, maybe lL's preferable Lo leL Lhe pro[ecL sllde compleLely. 0,: And whaL abouL crlLlcal facLors Lo Lhe pro[ecL's success? normally when lL sLarLs, are Lhey deflned? //: es, LhaL's one of Lhe Lhlngs LhaL should be done over Lhe flrsL assessmenL, we have already found a company LhaL does LhaL. When we wln a pro[ecL, over Lhe klck off meeLlng, glve a quesLlonnalre Lo Lhe cllenL Lo flll ouL ln Lhe beglnnlng of Lhe pro[ecL, wlLh quesLlons such as: WhaL are Lhe crlLerla you conslder more valuable on Lhls pro[ecL? Llke: Lhe esLlmaLed Llme, Lhe quallLy and robusLness of Lhe flnal producL, flnlshlng Lhe pro[ecL wlLhln Lhe predlcLable cosL. Several lndlcaLors! 1he cllenL manages lLs own expecLaLlons. Pe says l value Lhls and LhaL" and Lhen you can communlcaLe Lo all Lhe Leam, Lhe C Lo Lhe Leam and Lo Lhe Scrum MasLer, all undersLand whaL's lmporLanL Lo Lhe cllenL. Cn Lhe oLher day we had a meeLlng wlLh lrederlck, and Correla was saylng 'l'm worrled abouL Lhls, because we can do Lhls, ln Lhls way buL we are golng Lo have problems wlLh performance, lf we do lL llke LhaL we are golng Lo Lake Lhe longesL Llme, Lhe problem of synchronlzlng daLa bases lL's golng Lo Lake a very long Llme Lo do, on Lhe meeLlng we Lalked Lo Lhe cllenL whlch was lrederlck and explalned we can have performance problems here, how many cllenLs are Lhere? And oLher quesLlons. Lhls ls golng Lo hlL Sugar(C8M) a loL and we can have problems. lrederlck answered no problem. lf LhaL happens: more machlnes, more processlng capablllLy, l don'L Lhlnk lL ls golng Lo happen. lor me LhaL's noL a currenL problem." We deflned Lhe expecLaLlon he sald l'm noL worrled abouL Lhe performance, leL's do lL LhaL way." 0,: 1haL lnformaLlon ls lnserLed, lL's. //: More Lhan lnserLed, LhaL lnformaLlon was communlcaLed. Scrum doesn'L predlcL we documenL everyLhlng, obvlously lL may be reglsLered, send an e-mall wlLh LhaL lnformaLlon, LhaL's enough and l save LhaL e-mall aLLached somewhere ln Lhe lssue or ln Lhe descrlpLlon of Lhe lssue lLself lL says performance ls noL a lssue" Lhere lL musL be deflned LhaL over Lhe accepLance LesLs a performance ls noL consldered. 8uL
76 Lhere ls noL a formal way, Lhe maln aspecL ls communlcaLlon, Lhe reglsLerlng of Lhe evldence ls a defaulL. 0,: Ck, parameLers. 1he maln success facLors lsnL sLlll very developed, ls lL? 1he crlLlcal facLors? //: no, LhaL's someLhlng we are golng Lo develop, all Lhese sysLems you are Lalklng abouL are golng Lo be deepened, wlLh Ana aula, and Lhey are golng Lo be clearly deflned and documenLed. WhaL for? ln order Lo have a process where clearly we know we had a commerclal proposlLlonal LhaL was closed orlglnaLlng a pro[ecL, Lhe flrsL Lhlng we are golng Lo do ls undersLand whaL Lype of pro[ecL Lhls ls, consulLlng Lhe maLrlx, leL's creaLe Lhe pro[ecL, whlle dolng so we deflne Lhe bllllng lndexes, we creaLe Lhe pro[ecL on Lhls slde: flrsL sLage Lhe producL owner needs Lo do Lhls, second sLage he needs Lo do LhaL and lf Lhere are any dlfferences he needs Lo communlcaLe wlLh Lhe cllenL. We clearly undersLand every sLage and all Lhe acLlvlLles, Lhe people who need Lo do whaL, when Lhere are dlfferences, whlch ls perfecLly normal, whaL should be done, ln such a way we can manage Lhe whole Lhlng wlLh all lLs varlables, as we seen. 0,: 1hey are many. now you don'L have a way Lo ascerLaln lf Lhe pro[ecL was successful or noL. 1here aren'L meLrlcs. //: no, LhaL's a Lhlng LhaL dolng Lhe cosLumer survey ln Lhe beglnnlng and ln Lhe end we can sLarL Lo assess LhaL. We wanL Lo puL Lhese meLrlcs Lo Lhe cosLumer survey and send lL Lo Lhe cllenL uld you llke Lhe pro[ecL? es, no. WhaL dld you llke Lhe mosL abouL Lhe pro[ecL? Whlch are Lhe sLrongesL feaLures among all Lhe componenLs of Lhe pro[ecL? 1he people, Lhe Leam, Lhe Lechnology, Lhls, LhaL, whaL dld you llke Lhe mosL? 1he Llme frame, Lhe producL's quallLy?" We can evaluaLe a global perspecLlve ln llne wlLh Lhe lnlLlal expecLaLlon survey, ln order Lo undersLand lf Lhe expecLaLlons were fulfllled or noL. ln meLrlcs vlews we can say we have 100 pro[ecLs a year and from Lhose 80 have an accepLance level from Lhe cllenL's vlew as good, Lhe relaLlonshlp wlLh Lhe Leam ls specLacular, Lhen we can dlvulge Lhese lndlcaLors and evaluaLe more Lhan x of our cllenLs llke our Leam, more Lhan x are glad wlLh our work, we can manage Lo glve LhaL feedback Lo Lhe markeL, LhaL's whaL brlngs even more cllenLs. 0,: 1o Lhe markeL and lnLernally. //: lnLernally, Lhe Leam can undersLand Lhe work. Lhe recognlLlon Lhe cllenLs are havlng of Lhe work Lhey do, exLernally Lhe cllenLs, or oLher poLenLlal cllenLs wlll say 'LeLs work wlLh Lhls company because Lhey are good'. 0,: 1hese crlLerla, Lhese facLors aren'L deflned yeL. //: 1hey are golng Lo be all deflned.
77 0,: because Lhere was no Llme, ls lL baslcally a problem of Llme? //: es, lL ls golng Lo happen now. 0,: soon //: soon. 1hese sysLems LhaL we have been Lalklng abouL are all llnked. 1he way we are changlng Lhls ls noL leL's [usL sLarL here, because Lhere are some Lhemes we need Lo undersLand Loday and overcome Loday because Lhey'll brlng more producLlvlLy. now we are golng Lo focus on plannlng, here and here. We wanL Lo have esLlmaLed backlogs, compleLely esLlmaLed and LhaL Lhls value maLches Lhls value, whaL we sell Lo Lhe cllenL ls Lhe same as Lhe backlog, lf Lhere ls any dlfference leL's lmmedlaLely speak Lo Lhe cllenL, when we come here and we slL Lo plan Lhe sprlnLs we wanL Lo undersLand, are Lhere any dlfferences, lf so leL's speak Lo Lhe cllenL, and Lhe Leam needs Lo plck Lhls and work on a known Lhlng, and every day Lhey reglsLer Llme spenL / remalnlng. 1hls ls our maln worry. lf Lhls ls ln gear, Lhen yes. 1hen how pro[ecLs are broughL Lo us? 1hey are reglsLered here. Lxample: leL's add flelds Lo a Sugar(C8M). 1hls fleld ls feed by a servlce, Lhls fleld ls feed by anoLher and so on. 1hen aL a cerLaln Llme we are able Lo say: we have Lhe pro[ecL here belng feed by Lhese lndlcaLors, based on Llme remalnlng, spenL, and Lhe delLas of Lhe dlfferences from one week Lo Lhe oLher" and Lhen we are managlng Lo Lackle Lhe maln problem whlch ls Lhe Leams are produclng, lf we worry abouL Lhls and Lhe Leams are here aL Lhe place, we don'L have meLrlcs, we have noLhlng, we don'L know how Lhlngs occurred. 1hen Lhe quesLlonnalre we glve Lhe cllenL ln Lhe beglnnlng and Lhe one ln Lhe end Lhey are golng Lo be Lhe lasL Lhlngs Lo do, Lhey are noL Lhe flrsL. We baslcally already know Lhe cllenL llkes Lo geL Lhlngs on Llme, he llkes quallLy and Lhe budgeLs are meL, Lhese premlses we know LhaL Lhe cllenL always wanLs Lhem. Are Lhere a loL Lo evaluaLe we don'L evaluaLe now? es. We wanL Lo evaluaLe Lhem ln Lhe fuLure, buL Lhe prlorlLy ls Lo sLeer Lhe boaL. 0,: AlrlghL. now, abouL Lhe sysLem we are uslng !l8A and Lhe oLher componenLs, we don'L have a name. //: LM - LnLerprlse ro[ecL ManagemenL 0,: es, l've [usL seen LhaL name. Are Lhe LesLs belng successful? uo you Lhlnk so? ls lL a beL for Lhe fuLure? //: l belleve so. 1he nordlcs love lL, Lhey have glven us feedback, Lhey are lovlng lL. 1he sysLem ls lncredlbly powerful, 8lcardo and l are well lnformed abouL lLs ablllLles, we sLlll haven'L found Lhe Llme Lo lnform Lhe people we are sLarLlng Lo do Lhlngs Lhls way. 0,: ls Lhere a way?
