Professional Documents
Culture Documents
ForthisassignmentwedecidedtoobservemetrousersattheGalleryPlaceMetro
Stationaround6:30pmonThursdayFebruary11th.WechoseGalleryPlacebecausewe
knewitwasastationthatreceivesalotoftraffic;wechosethatdateandtimebecause
weknewtheVerizonCenterwashostingaDisneyonIceShowthatevening.Webelieved
theshowwouldbringmanynoviceuserswhoseinteractionswiththemachineswould
provideusinsights.Wewereatthemetrostationforaboutanhourandobserved
roughly200passengers.Themajoritywereobservedusingtheturnstilebutwealso
observedabout20peopleusingtheticketmachines.Westayedtothesideofboth
machineswhileobserving(Figure1)andmadesuretoonlytakenoteswhentherewasa
lullintraffic.Analternativeapproachwecouldhavetakenisconductedsurveys.One
benefitofsurveysisthatwecanexplicitlygetuserslikesanddislikesaboutthemetro.
Onedrawbacktosurveysisthatusersmaynotaccuratelyremembertheirinteractions.
Figure1:PanoramicviewsofoneentranceatGalleryPlace.Top:viewoftheticketmachines.
Bottom:Viewoftheturnstiles.Redarrowsrepresentapproximatelocationsofwherewestoodwhile
observingmetrousers.Usersthatwerepresentinourpicturehavebeenblackedoutforprivacy.
IfthisstudywasproposedattheUniversityofMaryland,itwouldfallunder
exemptreview.Thereasonforthisclassificationisbecausethisresearchinvolvedthe
lowestamountofrisktoparticipants,collectednon-identifiableinformation,and
passivelyobservedpublicbehavior[1].Wedonotbelievethistypeofresearchrequires
informedconsentasthegoalofthisresearchwaspassiveobservationandtherefore
shouldhavenoimpactonusers.Additionally,tryingtogetinformedconsentinthis
situationnotonlycouldchangepeoplesbehaviorbutalsoseemsimpracticalduringrush
hour.Nevertheless,justbecausethisresearchfallsunderexemptreviewdoesnotmean
therearenoethicalconcerns.Aresearchermuststillprotectindividualsprivacy,notbe
intrusive,andnotforcepeopletoparticipateorputtheminhighrisksituations.
For this assignment we chose to observe two tasks. The first task was adding the
exactamountrequiredforasinglefareonaSmarTripcard.Wechosethistaskbecauseit
is a lot more complicated than adding a bulk amount and because DC visitors generally
add onlywhatisneeded.Thesecondtaskwechosewasobservinguserspassthroughthe
turnstiles using their SmarTrip card. We chose this task becauseitiscentraltousingthe
metro and because WMATA is phasing out the farecards for a SmarTrip only system.
1
Figure2:StandardfaremachineusedtoaddmoneytoaSmarTripcard.Thekeycomponents
include:(A)FareTables,(B)TerminalPrompt,(C)Addvaluebuttons,and(D)SmarTripreader.
2. Onceatthemachine,examinethefarevendingtablesatthetopofthemachine
(Figure3).Determinewhetheryouaretravelingonoroffpeak(TableA)andthen
findyourstationfromthealphabeticallistinTableB.Combinethosetwopiecesof
informationtodeterminetheamountyouneedtoaddonyourSmarTrip.
Figure3:Tablesusedtocalculatefare.TableAinformsusersaboutpeakandoffpeakhourswhile
TableBinformsusersaboutpricesforone-wayfarestoindividualstations.Ausermustcombine
bothpiecesofinformationtodeterminehowmuchmoneytoadd.
3.
4.
5.
6.
TaptheSmarTripcardtothecircularSmarTripreadertobeginthetransaction.
PressBontheterminalprompttoaddvalue.
EitherinsertcashdirectlyorpressB/Ctopaybycreditordebitcard.
Ifpayingbycash,keepaddingmoneyuntiltheexactamountisreached.Then
skiptostep10.Ifpayingbycard,gotostep7.