78 //: no. 0,: lor example, yesLerday we were asklng where Lo load Llme and, Lhls opLlon or LhaL. //: Pe Lold me Lhls mornlng, and lL ls golng Lo be sLlpulaLed, Lhere'll be documenLaLlon on how Lo deposlL Llme, afLer sLlpulaLlng LhaL, whlch ls someLhlng LhaL ls lmporLanL for us, parL of Lhe dally acLlvlLy of everyone ls Lhe Llme reglsLraLlon and LhaL plan needs Lo be rapldly clarlfled so LhaL people sLarL reglsLerlng Lhe Llme on Lhe correcL places. And Lhe way we are dolng lL ls. lf we were golng Lo deflne Lhe whole process Lo load on Lo sysLem lL would Lake slx monLhs, and Lhen lL ls preferable Lo load Lhls, lL happened ln Lhe same way wlLh acLlvecollab unLll we were on gear, you need Lo glve Lralnlng Lo people. 1here was LhaL whole process buL we were already uslng lL, we have already senL lL forward. l Lhlnk LhaL ls an aglle corporaLlon, a company LhaL puLs lLself Lhrough challenges. LeL's sLarL dolng Lhls Lhls way, lL Lrembles a blL, we are a blL losL, buL slx monLhs from now, we are golng Lo be beLLer and one year from now, even beLLer, Lwo years, beLLer even. 1he perspecLlve needs Lo be Lhls 'conLlnuum lmprovemenL, whaL are we golng Lo aLLack now? WhaL are Lhe maln problems?' 1hen one of our maln ob[ecLlves we are Lrylng Lo obLaln ls backlogs, all pro[ecLs havlng an esLlmaLed backlog, even lf lL on hlgh level: Len, Len, Len, flfLeen, LwenLy, flfLeen, LwenLy. 1here! lL's esLlmaLed. uoes lL maLch wlLh Lhe number of days on Lhe proposlLlon? no. 1hen re-scheme Lo maLch. 0,: 8uL lL can'L be from such a helghL.. //: 8uL lL ls. lL may be. 1here's no problem, you don'L have any commlLmenL. WhaL do you know Loday? lf you do flve, Len, flve, Len, flve, Len and lL maLches Lhe number of days, perfecL. 0,: 8uL Lhere ls a rlsk of Lrylng Lo maLch Lhe esLlmaLes wlLh Lhe proposlLlon. //: es, because l know, lf l sold flfLy days, lL's because l need Lo do Lhls ln flfLy days, now when we sLarL breaklng lL down, we evaluaLe lf lL can be done ln flfLy days. lf noL, we renegoLlaLe wlLh Lhe cllenL. 1he cllenL boughL lL for flfLy, Lo say flfLy we had some assumpLlons, Lhen we need Lo presenL all Lhe lfs. SomeLlmes Lhe cllenL says l need a sysLem LhaL does such and such." 1hls ls hls requesL, how do you know? ou answer, Lelllng hlm LhlrLy days, buL Lhen he wanLs LhaL and LhaL. ou answer: WhaL we had ln mlnd for LhlrLy days was such and such, noL all you are asklng now, whaL you are saylng ls dlfferenL, lf you had Lold me LhaL ln Lhe beglnnlng we would have proposed sevenLy or hundred or one Lhousand". 0,: uo you Lhlnk., well l assume Lhls pro[ecL ls always ongolng?
79 //: es, flrsL sLage: have a basellne. 0,: Are we already on Lhe flrsL sLage? 1he second sLage? WhaL's Lhe sLaLus? //: We are on . LhaL's a blL hard Lo answer. We are well on Lhe way, we already have all Lhe appllcaLlons lnsLalled and are uslng Lhem, l could explaln all Lhls, because Lhey are Lhlngs we are already dolng. lL ls noL Luned, we are aL a sLage of LesLlng and Lunlng. 0,: dlscovery? //: 1haL l Lhlng ls always Lhere, Lhe same way we are worklng. Lvery Llme you do a plannlng for a seL of funcLlonallLles ln a sprlnL you are dolng dlscovery. When you are golng Lo break lL down. ou have a backlog and lL ls esLlmaLed. es! l have a backlog and l know Lhls ls whaL l wanL Lo do, dld l go Lo every deLall and reflned lL? no. 1here are some parLs where l'm more advanced and Lhere are oLhers where l'm more dlsLanL, buL l know whaL l wanL, and how much Llme do l have Lo do lL all. 1haL's why l can'L Lell you lf we are on a flrsL, second or Lhlrd sLage, because lL ls lLeraLlve. Pow many passages have we done? Several, we sLarLed by lnsLalllng all Lhe appllcaLlons, we sLarLed revlewlng Lhe meLhodologles we are golng Lo use, we've already done a prevlous sLudy on pro[ecL managemenL meLhodologles and sofLware developmenL and now we need Lo Lune lL. We need Lo ask: ln whaL way does Lhls adapL Lo our currenL model?" Assumlng LhaL our model lsn'L Lhe correcL one. 1here ls sLlll a loL Lo lmprove. 0,: Ck. We have spoken abouL Lhe Lool, ls Lhere anyLhlng Lhe Lool doesn'L. //: 1he Lool needs Lo manage requlremenLs, backlog and background of Lhe Lasks. 1he Lool needs Lo allow communlcaLlon and by reglsLerlng Lhlngs you are already communlcaLlng and by deflnlng esLlmaLes and Llmes (by Lhe Leam and C), you are already communlcaLlng expecLaLlons, you have Lhe Confluence LhaL allows you Lo descrlbe and documenL e everyLhlng, you have Lhe posslblllLy Lo make pro[ecL's documenLaLlon and analysls, glossarles lnslde Confluence, Lhe posslblllLy Lo make mock ups and dlagrams. 1he communlcaLlon plaLform ls sLlll mlsslng, we have preLLy much chosen one, buL we haven'L gone forward yeL. 8uL we are movlng on, because for Lhe revlews lLem, for Lhe sprlnL revlews lL ls very lmporLanL. As we do many remoLe revlews, for lnsLance for Lhe nordlcs, we are dolng a remoLe presenLaLlon we need Lo have a Webex where we can show, where we can record and recelve Lhose feedbacks afLer. 0,: ls LhaL someLhlng LhaL may be lnLegraLed? //: l don'L know yeL. We are Lrylng Lo explore ln whaL way lL can be lnLegraLed. 1he process ls lnLegraLed, l can say all sprlnL revlews done wlLh remoLe cllenLs, or wlLh
80 near-shorlng Leams are done Lhrough an appllcaLlon, l can do a presenLaLlon and recordlng so l can plck Lhe feedbacks and feed Lhe backlog based on Lhem. 1he LesLlng Lool: 1wo - Zephyr and 8onflre". Zephyr permlLs Lo creaLe LesL cases, LesL scrlpLs and LesL cycles Lo allow bulldlng: funcLlonal LesLs, a cycle LesL for cross browser for lL, llrefox Chrome, Lo have a LesL cycle of regresslon LesLs whlch are when we change someLhlng, aL leasL Lhese ones we need Lo LesL, a LesL cycle of unlLary LesLs of cllenL accepLance or user accepLance LesLs. 1hen 8onflre ad hoc LesLs you can surf Lhe appllcaLlon, Lhe producL you have [usL concelved, capLure screen shooLs and lmporL Lhem dlrecLly lnslde !l8A / CreenPopper. Make shooLs wlLh boxes and arrows and sLuff around. 0,: Saylng whaL? //: Saylng whaL Lhe problem ls, lL auLomaLlcally capLures operaLlng sysLem, browser verslon and a serles of Lhlngs.1he LesLlng parL ls covered. We covered requlremenLs, developmenLs, LesLs, documenLaLlon, dlagrams, mockups, communlcaLlon. l Lhlnk we have all Lhe Lools and condlLlons Lo do Lhe besL work posslble. l don'L see Lhls happenlng aL Lhe companles around us. l look aL Lhe companles around us, and see Lhem worklng ln Lhe model ln whlch u8l worked ln Lhe pasL, where Lhe lnformaLlon was exchanged orally, Lhere were no records and where people dldn'L have Lhe noLlon lf Lhey are performlng Lasks LhaL are really producLlve or belng. 0,: 1hey don'L know Lhe sLaLus quo. //: LxacLly, whaL we ulLlmaLely wanL ls a room full of screen across Lhe offlce where we have all Lhe lnformaLlon abouL our acLlvlLy, all we do on Lhese screens, lL's someLhlng you look whlle you pass, and lf you see a red screen you see LhaL pro[ecL lsn'L ok, someLhlng ls up, can l help? We are havlng a problem ln Lhe supporL area, how many LlckeLs arrlved Loday? Can l help? More pro[ecLs are arrlvlng, llke ln Lhe alrporL: Lhe plane A, 8, C are landlng. 0,: lL's lmporLanL Lo have Lhe noLlon of whaL's comlng.All LhaL aL '1orres do CrlenLe' rlghL? //: es. 0,: (LaughLer). //: noL Lhere, Lhere ls a llmlLaLlon whlch ls you can only have one person by each 10 square meLers, renLlng an offlce of 400 square meLers you could only have 40 people Lhere, LhaL's whaL we have here. 1here's a llmlLaLlon abouL emergency exlLs, so Lhere's Lhe llmlLaLlon LhaL each offlce only has one person for each 10 square meLers
81 and lL's a bruLal cosL. lf anyone here wlns Lhe loLLo and wanLs Lo renL Lhe offlces Lhere l won'L mlnd. 0,: Ck. lL's all. //: Are you enllghLened? uo you Lhlnk my vlslon.? 0,: l already had an ldea, buL l wanLed Lo have Lhese Lhemes Lo dlscuss wlLh Lhe professor and see, based on my research, buL Lhese weeks l arrlve home a blL Llred. //: ou have Lhe ldea of whaL's happenlng and wlll happen. Ana came here Lo glve Lhese Lralnlngs and lasL week 8lcardo and l had a reunlon wlLh her Lo Lry Lo undersLand whaL she LhoughL abouL Lhe Leams and she sald she LhoughL Lhe Leams had a loL more knowledge Lhan whaL she has observed elsewhere, when she glves Scrum Lralnlng Lhe people don'L know, don'L undersLand and never heard of lL, she has arrlved aL places where Lhey say we do Scrum, buL noLhlng llke Lhls". 0,: (LaughLer) //: And she undersLood we have a predlsposlLlon, we are all opened up, one of Lhe maln dlfflculLles on lmplemenLlng Lhese sorLs of Lhlngs ls reslsLlng changlng, buL she undersLood we are really orlenLlng on lmprovlng and we all wanL someLhlng much more organlzed, more loglcal and predlcLlve. 1hen whaL we agreed wlLh her ls LhaL we wlll conLlnue Lo do Lhls Lralnlng process, buL on Lhe [ob, Lackllng Lhe crlLlcal polnLs. We Lold her we were golng Lo Lry Lo guaranLee we have backlogs prlorlLlzed and esLlmaLed and we wanL you on Lhe plannlng meeLlngs Lo observe how do Lhe guys break down Lhe problems and reglsLers Lhls lnformaLlon so LhaL you can promoLe lmprovemenL on how we do LhaL. We weren'L used Lo lL, Lo breaklng down someLlmes we Lhlnk buL we don'L wrlLe l need Lo do Lhls, and Lhls, and Lhls." And lL works because when you wrlLe everyLhlng, you only need Lo. 0,: ass on? CC: ou have seen Lhe documenLs you have belng maklng for supporL. 1o me lL's loglcal. lf l need Lo pass Lhe lnsLrucLlon of deploymenL Lo someone, l wrlLe all of Lhem, and Lell Lhem where Lhey are, you share. 0,: l even llked, l hadn'L dlscovered yeL, LhaL l could puL Lhe checkboxes. //: 1hls ls done and Lhls Loo and so ls Lhls. And l already have a plug-ln LhaL ls '1alk' where you can puL commenLs ln llne llke you do on google docs, for lnsLance l don'L qulLe know Lhls" or aLLenLlon Lo Lhls". ou can communlcaLe wlLh everyone and our real ob[ecLlve ls Lo flnd a paraphernalla of Lhlngs, noL Lo subsLlLuLe verbal communlcaLlon, LhaL ls always lmpllclL, buL Lo reglsLer and save evldences abouL Lhe Lhlngs we do. under LhaL a secreL pro[ecL ls runnlng LhaL ls Lhe collage of our process Lo
82 CMMl, Lo be compllanL wlLh maLurlLy level ll aL leasL and one day be able Lo propose ourselves Lo cerLlflcaLlon. Why ls lL secreLly runnlng? 8ecause lL has Lermlnology a blL dlfferenL from Scrum and we don'L need Lo know. ou do Lhe acLlvlLy x1C of whaLever whlch ls Lake a phoLo and puL lL lnslde Lhe sysLem. 1he value lL brlngs even lf we don'L know how lL ls called on CMMl lL's LoLally lrrelevanL. All we need Lo know ls LhaL we have a screenshoL and we puL Lhe phoLos of our deslgns and sLuff saved elLher on Lhe lssue on !l8A or lnslde Lhe Lree of Lhe pro[ecL and so on. l Lhlnk wlLh Lhls we are leaplng forward ln quallLy, l'm clearly undersLandlng LhaL slx monLhs from now over Lo a year we are golng Lo be on a compleLely dlfferenL reglsLer. ou are here longer and you have observed, how we were a few years back, how we have been progresslng and l Lhlnk Lhls whaL lL does for us ls Lo belleve, Lhls lsn'L belng done by one or Lwo, Lhls ls belng done because everybody wanLs Lhls and ls movlng Lhere. We have accompllshed Lo undersLand and belleve Lhls works. 1hen look, puL Lhlngs ongolng and conLlnue Lo flghL and walL. LssenLlally we need Lo walL, walL knowlng we are dolng someLhlng. !usL walLlng, knowlng LhaL noLhlng ls happenlng, LhaL we are dolng noLhlng, no, LhaL ls demoLlvaLlng. SomeLlmes lL's noL so fasL, we need a longer break, Lhen we go forward qulcker. 0,: 1he way lL's always been: one sLep aL a Llme. //: lL needs Lo be solld and susLalnable, oLherwlse. 0,: Ck. //: l hope l've been clear. 0,: erfecLly. //: l'm sorry abouL Lhe delay on Lhe conversaLlon, buL Lhese have been crazy weeks. 0,: Ck, leL me [usL. a<;<=< (LIEFOJEU = ! lnLervlewer: Mlguel erelra ! lnLervlewed: uora 8allo ! uaLe: lrlday, 29 of november of 2012
0,: 1he flrsL quesLlon ls: who glves you Lhe lnformaLlon you need for Lhe new pro[ecLs? lrom Lhe commerclal area, ls lL you. &7: lrom Claudlo, (laughLer) baslcally. 0,: lL doesn'L come from 8yder, ls lL always Cludlo?
83 &7: no, 90 comes from Cludlo, exlsLlng buslness or new buslness LhaL arrlved Lhrough hlm and Lhen he passes lL on Lo me. 1hen lf lL ls a pro[ecL l am already worklng on llke 1ecnldelLa, lL's me who does Lhe requlremenLs assessmenL wlLh Lhe cllenL, Lhen l valldaLe lL wlLh Lhe asslgned commerclal, ln Lhls case 8ebelo or 8yder, ulogo (8ebelo) on Lhe knowledge and 8yder on Lhe bureaucraLlc process of proposals. When lL's really new, normally lL's always Claudlo. l'm golng Lo Lell you abouL Lhe cases, 1ecnldelLa ls Lhe one l've Lold you abouL, verLl l was presenLed Lhe pro[ecL flrsL, Lhls case was Cludlo, SlAS was Cludlo, Lhe llfLy-Slx l don'L know lf lL was Claudlo or aulo MrLlres because Lhey already had prlor knowledge. 0,: es, yes. &7: 1hey presenL lL Lo me, pass on Lhe lnformaLlon. no from SlAS lL was !orge 1elxelra da Sllva who pass lL on. 1hey pass on Lo me all Lhe pre-analysls work Lhey had done and l valldaLe lL, ask for some more componenLs and lL sLarLs ln Lhls manner. 0,: And Lhen? &7: When lL's wlLh Lhe cllenL, lL's Lhe cllenL LhaL glves me LhaL lnformaLlon. 0,: ln Lhe ones you were already a parL of? &7: LxacLly. 0,: Whlle seLLlng up Lhe pro[ecL, do you always sLarL by asklng LhaL lnformaLlon Lo Claudlo? 8uL evenLually you conLacL Lhe cllenL and do your own analysls or do you base yourself ln Lhe one Lhey geL you? &7: 1here ls always doubL, Lhen you have Lo Lalk Lo Lhe oLher conLacL, lL may noL be Lhe cllenL dlrecLly, buL someone responslble. how ls lL called, lL's noL Lhe buslness. 0,: lrom Lhe cllenL's slde? &7: yes. 0,: SLakeholder. &7: SLakeholder, exacLly. lL may be someone exLernal and ln Lhe SlAS case lL's SanLana, Lhen he speaks Lo MouLlnho and Lhey see lnLernally wlLh Lhe cllenL Lnglneer whaL's ls name. So Lhere are a loL of people. l also valldaLe lL lnLernally wlLh people who already knew Lhe pro[ecL lnslde. 8uL Lhe lnlLlal documenLaLlon ls never on Lhe level of deLall we need, so we always need Lo pose more quesLlons, your own Leam asks you quesLlons and you need Lo geL. 0,: Lhe lnformaLlon, yes. And do Lhe pro[ecLs always follow Lhose sLeps? Cr ls each pro[ecL dlfferenL? &7: ln everyLhlng ln llfe, each pro[ecL ls dlfferenL. 1here lsn'L a pro[ecL 100 llke oLher.
84 0,: ou can never ldenLlfy some sLeps LhaL. &7: no, Lhe sLeps are global, buL Lhey always have dlfferenL conLours. 8ecause of Lhe people lnvolved, for lnsLance ln Lhe case of verLl Lhere was a flrsL requlremenLs assessmenL done by Cuerra, l Lhlnk, and Lhen l spoke Lo Lhe cllenL and dld anoLher requlremenLs assessmenL myself. 0,: ou dld LhaL requlremenLs assessmenL yes. 8uL Lhen how Cludlo sald Lhere should be Lhose phases: conLrol and execuLlon, noL LhaL much Lo you, LhaL's more for Lhe developmenL crew, Lhe conLrol, Lhe lnlLlaLlon and Lhe dellvery sLages. lnslde Lhose sLages l don'L know lf you have LhaL noLlon. lf you Lhlnk llke LhaL. &7: 8lghL now l was [usL focuslng on whaL happens ln Lhe beglnnlng. 0,: 8uL whaL abouL along Lhe whole pro[ecL? &7: 1he whole pro[ecL. 0,: Are Lhere any sLages you ldenLlfy? &7: l ldenLlfy, l ldenLlfy myself wlLh Lhls way we are worklng now, ln Lhe sense Lhere are several lLeraLlons and we are dolng LhaL on all he pro[ecLs now. As l dld ln oLher pro[ecLs, where Lhe meLhodology was Lo show ln each Lhree monLhs or rlghL aL Lhe end of Lhe developmenL lL lsn'L golng Lo work, noL Lo say bad words 0,: 1haL's ok, 1hen l &7: ou fllLer Lhe bad words. 1he followlng parL on how we are dolng lL know obllges LhaL, every Lwo weeks or aL Lhe maxlmum every Lhree weeks, you show someLhlng Lo Lhe cllenL. lL's a way of havlng a rouLlne. 0,: Also monlLorlng &7: MonlLorlng whaL ls belng developed. And LhaL, once agaln, ls dlfferenL from cllenL Lo cllenL, for lnsLance verLl ln some subpro[ecL we can do LhaL really every Lwo weeks, because we had Lhe demos and we passed Lhem onLo Lhelr envlronmenL. WlLh 1ecnldelLa Lhere ls every 4 or 6 monLhs LhaL we do small new feaLures. ln Lhe SlAS case, every Llme we develop someLhlng we show Lhem Lo valldaLe lL, [usL now l've recelved and e-mall glvlng me feedback on whaL l had senL Lhem, LhaL's why l Look a llLLle longer. 0,: ou do Lhe monlLorlng and aL Lhe end ls lL always Lhe same way? Pow do you. ou glve lL Lo Lhe cllenL and. &7: Cnce more, lL depends on Lhe cllenL, from pro[ecL Lo pro[ecL. l'll glve you all Lhe examples of Lhe pro[ecLs l've worked on. Cn verLl's case, we had an envlronmenL LhaL we made avallable for Lhem Lo LesL, we had dlfferenL people wlLh whom we spoke Lo, Lhe demos we made LhroughouL developmenL were wlLh key users no, Lhe maln
83 sLakeholder, he valldaLed everyLhlng, dld Lhe LesLs, eLc. when he accepLed, Lhen we had a formaLlon wlLh Lhe key users and Lhey dld LesLs and Lhen we enLered producLlon and flnlshed. 1hey lnLernally dld more valldaLlon and LesLs. ln 1ecnldelLa's case, for example we do our LesL here, we make avallable a LesL envlronmenL, Lhe cllenL (Cludla) LesLs lL, someLlmes she doesn'L glve Lhe correcL feedback and Lhen we puL lL on producLlon. 1hen who glves feedback are Lhe operaLors, who glve her feedback LhaL Cludla sends Lo us. ln SlAS's case, l sLlll haven'L puL anyLhlng ln producLlon, buL lL's Lhe sLakeholder. 0,: lL ls golng Lo be llke LhaL. &7: We Lry Lo have aL leasL one LesLlng envlronmenL / quallLy and Lhen Lhe flnal - producLlon. 0,: 8eyond Lhe communlcaLlon you perform wlLh Lhe developmenL Leam, besldes LhaL wlLh whaL oLher deparLmenLs do you communlcaLe durlng Lhe developmenL sLage, how do you do lL? &7: lnLernally, here. 0,: es. &7: 1here's only one Lhlng LhaL perlodlcally we are quesLloned abouL, Lhe flnanclal deparLmenL normally when a preposlLlon ls made for Lhe pro[ecLs LhaL are bllled aL Lhe sLarL, ln oLher words Lhey have flxed prlce Lhere ls a clause LhaL says when Lhe pro[ecL ls bllled, x ln Lhe beglnnlng, x ln x monLhs and x someLlme afLer and x ln Lhe end. When LhaL Llme comes, Lhe flnanclal deparLmenL asks lf Lhey can blll Lhe cllenL or noL. 1haL's whaL l speak Lo Lhem. 0,: ls lL done verbally or. &7: Mall and verbally boLh. WlLh Lhe commerclal deparLmenL always when Lhere ls a requlremenL LhaL changes Lhe condlLlons, for example Lhe oLher day Lhere was a problem wlLh 1ecnldelLa's daLabase, lL crashed and Lhe cllenL spoke Lo me and Lo Lhe Lechnlcal (supporL) deparLmenL and l also spoke Lo Lhe commerclal area Lo valldaLe LhaL LhaL was conLracLually esLabllshed and lncluded on Lhe servlce, so LhaL l wouldn'L offer Lhe cllenL a servlce LhaL Lhey weren'L paylng for. And Lhus l Lhlnk l spoke abouL all Lhe u8l's deparLmenLs: SysLems supporL, commerclal and flnanclal. MarkeLlng, Lhey speak more Lo me Lhan l Lo Lhem. noL Lhe one you were Lhlnklng abouL, buL for lnsLance l rarely asked Lhem someLhlng for Lhe pro[ecLs, Lhey ask us more for Lhelr pro[ecLs noL Lhe personal parL you oughL Lo be Lhlnklng abouL llke newsleLLer and whaL noL .