7. Clickthe+/-$andbuttonslocatedbelowtheprompttochangetheamountof
moneyaddedtothecard.
8. Oncetheamountiscorrect,insertcardintothecardreader.
9. PressB[Yes]orC[No]dependingonwhetheryouwouldlikeareceipt.
10. RetaptheSmarTripcardonthereadertoconfirmthetransaction.
2
Challenges,Errors,andSuccesses
LowSeverityIssues
Wenoticedtwolowseverityissuesthatincreaseduserstimeandfrustration.The
firstissuewasalackofabackbuttonontheprompt:onceauserstartedanaction,he
eitherhadtocompleteitorcanceltheentiretransaction.Asyoucanimagine,userswere
annoyedwhentheyrealizedtheirerrorforcedthemtostartagainfromthetop.Asecond
issuewenoticedwasthedesignoftheaddvaluebuttons(Figure4).Theseoptionswere
(+/-$)and(+/-)withthe$buttonaddingorsubtracting$1andthebuttonaddingor
subtracting5.Becauseofthisdesign,goingfromthedefault$10toanamountlike$3.50
wouldtakesix$pressesand10presses--anobnoxiousamountofbuttonpressesfor
anyuser.Aswiththelackofacancelbutton,thisdesignagainincreasedusertime.
Figure4:Mihaidemonstratinghowtousetheaddvaluebuttons.The$buttonaddsordeducts$1
whilethebuttonaddsordeducts5.Becauseofthemanualnatureofthesebuttons,ausercould
beforcedtobepressthesebuttonsalargenumberoftimesbeforetheyreachtherightamount.
HighSeverityIssues
Wealsoobservedanumberofhighseverityissues.Themostcommonissuewe
observedwasthatmanyusershadanextremelyhardtimefiguringouthowmuch
moneytheyhadtoaddtotheircards.Althoughthemachinedoesgivepeakhoursand
fareinformation,itdoessoacrosstwotablesintinyfont.Evenworse,itforcesusersto
understandthetwotables,combinebothpiecesofinformation,andthensavethat
informationintheirmemorywhenaddingmoney.Giventheattentionrequired,the
potentialforputtingthewrongamountisveryhigh;manyuserseithertookalongtime
tofigureitoutorjustgaveupandwenttotalktothestationmanager.
Relatedtothefirstissue,wealsonoticedthatsomeusersattemptedtousethelist
ofstationsasameanstoreorientthemselves(reorientationproblem)withinthemetro.
Unfortunatelygiventhatthelistofstationsisinalphabeticalorder,usingthelistasade
factomapisaveryeasywaytomakeaverybigmistake.Ifauseristocorrectlyreorient
herself,shemustgotoamaplocatedelsewhere.
Finally,wealsonoticedthatseveralindividualswereconfusedastowhich
machinetouse.Asitturnsout,thereweretwomachinespresentatthestation:the
standardmachineandthefarecardtoSmarTripmachine(Figure5).Whilethestandard
machinehadexpectedfunctionality,thefarecardmachinehadalimitedterminal
promptandeitheracceptedcashortookinafarecardandtransferredtheamounttoa
SmarTrip.However,becausebothmachineshadaSmarTripreader,someuserswent
directlytothefarcardmachineandendedupgettingextremelyconfusedbythelimited
interface.Itcertainlydidnothelpmattersthattheinstructionsonthemachinewere
aboutbuyingafarecardandnotSmarTrip!
Figure5:Comparisonbetweenfarecardconvertermachine(left)andstandardmachine(right).Both
machineshaveaSmarTripreaderandbothcanbeusedtoaddmoneytoaSmarTrip.However,the
machineontheleftnotonlyhaslimitedfunctionality(e.g.onlyacceptscash;hasalimitedterminal
prompt)butthemaininstructions(highlightedinred)havenothingtodowithSmarTripcards!
RecommendationsForDesign
Figure6showsourrecommendationsforredesignoftheticketmachine.