86 0,: es, buL lL wasn'L all. LhaL and for Lhem Lo publlclze Lhe pro[ecLs LhaL were done. &7: es, buL Lhey can also ask, as Lhey already dld, when Lhey wanL a case LesL approved by Lhe deparLmenL. 0,: now, maybe Lhe mosL dlfflculL quesLlon: WhaL are Lhe maln dlfflculLles you encounLer on managlng Loday? &7: 1he pro[ecLs? 0,: es. &7: 8e able Lo documenL, now wlLh Confluence lL ls a preclous help. 8uL lL was Lo be able Lo documenL convenlenLly Lhe pro[ecLs and l'll explaln whaL l mean wlLh convenlenLly documenL pro[ecLs. Why ls Lhls? l have someLhlng LhaL l need Lo descrlbe, someLhlng l wanL Lo alerL Lo, someLhlng l [usL wanL Lo polnL ouL, someLhlng Lo keep Lo show Lhe cllenL ln Lhe fuLure and each Llme l had Lhese needs l wroLe Lhem ln a doc or excel or google share doc someLhlng llke LhaL and 1wo Lhlngs: 1he flrsL: l'd lose Llme knowlng how Lo organlze my own lapLop, whaL klnd of lnformaLlon LhaL was, Lhey should be lnslde a folder called whaLever for me Lo be able Lo flnd lL laLer, Lhen l'd lose Llme over Lhe formaL, l need lL Lo be preLLy, Can l make lL lnLo a Lable, lf l come here afLer Lhree monLhs am l golng Lo recall how Lhls was wrlLLen? So, Lhere were always Lwo Lhorns ln Lhe slde LhaL prevenLed me Lo reach Lhe goal more rapldly whlch was Lo effecLlvely descrlbe whaL l wanLed Lo descrlbe. WlLh Confluence, LhaL helps a loL because l can do baslc formaLLlng a la wlkl, lL doesn'L maLLer lL ls always avallable Lo everyone, so l don'L need Lo worry abouL where lL ls, l can search for lL ln any way LhaL l wlll flnd lL and laLLer l can exporL lL Lo pdf or puL onLo a mall l can do whaLever l wanL. l also can reuse LhaL lnformaLlon Lo everyLhlng l wanL. Cne of Lhe greaLer dlfflculLles LhaL l had slnce my oLher company and now here ln Lhe beglnnlng was, Lhe lnlLlal process ln every pro[ecL ls always a meeLlng, mlnuLes of Lhe meeLlng, Lhen a drafL of Lhe proposal, Lhen Lhe effecLlve proposal and Lhem Lhe requlremenLs deLall and Lhls ls all lnformaLlon LhaL before was only on docs, Lhen Lhe requlremenL was noL LhaL one, lL had Lhe number of Lhe oLher requlremenL and Lhem l had Lo wrlLe over many Llmes Lhe same Lhlng, someLlmes Lhe LexL wasn'L exacLly Lhe same Lhlng, when l only wanLed Lo add deLall or read[usL. 1haL goL ln Lhe way, because lL was Llme consumlng and lL perverLed Lhe Llme l had Lo profound Lhe requlremenL deLall ln each deal, now lL ls much beLLer. l can reuse everyLhlng l wrlLe once wrlLLen lL may be llnked Lo varlous places, llnked Confluence Lo !l8A and lL has been bruLal. l
87 make beLLer use of my Llme and Lhe documenLaLlon ls always avallable Lo everyone. l'm speaklng mosL abouL Lhe Lools because Lhey have been Lhe greaLesL help. 0,: 8uL your problem was documenLaLlon. &7: ln Lhe sense LhaL, l'd do lL, spend Llme dolng lL and Lhen whaL were Lhe galns l had for dolng lL? l could geL Lhem, buL lL Look Loo long Lo geL Lhem, because lL was Loo scaLLered. 1hey were Lhe maln problems. Whlle you were speaklng Lo someone, whaL was Lhe verslon of Lhe documenL? Was lL x or ? l can'L open lL, because l don'L have docx, lf lL was ln excel. send me LhaL ln pdf, now l can'L see Lhe Lable. send me LhaL on an e-mall. A loL of Llme ln crap, ln LroubleshooLlng lnformaLlon access and llLLle Llme Lo analyze whaL you really wanL. AL any Llme you can share, and Lhe facL LhaL you share doesn'L dupllcaLe Lhe LexL. 0,: 1he lnformaLlon yes. &7: When you do lL by e-mall, dlrecL e-mall, you wroLe lL once and Lhen you hlL reply and Lhe people ls already on anoLher e-mall, flne lL's all connecLed buL you have LhaL wrlLLen a loL of Llmes on Lhe e-mall. When you share your LexL ls only Lhere, you only polnL ouL people Lo access and commenL LhaL lnformaLlon. 1he lnformaLlon doesn'L geL losL. 0,: 1haL was your maln problem, looslng and lnformaLlon managemenL. &7: and Lhe Llme losL ln all LhaL. 0,: ou've already answered Lo Lwo quesLlons ln one. &7: CreaL 0,: ou are on half Lhe Llme now. &7: CreaL (long slgh) 0,: Are Lhe esLlmaLes made for Lhe developmenL currenLly compared Lo whaL was done? &7: wlLh whaL was spenL by Lhe end of Lhe pro[ecL. 0,: es. &7: Ah. l sLlll dldn'L have Lhe opporLunlLy Lo do lL, slnce we are uslng Lhe Lool, on Lhe oLher pro[ecLs l had never done LhaL Lype of analysls, or l end up dolng lL when l spenL more Llme, because lL sLands ouL, Lhe quesLlon was [usL abouL analysls, ls LhaL lL? 0,: es, analysls. lf you do lL or noL. &7: 1here are people who do lL for me necessarlly, whlle dolng Lhe bllls Lhey see how much Llme l spenL. And now lL ls much easler Lo be on Lop of LhaL sorL of maLLer, because Lhlngs sLand ouL much more clearly. 0,: So before you dldn'L.
88 &7: 8efore l dldn'L do lL. 8y my own lnlLlaLlve, noL dlrecLly, Lhe resulLs appeared Lo me. 0,: And now wlLh Lhe Lools avallable, do you Lhlnk you wlll be able? &7: no. 0,: 1haL belng, you cannoL have any ldea on whaL Lhe acLlvlLles LhaL are underesLlmaLed. &7: l have an ldea, because erroneously or mlsLakenly, when you do Lhe flrsL proposal and when a commerclal Lhlnks abouL someLhlng Lo glve Lo Lhe cllenL, Lhere are several Lhlngs he doesn'L Lake lnLo conslderaLlon. More and more l have been searchlng for Lhose polnLs and lncludlng Lhem. And we reach Lhe concluslon LhaL a pro[ecL of 100k, lL's dlfflculL someLlmes for cllenLs Lo undersLand lL, you spenL abouL 40 or 30 effecLlvely on developmenL, all Lhe resL ls whaL lnvolves handlng ln Lhe pro[ecL, all sorLs of managemenL noL only pro[ecL managemenL buL managlng Lhe Leam, Lhe lssues, noL [usL developmenL. 0,: es all Lhe communlcaLlons, all Lhose Lhlngs and preparlng. &7: And Lhen Lhe oLher slxLy aren'L [usL Lhls, lL's all Lhe LesLlng, Lralnlng and documenLlng. All LhaL ls nearly half Lhe pro[ecL. 0,: 1he nexL sLep ls do you generally have deflned Lhe crlLlcal facLor for lL Lo be consldered successful? &7: 1he maln feaLures belng avallable. 0,: uo you deflne aL Lhe sLarL, whlch are Lhe maln feaLures for lL Lo be ok? 1hls because l don'L have as much percepLlon abouL sugar pro[ecLs. AL a slLe, ls Lhe slLe belng operaLlonal, buL maybe lL's noL [usL LhaL. AL a sugar pro[ecL lL may be dlfferenL. &7: l'll glve you an example, 1ecnldelLa, for Lhe pro[ecL Lo be worklng Lhere are several modules and normally Lhose modules are ldenLlfled aL Lhe proposal, lf Lhe proposal says sLorage managemenL, Lhen Lhe sLorage managemenL needs Lo be avallable. now are all Lhe feaLures presenL? 1haL ls a posLerlor breaklng down. 8uL for Lhe pro[ecL Lo be a success ls, really, Lhe users havlng a worklng sLorage managemenL feaLure. 0,: Ck &7: Lasy. lL's llke LhaL. Why does Lhe cllenL come Lo us Lo develop someLhlng? Pe has a need and he can'L do lL some oLher way or does lL ln an archalc manner, we have Lo guaranLee LhaL whaL he couldn'L do or dld ln an archalc manner, can now do wlLh our producL, and LhaL's Lhe declslve facLor. AL slLes, however, l can glve you verLl's case, Lhere were several slLes, acLually lL ls an appllcaLlon ln urupal's case, and
89 Lhey wanLed, ln Lhe case of Lhe occurrences, a porLal where all Lhe company could submlL conLenLs and have Lhe P8 deparLmenL Lo manage Lhose conLenLs, from Lhe momenL Lhey glve feedback and Lhey can manage LhaL feedback lL's success. 0,: ls Lhere a comparLmenLallzaLlon of Lhe polnLs where Lhe modules are lnsLalled Lo deLermlne, for lnsLance. &7: 1he cllenL doesn'L wanL Lo know whaL modules are lnsLalled. 0,: 1haL's noL Lhe quesLlon, from Lhe feaLures, for lnsLance when lL ls esLlmaLed a cerLaln Llme and acLually Lhe developmenL requlres a longer Llme, aL leasL aL Aglle lL ls sLaLed, Lhen see wlLh Lhe Llme avallable whaL can. &7: And LhaL ls someLhlng, slncerely ln every pro[ecL l've parLlclpaLed up Llll now, l sLlll haven'L had Lhe opporLunlLy Lo do and LhaL ls always golng Lo depend upon Lhe cllenL's menLallLy and lf lL ls closed aL Lhe beglnnlng or noL, lf lL's Llme and maLerlals. And whaL happens always ls LhaL Lhere are Lhree or four blocks of feaLures whlch are conLemplaLed aL Lhe proposal and was accepLed by Lhe cllenL, Lhen you are golng Lo deLall, lL ls besL Lo be very deLalled Lo know whaL lL conLemplaLes, when lL's noL, you have Lo deLall afLerwards. 1here are cllenLs LhaL don'L wanL Lo know more, you develop whaL you compeLed for and lL's Lhe pro[ecL manager LhaL guaranLees LhaL Lhere ls all LhaL was supposedly sold, some oLher cllenLs are lnvolved and Lhey ask for more and here Lhere are Lwo ways: one - lL was noL conLemplaLed, lL Lransfers Lo anoLher pro[ecL LhaL you wanL Lo pay afLer, normally on Lhe pro[ecLs LhaL l'm lnvolved wlLh, lL's Lhls LhaL happens even ln my former company. lf lL's ouL of Lhe pro[ecL's scope, we'd do a proposal of conLlnuous malnLenance and upgradlng or a new pro[ecL Lo whlch we would assemble Lhose funcLlonallLles. now l always Lry, lf lL ls posslble for Lhe cllenL, because Lhey have closed budgeLs and lL lsn'L always as easy or as agllely as Aglle wanLs. WhaL Aglle says ls ok you need Lo have all Lhe feaLures or he asks for one or one Lakes longer, Lhe lasL one drops and Lhen you have Lwo ways or lL dles ouL and lL ls never done or he pays more and you feed lL blgger. 0,: AcLlvlLles, success. So baslcally we can say LhaL Lhe pro[ecLs we have worked on are all consldered successes. &7: ln whlch parameLers? 0,: 8y your parameLers. &7: noLhlng ls losL, all ls Lransformed. Lven a bad experlence a pro[ecL you consldered wenL wrong, why because we spenL longer or efforL Lo develop and we losL money wlLh lL, because Lhe cllenL send lL Lo Lhe Lrash can, Lhe several reasons why Lhe pro[ecL may have wenL wrong, you always learnL and Lhe nexL pro[ecL you'll do you wlll
90 be able Lo avold Lhose Lypes of slLuaLlons, avold some confronLaLlons. A pro[ecL wenL wrong, because you aggravaLed yourself wlLh Lhe cllenL and you had an argumenL and he wenL away, you now, for Lhe nexL pro[ecL, Lhe llmlLs you have Lo guaranLee for LhaL noL Lo happen agaln, so Lo me ln global Lerms all ls success, beLLer off, noL success, wlLh every pro[ecL wheLher Lhey are good or bad you wlll learn and LhaL ls always a success. 0,: And for Lhose problems, Lhey are always Llme problems or ls Lhere any oLher problems LhaL condlLlons. &7: no, lL can be Lechnology or lack of vlslon from Lhe cllenL's parL or from our parL, we don'L undersLand someLhlng, LhaL's noL Lyplfled l guess. 0,: uo you have any ldea on Lhose problems? 1he Llme, Lhe cllenL's vlslon. &7: Pere, as you now, l have eleven years' experlence from anoLher company and l can glve you several examples of Lhlngs LhaL wenL wrong and why. Pere, from Lhe pro[ecLs l'm ln and l can say someLhlng has gone wrong. When we say wenL wrong whaL do you mean? ls noL worklng wlLh Lhem agaln or Lhe cllenL belng unsaLlsfled? lrom verLl he was saLlsfled and hasn'L conLlnued, because Lhey have no budgeL aL Lhe momenL, 1ecnldelLa ls glad and conLlnues Lo works wlLh us, Lhere are some problems on Lhe cllenL's slde and on communlcaLlon wlLh oLher parLners LhaL need Lo be done, SlAS Lhere ls no problem, llfLy-slx Lhe problem LhaL may exlsL ls Lhe mlnor prlorlLy LhaL has delayed Lhe pro[ecL. now, oLher pro[ecLs, Lhlngs gone wrong ln my oLher experlences: ou have much efforL on Lhe pro[ecL and suddenly Lhe cllenL ends Lhelr budgeL you need Lo drop lL, lL dled, on anoLher cllenL Lhe pro[ecL ls ok, flne, you dellver lL, Lhe cllenL changed Lhe dlrecLor and he does lL all agaln wlLh anoLher company, for you lL's a success because you dld lL and dellvered lL buL lL's done. 0,: uld you geL pald? &7: es, lL's Lhe same Lhlng lf you buy a 1v, ok l don'L llke lL, lL shouldn'L be black, buL lL ls whlLe. bummer, your speclflcaLlons sald you wanLed lL black you wanL a whlLe one, buy anoLher. lL wenL wrong, we dldn'L do anyLhlng Lhe cllenL wanLed, ls LhaL correcL, and ls lL wrlLLen you valldaLe lL and you flnd Lhe cllenL was rlghL you do lL agaln or Lhe cllenL doesn'L pay and we are sLlll ln good Lerms, lL depends. 8uL Lhe quesLlon, l am Lalklng Lo you abouL Lhe negaLlve aspecLs, buL your quesLlon was abouL Lhe poslLlve aspecLs. 0,: no, lL was: WhaL are Lhe maln problems LhaL lead a pro[ecL noL Lo be deemed as a success? lL's exacLly whaL you were saylng.
91 &7: Success for whom? And whaL are Lhe parameLers? lf you are from Lhe flnanclal area, Lhe pro[ecL ls noL successful lf you have 0 or loss. 1he Llme consumed and Lhe esLlmaLed was Lhe same lnsLead of belng lucraLlve, a cenL earned, a cenL spenL", you goL experlence and noLorleLy, you can say you dld LhaL pro[ecL, you can earn more pro[ecLs for LhaL cllenL, Lhe cllenL ls very demandlng and lL Lakes Loo long Lo pay. lL depends on whaL's poslLlve or noL ln Lhe flnanclal area, ln Lhe commerclal area, you dld lL, lL was ok, buL lL doesn'L sulL you Lo say you dld lL. 0,: es, Lhls ls [usL from a managerlal polnL of vlew. &7: ln pro[ecL managemenL l Lhlnk lL ls never a fall, lL's a fall. l have already had cllenLs, noL here buL elsewhere, we dld lL, lL's ln producLlon, and Lhe people are uslng lL, buL l never wanL Lo see LhaL cllenL agaln because he ls very annoylng. ou can'L say Lhe pro[ecL was noL a success, Lhe personal relaLlonshlp may someLlmes spoll Lhe pro[ecL managemenL. lL's Loo sLressful. l forgoL ConLrollnvesL: we developed Lhe Lhlngs are ln producLlon, Lhe cllenL ls a blL annoylng buL ok. So. 0,: So Lhere's noL a frequenL problem. &7: 1he recurrenL problem Lo Lhe pro[ecLs ls Lo puL a break on whaL Lhe cllenL asks whlle Lhe pro[ecL ls runnlng. 8ecause he always wanLs more, he wanLs everyLhlng and even lf lL's noL conLemplaLed lf you glve hlm a candy, he wlll Lake lL, buL he rarely wanLs Lo pay for Lhe candy. lf you don'L puL a sLop Lo Lhese demands, you end up glvlng wlLhouL hlm recognlzlng he's geLLlng lL for free. 0,: So lf you have a sysLem Lo supporL you, lL helps you on LhaL. &7: 1o supporL Lhls, lL's whaL happens wlLh Lhe producL backlog, now l puL lL lnslde and ok you say LhaL" lL's noL lnslde Lhe pro[ecL (scope) buL l undersLand your need, aL Lhls momenL we are focuslng on developlng whaL we commlLLed Lo do, when we flnlsh Lhls pro[ecL we speak abouL lL. SomeLlmes Lhey forgeL and never speak of lL agaln. 0,: (laughLer) &7: buL lL's a moLlve Lo collecL all and aL Lhe end Lhey pay for LhaL one and Lhen you remlnd hlm and propose, lf he says yes you glve hlm a budgeL and he accepL or noL. 0,: Ck, aL Lhe Llme Lhls was made, !l8A wasn'L very developed sLlll. Pave you already LesLed !l8A? &7: es, and l'm lovlng Confluence. 0,: Ck.
92 &7: And !l8A Loo. 0,: l've used lL now for Lhe Lralnlng and lL came ln handy. uo you Lhlnk Lhere ls someLhlng LhaL needs Lo be lmproved? Can you polnL ouL anyLhlng LhaL needs Lo be lmproved? &7: ln whaL? 0,: ln !l8A or Confluence, ln Lhe sysLem. &7: We are sLlll ln a learnlng perlod and knowledge, we have already deLecLed some mlsLakes LhaL are appearlng ln Lhe reporLs. (lnLerrupLlon - phone call) 0,: We were Lalklng abouL problems LhaL may be happenlng aL !l8A. &7: 1hey were [usL Lhose ones. 0,: WhaL were Lhey? &7: 1he reporLs aren'L reporLlng some Lhlngs correcLly, some Lhlngs LhaL you feel lnLulLlvely lL should be capable Lo do buL lL doesn'L, Lhen you can do lL ln anoLher way. lor lnsLance, a masslve copy of lssues from one pro[ecL Lo Lhe oLher you can'L do, you have Lo copy lssue by lssue, a masslve updaLe, and Lhlngs llke LhaL, buL lL's [usL deLalls. 0,: uo you Lhlnk you can sLlll beneflL even more on Lhe way you can work wlLh Confluence? &7: 8eneflL more, yes, yes. and over Llme even more. 0,: Lven more. &7: es, l'm sLlll a blL green, 20 of whaL l can Lake ouL of resulLs. 0,: uo you Lhlnk ln Lhe Lool, ln Lhe LM, Lhere ls someLhlng mlsslng? &7: l can'L answer LhaL because l don'L know all of lL, do you undersLand? l have noLlced we have some wldgeLs / macros we can lnclude and Lhere aren'L all Lhere, so we need Lo search on Lhe neL whlch ones come ln handy and geL Lhem. l don'L know all of Lhem, l LhoughL oh Lhls one would be good, Lhls ls beLLer Lhan LhaL and LhaL one could do LhaL. buL l cannoL answer LhaL wlLh cerLalnLy from whaL l know. WhaL l can Lell you ls my flrsL commenL abouL Lhe ALlasslan Confluence / !l8A was 'uh! Pow ugly!' 8ecause l dld a loL of research and l was used Lo seelng Lhlngs wlLh a cuLer graphlc lnLerface, buL LhaL's noL Lo be cuLe, lL's Lo be funcLlonal so really as cuLer lL ls, Lhe slower lL geLs Loo, normally. So LhaL ls exacLly wlLh Lhe conLenLs you need, when you need Lhem, wlLh Lhe permlsslons you need, wlLh Lhe shorLcuLs you need and wlLh Lhe llnks you need, lL can ad[usL and l really llke Lo ad[usL wlndows Lo be able Lo see Lhe mosL lnformaLlon l posslbly can and lL has cerLaln llnk poslLlonlng LhaL does lL for you.