Figure6:Ourredesignedmachinewithaminimalistapproach.Informationthatwaspresentonthe
previousmachines(seeFigure5)hasbeenreplacedbyalargetouchscreen.Allinstructionsare
withinthetouchscreen.Paymentisnowallononeside(insteadofcashonleftandcardonright)
andtheSmarTripreaderisnowaslot.
WhencomparingourdesigninFigure6withthecurrentmachinesinFigure5,
obviousdifferencesimmediatelyjumpout.Unlikethecurrentmachines,oursis
minimalistindesignwiththemajorcomponentbeingthetouchscreen.Allofthe
previousclutteredtext,tables,instructions,andcolorshavebeenremovedfromthe
outsidesotheuserisnotbombardedwithinformationbutinsteadknowsexactlywhere
tolooktobegin.Payment(eithercashorcard)havebeenplacedonthesamesidefor
balance.Finally,theSmarTripreaderhasnowturnedintoaslotthatusersinserttheir
cardinto.Thischangeiscriticalasitpreventsusersfromforgettingtotapouttheir
card--anerrorthatisverypossibleinthecurrentsystem.
Ofcourse,withatouchscreencomesthepossibilityofnewinterfaces.Figure7
illustratessomeofthenewinterfacesthatwouldbeavailabletousers.
Figure7:InterfacesavailabletouserswhileaddingmoneytotheirSmarTripcard.(A)Afterinserting
theirSmarTrip,userswillseetheircurrentbalanceandbeallowedtoaddvalue,pickadestination,
orendtransaction.(B)IfAddvalueisselectedin(A),userswillbetakentothisscreenwherethey
canquicklyaddtheamounttheywant.(C)IfSingleDestinationwasselected,userswillbetakento
amapwheretheycanselecttheirdestination.Mapallowsforzoomandslides.
PanelsA-CinFigure7showthedigitalinterfacesauserwouldinteractwithafter
theyinsertedtheirSmarTripcard.PanelAshowsthemainoptionsavailabletotheuser;
now,alloftheoptionsavailableinthecurrentmachinearepresentedinaclearfashion.
Ifaddvalueisselected,usersaretakentoPanelBwheretheycanaddanyamountthey
wish.Moreimportantly,however,themechanicaladdvaluebuttonshavebeenreplaced
byacleaninterfacethatwillsaveuserstime.PanelCshowshowwehavecompletely
redesignedtheSmarTripsinglefareexperience.Ratherthanforcetheusertolookat
tablestodeterminehowmuchmoneytoput,theuserwillbepresentedwiththemetro
mapandbeaskedtoselectherdestination.Onceastationhasbeenselected,the
machinewillautomaticallycalculatetheusersfare.Byreducinguserscognitiveload,
thisdesignshouldspeeduptransactionsanddecreasefrustrationanderror.The
presenceofamapalsosolvesthereorientationproblemdescribedpreviously.By
presentingthemapatthemachine,userswillbeabletoknowexactlywheretheyarein
thesystemwithouthavingtoleavethemachinetosearchforamap.Finally,allscreens
haveabackandcancelbuttontomakeitsimpleforuserstoundotheirmistakesatany
pointintheprocess.
5
Task2:Enter/ExittheMetroStationviaSmarTrip
Figure8:Thestandardturnstilesatanymetrostation.(A)Thegreenarrowdenoteswhetherornota
turnstileisforenteringorexiting.ThecardreaderontopiswherepassengerstaptheirSmarTrip
card.(B)Theredarrowillustratesaturnstilethatapassengercannotenterthrough.
Theprocedureforenteringandexitingametrostationisidenticalineverythingbutthe
directiontheusermustbewalking(eithertowardstheexitorintothemetrostation).
1. Walktowardsaturnstilethatcontainsagreenarrow(Figure8A);avoidturnstiles
withtheredsymbol(Figure8B).
2. Ifsomeoneisaheadofyou,waitfortheturnstiletoclosebehindhimorherbefore
proceeding.