93 So you can have ln your vlslon fleld all Lhe lnformaLlon you need Lo solve a cerLaln lssue. We wln by noL belng Loo complex, Lhe slmplesL lL can be. lL could be more appeallng, buL lL has whaL lL needs and no more. lL's greaL. 0,: So a Lek sesslon soon. &7: Are you dolng lL? 0,: no, you are. &7: (laughLer) .. uone? 0,: uone. &7: bye! See you Lomorrow 0,: 1hank you a<;<>< (LIEFOJEU > ! lnLervlewer: Mlguel erelra ! lnLervlewed: aulo MrLlres ! uaLe: lrlday, 7Lh of uecember of 2012
0,: llrsL quesLlon, well l'll ask. lrom Lhe pro[ecL managemenL polnL of vlew, how does Lhe lnformaLlon geL Lo you? Where does lL come from? uoes Lhe commerclal area pass Lhe lnformaLlon on Lo you? Cr whlle Lhe commerclal area ls deallng wlLh Lhe cllenL do you go Lalk Lo hlm? ,0: lL depends, you have boLh cases. 1here are pro[ecLs LhaL arrlve preLLy much already deflned by Lhe commerclal area and Lhe cllenL, buL Lhese are rare, normally whaL arrlves ln an lnlLlal sLage ls very llLLle lnformaLlon and mosL of Lhe Llmes Lhe pro[ecL ls already sold. 1hen on Lhe oLher hand, more and more even before a proposal ls made, who manages Lhe pro[ecL or who conLrlbuLes Lhe mosL sLraLeglcally ln an lnlLlal sLage ls presenL durlng Lhese flrsL meeLlngs and geLs much more lnformaLlon aL LhaL Llme. Anyway we are always speaklng abouL very global lnformaLlon, only from Lhe momenL when you sLarL Lo do Lhe pro[ecL wlll you effecLlvely pull ouL whaL you really need. And ln Lhe ma[orlLy of Lhe pro[ecLs Lhe lnformaLlon lsn'L glven Lo you, you need Lo flnd lL ouL. 8ecause elLher Lhe cllenL doesn'L know or nobody LhoughL of lL. ln a general manner you never geL Lhe lnformaLlon you need ln Lhe beglnnlng of Lhe pro[ecL. 0,: ou go and geL lL dlrecLly wlLh Lhe cllenL? ,0: es, lL's usually so.
94 0,: So baslcally Lhe flow of Lhe pro[ecL ls commerclal area, Lhen pro[ecL managemenL and flnally developmenL? ,0: lf you Lhlnk of lL ln a normal pro[ecL managemenL perspecLlve, lL's llke LhaL. 8uL ln my case, l'm presenL, mosL of Lhe Llmes, when Lhe pro[ecL lLself doesn'L exlsL, yeL. l access lnformaLlon aL a sLage when lf l was merely a pro[ecL manager ln Lhe normal flow l wouldn'L. lncreaslngly less l'm Laken by surprlse. 0,: uo you Lhlnk LhaL ls our way Lo do lL, LhaL you are Lhere? Cr are you seelng lL from Lhe ux perspecLlve? M: lL lsn'L abouL Lhe ux perspecLlve, lL's really belng called. More and more lL ls Lrled Lo do lL llke LhaL: Lhe Leams, Lhe pro[ecL managers or sLraLeglsLs Lo be presenL aL an lnlLlal sLage of Lhe pro[ecL. lrequenLly Lhe pro[ecL doesn'L exlsL, yeL. We are only lnvesLlgaLlng and make proposal or gaLher lnformaLlon Lo do a proposal, buL l Lhlnk Lhere has been an efforL Lo Lry Lo lnvolve Lhe Leam aL Lhe early sLages of Lhe pro[ecL. lL lsn'L enough Lo have Lhe manager or C or whomever wlLh all Lhe lnformaLlon, you need Lo have Lhe whole Leam or else lL affecLs everyLhlng. l Lhlnk Lhere has been an efforL accordlngly. ln my parLlcular case l geL lnvolved ln Lhe beglnnlng, because l care, buL Lhere ls an efforL from Lhe company ln LhaL dlrecLlon. 0,: So Lhe pro[ecL preparaLlon, how would you say lL ls? Pow ls Lhe process of beglnnlng a pro[ecL? 1he meeLlngs. ls Lhere any maLerlal already made? lor example aL Lhe Aldelas" you have Lhe lnformaLlon archlLecLure, buL LhaL ls ln a more advanced sLage. how do you do Lhe lnlLlal seLup Lo sLarL? ,0: lL varles ln each pro[ecL, ln Lhe case of Aldelas" lL's a dlfferenL pro[ecL, you know you have Lo do Lhree componenLs, you LhaL are some obvlous Lhlngs LhaL need Lo be a parL of Lhose componenLs buL Lhe mosL of Lhe lnlLlal sLage of Lhe pro[ecL ls Lo undersLand exacLly whaL Lhe problem was, aparL from whaL Lhe cllenL LhoughL lL was and Lhen have a sLraLeglc approach Lo come up wlLh a soluLlon. So Lhe pro[ecL's scope changed afLer lnvesLlgaLlng and dolng a workshop, and undersLandlng Lhe problem wasn'L such and such, buL really LhaL. So here Lhere was a dlscrepancy ln Lerms of Lhe proposal, whaL Lhey wanLed was a sLore and a communlLy and Lhe porLal wasn'L all LhaL lmporLanL, buL Lhelr maln problem was really Lhe porLal. 1hey need a sLep Lo geL Lo people, so Lhe Lhlngs were lnverLed. 8uL Lhen you have pro[ecLs where lL ls much more llnear. ou know exacLly whaL you have Lo bulld, you may noL know every llLLle deLall and dlscover Lhem on LhaL flrsL sLage or along Lhe pro[ecL. lL varles a loL, Lhere lsn'L exacLly a model. And moreover by a Llme consLraln and laLely by meLhodology you don'L need everyLhlng aL flrsL Lo
93 sLarL developlng. ou Lry Lo have Lhe essenLlal, you need Lo know Lhe way and Lhe seL of feaLures LhaL are Lhe basls, buL rapldly we can sLarL worklng, whlle before we sLarLed worklng wlLhouL knowlng exacLly whaL we were golng Lo do. 0,: ou deflne a guldellne and Lhen aL each lLeraLlon you. ,0: LxacLly. 0,: 8reak down. ,0: lf you really llke Lo be Scrum or follow a more Aglle sLyle you need Lo follow LhaL paLh. ou may have a dlscovery phase, you should have lL, lL can be more or less sLrucLured and based on LhaL lnlLlal dlscovery sLage, you may call lL sprlnL zero or whaLever, you bulld a more asserLlve backlog. And supposedly Lhen Lhe Leam has Lhe condlLlons Lo sLarL worklng, lL's noL exacLly whaL happens. lL doesn'L necessarlly happen llke LhaL all Lhe Llme buL Lhe efforL ls Lo walk LhaL paLh? 0,: lsn'L Lhere a sLrucLure for LhaL sprlnL zero? ,0: noL yeL. 0,: noL yeL. 8uL ls lL planned? ,0: ln my parLlcular case l'm Lhlnklng of lL, because Lhe web area has cerLaln parLlcularlLles LhaL maybe Sugar and oLher Lhlngs don'L, buL l Lhlnk so lL's belng LhoughL of. 0,: Cludlo spoke of all Lhe sLeps LhaL are ln a pro[ecL: Lhe beglnnlng. l don'L have lL deLalled here aL Lhe momenL. Lhe second Lo lasL ls Lhe developmenL and Lhen Lhe dellvery and Lhe follow up, l can'L remember Lhe flrsL ones now. 8uL Lhe way you manage Lhe pro[ecL, are all Lhe sLages dlfferenL? ls Lhere anyLhlng ln common? M: es, ln a general manner all Lhe pro[ecL have Lhe same sLrucLure. ou have Lhe lnlLlal sLage of dlscovery, requlremenLs' assessmenL, sLudy soluLlons eLc. 1hen you have Lhe acLual developmenL: Lhe deslgn, Lhe ux, Lhe programmlng, Lhe LesLs. lf you are worklng ln an aglle sLyle you don'L do lL, you don'L LesL [usL ln Lhe end, Lhey are belng LesLed. 0,: es, aL Lhe end of each lLeraLlon. ,0: LxacLly. 0,: AlmosL aL Lhe end of each lLeraLlon. ,0: 1hen supposedly, you have a deploymenL sLage and Lhere you perform measuremenLs and Lhe upgrades or follow up. lL's one of Lhose Lhlngs, lL depends on Lhe way Lhlngs are sLrucLured and sold. 8uL Lo me, lL makes sense LhaL when you have a pro[ecL, rlghL ln Lhe beglnnlng, you deflne whaL Lhe mlnlmum producL ls, or
96 whaL ls essenLlal you dellver LhaL and you save Llme and money from Lhe pro[ecL Lo keep on maklng lL beLLer. WhaL ls normally done ls you develop Lhe pro[ecL wlLhln LhaL budgeL and dellver lL and Lhen Lhe lmprovemenLs, noL a supporL, buL Lhe lmprovemenLs are a new pro[ecL, wlLh a surcharge. 1he ldeal was Lo make a budgeL, dellver Lhe pro[ecL under Lhe budgeL and Lhen save Llme Lo lmprove. lrom Lhe momenL you have Lhe pro[ecL onllne or belng used, you can pull meLrlcs and Lhlngs Lo lmprove on, up unLll Lhen you're guesslng. 0,: 8uL LhaL way of Lhe llfecycle of Lhe pro[ecL ls always Lhe same. lndependenLly from Lhe level, l don'L know how Lo call lL, more or less complex or wlLh a larger budgeL. ,0: AL flrsL, yes. lf lL's a pro[ecL, noL [usL move someLhlng qulckly, you always have LhaL. ou always have Lhe lnlLlal sLage of dlscovery and preparaLlon, Lhen Lhe acLual developmenL and lnslde lL you may have pro[ecLs wlLh a large deslgn componenLs and oLher noL really, you have oLhers wlLh more fronLend, oLhers sLlll wlLh more fronLend and backend, buL lf you conslder lL Lo be developmenL, all you are dolng ls producLlon, all have exacLly LhaL, whlch ls l have Lhe dlscovery, Lhe execuLlon and aL Lhe end Lhe deploy, LesLs and follow up. 0,: ldeally lL ls really Lo dellver Lhe documenLaLlon and Lhe LesLs . ,0: LxacLly. 0,: Ck. l had Lhe quesLlons ln Lhree groups, from Lhls group. 8esldes Lhe developmenL sLage, whlle you are ln conLacL wlLh Lhe programmers, well you musL always be ln conLacL, rlghL? ,0: Always. 0,: Pow ls Lhe communlcaLlon wlLh Lhe remalnlng deparLmenLs done? 1he flnanclal or Lhe commerclal area. ls lL done verbally, by e-mall, ls Lhere a supporL llke !l8A or.? ,0: no, Lhere lsn'L. lL's done verbally and by e-mall. 1here lsn'L any oLher. well, Lhere are meeLlngs, follow up meeLlngs. 0,: llnances? ,0: es, when you have Lhe C meeLlng Lhere ls someone from flnanclal area Lhere Lo see how Lhe pro[ecLs are golng Lo percelve when Lhey can blll. 0,: Ck. ,0: 1haL klnd of Lhlnk, buL oLherwlse lL ls done ln an arLlsanal way. SomeLlmes someone asks abouL lL and you say lL's Lhls or LhaL. 0,: uo you Lhlnk.
97 ,0: We are walLlng Lo close by Lhls Llme. usually Lhere lsn'L much communlcaLlon, wlLh Lhe commerclal area Lhen lL's very rare, unless Lhe cllenL annoys Lhe commerclal, or Lhe commerclal needs Lo know a concreLe lLem. WlLh Lhe flnanclal area Lhere ls usually a conLacL when, from Lhelr parL, Lhey ask us lf Lhey can blll or lf Lhey can send Lhe lnLermedlaLe paymenL or someLhlng llke LhaL. 8uL Lhere lsn'L a process, lL's noL a regular Lhlng, you're noL always Lalklng wlLh Lhe flnanclal area Lo accounL for. 0,: Aren'L Lhe paymenLs done aL Lhe end of each sprlnL or someLhlng? ,0: no, you have a paymenL plan whlch ls agreed wlLh Lhe cllenL and lL varles accordlngly Lo each pro[ecL. 1here are cllenLs LhaL have moLlves, llke for lnsLance ln Lhe case of Aldelas, Lhey need Lo make Lhe paymenLs ln cerLaln Llmes when Lhey geL Lhe money from Lhe communlLy funds. 1here are oLhers, LhaL lL's ln Lhe beglnnlng, Lhen 30 more whenever and Lhe resL ln Lhe end, or forLy, forLy, LwenLy. Cr every monLh Lhey pay x, l Lhlnk lL happened noL LhaL ofLen. 1here are several models LhaL can be ad[usLed Lo Lhe cllenL. 0,: Ck. ,0: 8uL you don'L deal much wlLh LhaL, as a C you deal malnly wlLh esLabllshlng a budgeL, obLaln an execuLlon schedule and your concern ls LhaL, Lo keep Lhe schedule and lnslde Lhe budgeL. 0,: AfLer all LhaL, whaL's Lhe Lyplcal dlfflculLy you flnd whlle managlng pro[ecLs? ,0: Pere parLlcularly lL ls managlng Llmes, you have a greaL dlfflculLy knowlng exacLly how long you sLlll have for Lhe pro[ecL. lL's belng mlLlgaLed, buL lL was a dlfflculLy, because l remember LhaL for a very long Llme we couldn'L Lake ouL Sugar reporLs for Lhe pro[ecL. l don'L know how long was already spenL or l do Lhe maLh by hand or l don'L know. Lven now Lo do lL, lL ls almosL by hand Lo undersLand exacLly. Cn Lhe oLher hand you have esLlmaLes done ln a cerLaln way, Lhere are some Lhlngs LhaL aren'L Laken lnLo accounL, Lhere are hours LhaL geL ln laLe, people have Lo flll ouL LlmesheeLs and someLlmes lL's a blL confuslng and we don'L undersLand exacLly where have so many days gone lnLo developmenL, when ln effecL you know lL wasn'L exacLly llke LhaL. 1he dlfflculLy ls Lo know exacLly how long you have spenL, how long you have lefL, ln whaL sLage exacLly ls Lhe pro[ecL. Cne Lhlng ls knowlng you have LwenLy days buL Lhe pro[ecL ls 90 done and Lhe oLher ls knowlng you have LwenLy days buL Lhe pro[ecL ls 2, so . you sLarL havlng mechanlsms Lo undersLand buL LhaL's noL very easy yeL.
98 0,: usually when people answer Lhe flrsL, Lhey Lend Lo answer Lhe second: Pow can you mlLlgaLe Lhese problems? lL's belng made. ls lL wlLh Lhe supporL from Lhe Lool? ,0: LxacLly, you have !l8A and everyLhlng you work on should have an lssue and LhaL lssue musL have an esLlmaLe Llme asslgned and Lhus, Lhe Llme should be auLomaLlcally assumed and lL should glve you Lhose meLrlcs rlghL away. lL's noL exacLly so, buL Lhere ls an efforL belng done ln LhaL dlrecLlon. 0,: es, buL Lhese esLlmaLes are always more complex, Lhey requlre a sLrucLure. ,0: 1he problem ls LhaL Lhere ls normally a dlfference, someLlmes large, beLween Lhe esLlmaLes Lo provlde a budgeL ln a sLage you sLlll don'L know much abouL Lhe pro[ecL, buL Lhen lL ends up belng a reference for Lhe Llme you have Lo execuLe lL. And Lhere resldes Lhe dlfflculLy ln marrylng Lhe Lwo. SomeLlmes Lhe pro[ecL ls way more complex and you need Lo do more Lhlngs and lL lsn'L posslble or you have Lo flL ln wlLh whaL was sold. Someone sold 30 days, buL Lhe problem ls LhaL Lhose aren'L 30 calendar days buL 30 l1Ls, LhaL ln effecL can be one week lf you have slx people worklng. 0,: So, accordlngly Lo Lhe way esLlmaLes are done, aL Lhe momenL, you have no way of comparlng Lhe esLlmaLes wlLh Lhe real execuLlon Llme by Lask. ,0: 8y Lask you are sLarLlng Lo have lL, before Lhere was no way. ou had a very global ldea by Lype of acLlvlLy or sLage. 1he deslgn used up x days, Lhe developmenL used up whaLever, buL you had no way of knowlng, ok Lo seL up a CAS Look abouL x Llme, Lo make Lhe fronL end lL Look whaLever, buL now you sLarL Lo have a way. ou can undersLand LhaL maybe your esLlmaLe whlch was sold and you made Lhe pro[ecL from doesn'L maLch aL all wlLh Lhe real execuLlon. lL's a quesLlon of undersLandlng whaL falled: was your esLlmaLe lncorrecL? Cr who was coordlnaLlng Lhe lmplemenLaLlon wanLed Lo do more Lhan whaL lL was supposed Lo? lL's always compllcaLed. Cn Lhe oLher hand, lL's a quesLlon of orlenLlng by Lhe value Lo Lhe cllenL. Always Lrylng Lo keep on budgeL evldenLly, and Lhe quesLlon of whaL's good for Lhe company buL orlenLed Loward Lhe cllenL. ou normally have Lhe Lendency Lo puL more Lhan., alLhough you don'L have Llme you dellver more ln Lhe besL posslble way, lL makes sense for Lhe pro[ecL and Lhe cllenL, lL's a quesLlon of vlslon. l don'L Lake parL ln LhaL, Lhere are Lhose who Lhlnk we have a budgeL and lf lL fall under ok, lf lL doesn'L Lhough luck. We are noL worklng Lo no avall.
99 0,: So based on. you don'L have Lhe posslblllLy Lo see, yeL or you do. uo you have any ldea on whaL are Lhe acLlvlLles LhaL are usually underesLlmaLed? ,0: es, normally Lhe bulk ls Lhe deslgn parL, LhaL lL's noL underesLlmaLed. 0,: lL's underesLlmaLed. ,0: l mean LhaL lL ls underesLlmaLed noL because people Lhlnk lL Lakes less Llme, buL because normally ln commerclal Lerms ls sold by less or Lhey dldn'L accepL more, lL's geLLlng beLLer. normally Lhe deslgn parL Lakes longer, because lL's a collaboraLlve labor and lL lnvolves more people. lL's a blL underesLlmaLed. 1he developmenL parL, Loo. 0,: ls lL underesLlmaLed? ,0: l Lhlnk so. l Lhlnk lL's normally esLlmaLed accordlng Lo Lhe feaLures on Lhelr baslc componenLs, l need Lo geL lL worklng and lL wlll Lake x Llme and normally lL lacks Llme Lo reflne lL. Cne Lhlng ls havlng Lhe form sendlng an e-mall, anoLher ls havlng a preLLy form wlLh error messages, Lhe preLLy e-malls and all Lhose Lhlnks. 8aslcally lL's all Lhe same feaLure, Lo have a form, buL Lhey only Lhlnk Lo have Lhe form l Lake x Llme, buL Lhen lL comes Lhe C or someone and Lells you lL's lacklng Lhe error messages, Lhls or LhaL. ln a general manner l Lhlnk everyone underesLlmaLes Lhe Lasks. 1hen you have oLhers LhaL Lhrow lL upwards, as Lhey have no ldea and Lhey don'L wanL Lo geL burnL, Lhey say LhaL's a whole week and when Lhe Llme comes lL's done ln a day or less, buL ok. 1haL's a quesLlon of Llme, experlence, havlng meLrlcs LhaL more or less allow you Lo undersLand LhaL. 1he ldeal ls for example, worklng on a pro[ecL wlLh urupal and you geL LhaL afLer 4 or 3 pro[ecLs LhaL on average LhaL seL of feaLures Lakes x long. ln Lhe beglnnlng, when Lhe C ls dolng Lhe lnlLlal esLlmaLes or Lhe plannlng sprlnL wlLh Lhe Leam already has Lhe noLlon LhaL Lhey usually Lake x long Lo do Lhls, somewhere ln beLween Lhe elghL and Lhe elghLy, ln Lhe mlddle Lhere's an average. AL Lhls Llme, we can'L make any calculaLlon of LhaL sorL. 0,: ls Lhere a crlLlcal success facLor deflned by you or oLhers for each pro[ecL.? lf Lhls lsn'L meL Lhe pro[ecL lsn'L. ,0: no, Lhere lsn'L any sLrucLured Lhlng. lL's obvlous LhaL lf you're dolng a porLal or an appllcaLlon Lhey have Lo work, lf Lhey don'L work you have a problem, buL you don'L have crlLerla. 0,: lL's noL commerclally deflned... ,0: ou don'L have anyone who says, Lhls needs Lo work so and so or lL's noL vlable. Who deflnes Lhe crlLerla lL's us and Lhe cllenL.