3. Whenyoureachtheturnstile,slideyourSmarTripoverthereader:
a. Ifeverythingiscorrect,thedisplaywillshowyourSmarTripvalueaswell
asyourtripcostifyouareexiting(Figure9A);thegateswillopenandyou
canproceedforward.
b. Ifthereisanerror,thedisplaywillshowanerrormessageandyoumust
gotalktothestationmanager(Figure9B).
Figure9:(A)IfeverythingiscorrectwhentheSmarTripisread,theterminalwilldisplayauser's
SmarTripvalue(andtripcostifexiting)asthegatesopen.(B)Ifanerroroccurs,theterminalwill
telltheusertogotalktothestationmanager;thegateswillnotopen.
Challenges,Errors,andSuccesses
LowSeverityIssues
Weidentifiedthreelowseveritydesignflawswiththecurrentturnstilesystemat
GalleryPlace.Thefirstissuewenoticedwasthatusersdidnotknowwhentoscantheir
cardwhenanotherpersonprecedethemattheturnstile.Becausetheturnstileshadno
Readysignal,someusershesitatedacoupleofsecondsbeforescanning.Thesecond
flawwenoticedwastheawkwardnessofuseraccessories.Becausesscannersarelocated
ontherighthandside,userscarryingalargeaccessory(e.g.luggage)withtheirhand
(especiallyrighthand)hadaveryhardtimegettingthroughtheturnstileinonesmooth
motion.Thethirdflawwenoticedwasthatusershadtolookdowntoviewhowmuch
moneytheyhadleftontheirSmarTripafterscanningit(Figure10).Becausethedisplay
isinanon-ergonomiclocation,theactoflookingdownwasnotonlyawkwardfor
passengersbutalsostoppedtheminmotion.Unfortunately,allthreeoftheseflaws
disruptedtheconstantflowofpassengersthroughtheturnstilesystemandcontributed
tolonglinesathightrafficmoments.
Figure10:MihailookingdowntocheckthebalanceonhisSmarTripcard.Theplacementofthe
displayforcespassengerstoturnandlookdowninordertochecktheirbalance,slowingthemdown.
HighSeverityIssues
Wealsoidentifiedtwohighseverityissueswiththeturnstilesystem.Thefirstflaw
wenoticedwasuserinconsistencywhentravellingwithchildren.Becausetherewasno
clearindicationwhetherornotchildrenshouldbetreatedasnormalpassengers,parents
displayedawidevarietyofactions.Someparentshadacardforeachchild,others
crammedalloftheirchildrenthroughtheturnstileononeswipe,andstillothershad
theirchildrencrawlunderneaththegate.Thelackofclearinformationregarding
childrenmakesitsignificantlyharderforfamiliestousetheturnstilesandlosesmetro
money.Thesecondmajorflawwenoticedwasthatsomeusersdidnotunderstand
metrostap-in,tap-outsystem.Onmorethanoneoccasion,wesawuserstryingtopass
theirSmarTripcardbacktotheirfriendsstuckbehindthegateoruserstryingtoscana
numberofdifferentcardstoopenthegate.Unfortunately,neitherofthesemethods
worked.Theconsequenceofthisflawwasthatmanyusers--especiallynoviceusers--had
anincrediblyhardtimeusingtheturnstiles,wastedtheirtime,andhadtoaskforhelp.
7
RecommendationsForDesign
Figure11showsourfirstredesignoftheturnstilemachine.
Figure11:Ourredesignsolvestheproblemsofpoorerrormessages,uncertaintyaboutwhentoproceed,
anduserscarryingitems.(A)Lightbarsonthefrontofthegategloweithergreenorredtolettheuser
knowtheycanproceedforwardtoscantheircard.(B)Scannersonbothsidestoallowuserstouseeither
hand(C)Displayswillnowbepositionedatthetoptoallowuserstolookstraightaheadastheyscantheir
card.