100 0,: 8aslcally lL's: Lhls exlsLs, all needs Lo work. ,0: LxacLly, lL's a quesLlon of commonsense and Leam ad[usLmenL. lL's llke Lhe form sLory, you can have a form worklng correcLly wlLh Lhe error messages and all. lL works, lL's all done, and Lhere ls noLhlng whlch lsn'L done. ln mere Lerms Lhls ls an accepLance crlLerlon, lL works, we can send an e-mall, lL reglsLered, lL saves on Lhe daLabase, Lhe user goL Lhe e-mall, he cllcked, conflrmed lL's all worklng. 8uL Lhen you can have, lf Lhere ls Llme and wlll, you can have exLras Lhls wlll do lL ln A[ax or whaLever, buL lL's noL deflned. lL's one of Lhe Lhlngs LhaL needs Lo sLarL belng deflned by Lhe C, Lhe accepLance crlLerlon of a cerLaln Lask. lf Lhe crlLerlon ls LhaL Lhe form needs Lo be ln A[ax and change color, lL was deflned by hlm, or Lhe Leam explalns why Lhey can'L do lL ln LhaL manner or lf Lhey don'L, he has a rlghL Lo say lL lsn'L done, and geL back. usually, on Lhe cllenL's slde, LhaL doesn'L exlsL. 1he cllenL doesn'L deflne Lhe accepLance crlLerlon, lL's noL expressed. 0,: So lL's more a C's declslon Lhen Lhe cllenL's. ,0: es. 0,: 1he cllenL ls [usL Lhe maln scope. ,0: LxacLly. 1he C needs Lo have Lhe ablllLy Lo undersLand when someLhlng ls ready Lo be dellvered Lo Lhe cllenL, and ln advance he already knows lf Lhe cllenL accepLs lL or noL. up unLll now, Lhere weren'L exacLly Lhlngs belng reLurned. 0,: Ck. ,0: And Lhe ones whlch dld weren'L our faulL. lL's because he changed hls mlnd or he expressed hlmself lncorrecLly. 1here was someLhlng ln Lhe mlddle, an exLernal facLor whlch was noL relaLed Lo Lhe pro[ecL. 8uL Lhe one who makes LhaL fllLerlng should be Lhe C. As lf you have a Lechnlcal responslble for a Leam or deparLmenL lL should be someone who looks aL Lhe pro[ecL and sLaLes LhaL from Lhe Lechnlcal polnL of vlew lL's correcL and ready Lo be shown or puL lnLo producLlon or whaLever. 0,: So Lhere never was any pro[ecL. you referred some LhaL were reLurned. 8uL Lhe maln problems are from Lhe cllenL's parL or Lhe communlcaLlon wlLh Lhe cllenL? ,0: 1here was never someLhlng whlch was developed, aL our speclflc case everyLhlng LhaL was reLurned was by lssues wlLh Lhe deslgn noL funcLlonallLy. lL's noL a quesLlon of Lhe porLal belng poorly made, or Lhe speclflcaLlons weren'L correcL buL slmply because Lhe cllenL changed hls mlnd or dldn'L even remembered approvlng lL. up unLll now, Lhe problem has always been on Lhe cllenL's slde.
101 0,: Ck. ,0: 1here ls a loL belng reLurned lnslde Lhe Leam. 0,: ln LesLs. ,0: 1he reLurn here ls debaLable. ou conslder someLhlng ls done, someone looks and says lL lacks someLhlng or lL doesn'L comply wlLh Lhe deflned crlLerla or adopLed sLandard buL LhaL's normal. Malnly because Lhere lsn'L a deflnlLlon of Lhe sLandards anywhere. lL needs Lo be deflned. 0,: 1here are some, buL. ,0: As l refer Lo sLandards, lL's Lhe accepLance crlLerlon of whaL ls consldered done, made. lL's easler Lo wrlLe l need a form Lhen l need a form LhaL does LhaL, and lf LhaL doesn'L happen, Lhls needs Lo happen so my accepLance crlLerla are Lhese. lL needs Lo send malls, Lhe oLher needs Lo recelve and lf lL doesn'L lL needs Lo do whaLever. ou already know LhaL someone who ls developlng a Lask knows exacLly whaL lL needs Lo do, whaL means belng done and whaL's accepLed or noL. 0,: And whaL abouL Lhe quesLlon of Lhe cllenL saylng he doesn'L remember, ls Lhere anyLhlng belng done Lo mlLlgaLe Lhose problems? ,0: ou have no process and no Lools Lo do LhaL offlclally, as Lhe communlcaLlon wlLh Lhe cllenLs happens malnly ln person or by e-mall, Lhey are Lhe Lwo . 0,: Lhe Lwo channels. ,0: ou don'L exacLly have a collaboraLlve plaLform where Lhe e-malls LhaL geL ouL or come ln are always reglsLered for Lhe Leam knowledge. 1here ls Lhe commonsense pracLlce LhaL lf Lhere ls someLhlng asked vla Lelephone LhaL alLers or may go wrong, Lhere ls Lhe senslLlvlLy Lo reglsLer lL ln wrlLlng. Accordlng Lo whaL was spoken Lhere ls an e-mall wlLh lL and by prlnclple lf no one conLesLs lL, whlch was accepLed. uocumenL Lhe mosL. 0,: Lhe declslons. ,0: Lhe declslons. vla e-mall or you send a deslgn and you geL someLhlng saylng l llke lL very much, buL LhaL doesn'L work for you. WhaL works ls expresslng lL ls accepLed, go ahead or someone saylng yes Lo lL. ou don'L have a place where LhaL ls esLabllshed, saved or documenLed Lo everyone. We Lrled wlLh (acLlve)Colab buL. 0,: 8ecause Lhe cllenL dldn'L. ,0: 1here you go, lf you had someLhlng capable, Colab lsn'L Lhe besL example of lL. ou have a number of appllcaLlons where Lhe Leams collaboraLe wlLh Lhe
102 cllenLs and Lhe cllenLs go Lhere wlLhouL any problem. Maybe Lhe Lool wasn'L Lhe besL for LhaL Lo happen, lL depends of Lhe cllenLs. 0,: uo you Lhlnk Lhls one has Lhose feaLures? ,0: up unLll now l haven'L seen any of LhaL. 1he place where Lhe documenLaLlon ls, lsn'L avallable Lo lnLeracL wlLh Lhe cllenL. l have noL seen anywhere l can communlcaLe wlLh Lhe cllenL and lL ls reglsLered. So, no. 0,: Ck. AL Lhe end. WhaL abouL Lhe sysLem lLself? ou have LesLed lL lmmensely, now. When lL was belng developed dld you had a say ln lL? ,0: es, l had some lnpuLs. 1he Lhlngs were belng lmplemenLed ln a cerLaln way, llLLle by llLLle, geLLlng a loL of lnpuL from Lhe people. lL changed, you have a meLhodology whlch ls belng applled ln a cerLaln way, more forcefully. 1here ls sLlll Lhe lnpuL gaLherlng, people say whaL Lhey need Lo say, buL up unLll now Lhe meLhodology ls belng applled forcefully. 0,: uo you Lhlnk lL's worklng well? ,0: lL's sLlll early Lo Lell. 1here are some Lhlngs LhaL cannoL be applled exacLly as books Lell you Lo, because Lhe Leams are dlfferenL, Lhe people are dlfferenL and Lhe companles are dlfferenL. So l don'L belleve Lhere ls one slngle model appllcable Lo all. LeL's see. up Llll now Lhose changes weren'L lncorporaLed. WhaL Lhe Leams are saylng ls Lrylng Lo be mlLlgaLed lnslde Lhe consLralnL of Lhe meLhodology, buL Lhere were no changes. 0,: ln Lerms of meLhodology, abouL Lhe sysLem do you Lhlnk lL needs Lo be alLered? ,0: WhaL sysLem? 0,: !l8A, Confluence. ,0: l Lhlnk no Lool ls perfecL, none can predlcL every case. ou Lry Lo ad[usL accordlng Lo Lhe Lool you have, as you can'L change, you work wlLh lL and LhaL l Lhlnk people are dolng, whlle Lhey geL Lo know Lhe plaLforms. l geL LhaL ln Lerms of collaboraLlon and documenLaLlon Confluence ls good and brlngs advanLages and we are noL even geLLlng half Lhe producLlvlLy we could. AbouL !l8A, lL has some good lLems buL Lhere are oLhers, LhaL elLher by lssues of conflgurlng or lack of knowledge Lransforms lL ln a hurdle. Cn Lhe oLher hand you have a place where you can see all Lhe pro[ecLs and a vlsual represenLaLlon of whaL ls happenlng or ls golng Lo happen, people can log ln Lhe Llmes wlLh Lhe advanLages and dlsadvanLages LhaL brlngs, lL's a blL odd you can only log Llme on Lhlngs avallable aL !l8A, lL doesn'L glve you rellable meLrlcs. Cr you log Loo much Llme ln lssues LhaL really dldn'L Lake LhaL long, because
103 you need Lo seL 8 hours, well aL leasL 6, because 2 you can puL elsewhere, when for real Lhlngs don'L work ln LhaL way. Where do l puL Lhls half an hour here? uo l puL lL ln Lhe Lwo hours l have Lo speak wlLh my colleagues, buL whaL lf l have already spenL Lhose, because of oLher Lasks? 8uL Lhese are Lhlngs LhaL need Lo be Luned. ln a general way, !l8A ls beLLer Lhan whaL we had. 1he sulLe ls beLLer, now we wlll see. 0,: uo you Lhlnk lL's sLlll very much aL Lhe beglnnlng? AL whaL percenLage do you say we are aL? ,0: ou sLlll dldn'L have a whole complex pro[ecL belng managed Lhere. lf you really reallze, we are only uslng !l8A aL abouL one monLh, even less offlclally ls Lhe only Lool ln pro[ecL managemenL slnce november 14 Lh , aL Lhree weeks. 1here were already some pro[ecLs uslng lL before, buL Lhe LlmesheeLs and Llmes where elsewhere. l guess Lhree weeks or a monLh of lnLenslve usage lsn'L enough Lo draw concluslons. 0,: uo you conslder LhaL ls helps you wlLh Lhe Lasks you regularly have? Are you belng more efflclenL? ,0: l Lhlnk l'm belng more efflclenL, because l have less responslblllLy", one week sprlnLs wlLh four hour meeLlngs, no. 1he greaLesL advanLage of Scrum or Aglle ls Lhe collaboraLlve envlronmenL among Lhe Leams, and so you should need so much Llme wlLh rlLuals, because Lhlngs should happen organlcally, durlng Lhe sprlnL. 8uL wlLh one week sprlnLs, you can'L be organlc. ou sLarLed Lhe sprlnL and suddenly lL has ended. We can'L even undersLand much. When we have sprlnLs of Lwo, Lhree weeks lL wlll dlfferenL, posslbly. ou have a meeLlng sprlnL plannlng reLrospecLlve LhaL Lakes half a day, or even a whole day, LhaL happens someLlmes, Lhen you sLarL a sprlnL wlLh [usL four days, you don'L have Llme Lo enLer.. 0,: Lhe developmenL. ,0:. Lhe reglme. 8uL Lhese are Lhlngs. 0,: Lo lmprove. Ck. lL's all. ,0: uone? 0,: uone.