Withourfirstdesign,wewantedtosolvetheproblemsofpoordisplayfeatures,
usershesitatingbeforeproceedingthroughtheturnstilewhenfollowinganother
passenger,andtheawkwardnessofauseraccessory.Toachievethisgoal,wemadethe
followingchanges:first,weaddedglowinggreenandredbars(Figure10A)tomakeit
easierforuserstoknowwhentheyshouldproceedforwardtoscantheircard.Second,to
makeiteasierforleft-handedusersanduserscarryingalargeitem,weputscannerson
bothsideoftheturnstile(Figure10B);nowuserscanuseeitherhandtoscantheircard
insteadofhavingtoreachovertheirbodyorshufflewhattheyareholding.Finally,we
movedthedisplaytothetopoftheturnstiletoallowuserstolookstraightaheadwhen
scanningtheircard(Figure10C),allowinguserstocontinuewalkingwhilereadingtheir
card.Displayswillalsocontaindetailedmessagestoletpassengersknowexactlywhatis
wrongshouldanerroroccur.Webelieveallofthesechangesshouldmaketheprocessof
goingthroughtheturnstilesignificantlyeasierfortheuser.
Figure12showsoursecondredesignoftheturnstilemachine:
Figure12:ThisredesignremovestheneedforusingtheSmarTripcardonexit.(A)Thenewexit
turnstileopensandclosesbasedoffamotionsensor.
Oursecondredesignisamuchmoreextensiveplan,targetingnotonlythe
turnstilesbutthepaymentsystemaswell.Asdescribedabove,oneofthemajorissuesis
usersforgettingthatthemetrohasatap-in,tap-outsystem:ausermustusethesame
SmarTripcardtoenterandexitthemetro.Thisredesignattemptstoremovethis
problembyshiftingthesystemtosinglepaymentstructure.Specifically,weproposethat
userswouldonlyhavetoscantheirSmarTripcardwhentheyenterthemetro;they
wouldnolongerhavetoscantheircardwhentheyexit.
Toimplementthisideaacoupleofthingsmustchange.First,becauseusersno
longerhavetotapouttoexit,thesystemofchargingusersbasedondistancemustbe
replacedbyasingleonetimefeeattheentrance.Oneimmediatebenefitofthisideais
thatitsavespassengersthehassleofcalculatinghowmuchmoneytheyneedbecausethe
amountisalwaysconstant.Besidesupdatingthefaresystem,oursecondredesignalso
requireschangestotheturnstiles.Specifically,ourrecommendationrequiresturnstiles
tobesegregatedintoanentrancegroupandanexitgroup.Theentranceturnstilescan
stillbetheturnstilescurrentlyinuseortheycouldbereplacedbythesuperiorturnstiles
weproposedinFigure11.Ontheotherhand,theturnstilesusedforexitwouldbe
replacedbytheonesinFigure12.Becauseusersnolongerhavetotap-outonexit,the
onlythingrequiredontheseturnstileswouldbeamotionsensortoallowuserstopass
through.Topreventpeoplefromtryingtocheatthesystem,themotionsensorwould
onlybeontheexitsideoftheturnstile;thegatewouldnotopentouserscomingfromthe
entrance.Moreimportantly,thelackoftap-outwouldsaveusersanenormousamount
oftimeonexit.Inconclusion,webelievethatbothofourturnstilerecommendations
wouldalleviateanumberofuserproblemsandmakethesystemaloteasiertouse.
ConcludingRemarks
Fromdoingthisassignment,SeanandIlearnedthatsystemswehaveno
complaintsaboutactuallyhaveplentyofroomforimprovement.Onethingwedidnot
likeaboutthisassignmentwasobservingpeopleusingtheturnstiles;wefoundthat
observationalotharderthanobservingpeopleusingthemachine.Finally,onewayto
improvetheassignmentistointerviewstationmanagersabouttheirexperiences;they
probablyhavealotofgoodinsightsthatcouldhaveimprovedourdesigns!
Citations
[1]IRBProcess.(2016).RetrievedFebruary16,2016,fromUniversityofMarylandResearch
ComplianceOfficewebsite:http://www.umresearch.umd.edu/RCO/New/IRBProcess.html