You are on page 1of 45

Software Intelligence for Digital Leaders

Software Intelligence as a Service


Communicate – Decide – Measure – Protect – Discover – Improve

Sriram Srinivasan
VP & Global Client Partner

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Typical Service Management Challenges

Application Health
• What is the overall health of my portfolio? How is it trending – over time and by release?
• What is its structural integrity? Is it compliant with enterprise architecture and quality standards?
• How stable will my next release likely be?
• How secure is my application portfolio? Is it vulnerable to hacking?

Support & Development


• Can I benchmark my application portfolio with the industry?
• How do I quantify the value my application services partners are bringing to me?
• Are tickets simply being ‘closed’ or is the root cause also being addressed?
• Can you help me build a roadmap for application transformation?

Confidential 2 SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Why do I need CAST?
“I need to compare my “I want the same output, “I want a consistent “I want to measure
suppliers against industry but I need to do it with assessment of code quality, productivity, quality,
benchmarks, to ensure the fewer resources, to match and to find out potential innovation, development
best value” shrinking budgets.” problem areas – before effort and agility, so we
deployment.” are Best-in-Class.”

Vendor Management CFO QA CIO

“Through 2016, despite being a top priority of CIOs, less than 15% of
vendor management programs will leverage data analytics and business
intelligence to identify and manage vendor risk”
- Software Analytics & ADM Vendor Risk Management

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Significance of System-Level Structural Flaws

90%
SYSTEM LEVEL FLAWS

8% 48% Downtime caused


Of total by system-level flaws!
repair 8% of defects account for
effort
48% of repair effort and 90%
of downtime
92% 52%
of all of total 10%
defects repair
effort UNIT LEVEL FLAWS

Customer satisfaction impact 65%


Time to market impact 47%
6x
Cost to Fix

Shift Left
3x
2x
1x
Source: Economics of Software Quality – Jones & Bonsignour; Software
Defect Analysis: An Empirical Study of Causes and Costs in the IT Industry –
Kristiansen, NTNU; CAST Report on Application Software Health (CRASH
2012) – Curtis, Sappidi & Szynkarski

Confidential 4 SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Defining Software Intelligence

Software Intelligence is insight into complex software structure.


It consists of…

Software Software Software Software


Health Size Flaws Blueprints

Robustness, Efficiency, Security, Changeability, Functional and Engineering patterns that Map and visualize
Transferability, Complexity risks and Cloud technical sizing to assess create potential outages, security breaches, and software structure
Readiness quantity of work data corruption and architecture

Software Benchmarks
Accurate, consistent analytics by industry and geo, normalized by functional size, based on
structural data from thousands of software systems

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


What can you do with Software Intelligence?

Visualize ADM Business Application Assessment


One platform to run your ADM business For wise decisions or troubleshooting

Improve ADM Quality Discovery & Blueprinting


Deliver super software Legacy applications at light speed

Drive ADM Efficiency Reduce Costs


Boost team performance ADM run rate or MIPS

Supply Chain Transparency CISQ Audit


Get the most from partnerships Verify compliance to industry norms

Intercept Software & Security Flaws Application Benchmarking


Find and fix software monsters Know your software competitive advantage

Accelerate Modernization M&A Due Diligence


Master the transformation Intelligent due diligence and safer integration

Cloud Readiness and Transformation Protect Data


Spot boosters and blockers Comply with GDPR and privacy regulations

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


CAST ecosystem

CAST APPLICATION INTELLIGENCE PLATFORM

CAST Highlight Health Dashboard Engineering Dashboard CAST Imaging CAST Appmarq

• Portfolio health measures • Software health measures • Software flaw identification • Architecture analysis • Industry, technology & geography
benchmarks
• Potential risk • Trend analysis • Security and risk identification • Blueprinting
• +2500 apps, +3B LOC
• Savings opportunities • Flaws and technical debt • Architecture governance • Impact analysis
• Annual industry reports
• Cloud-ready assessment • Software sizing • Engineering best practices • Application discovery
• Custom benchmarking
• Industry benchmarks • Application security • Remediation planning

Software Intelligence as a Service


Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS
CAST solutions answer two common management questions

CAST Highlight at Portfolio Level


• For Due Diligence, IT Rationalization, Portfolio Assessment in Simplify IT
Where should I look? 1 • Rapid Business/IT Risk/Paas Readiness combined Apps portfolios (50 apps/ week)

Now I know where to look

CAST Application Intelligence Platform on targeted applications


• For critical apps: risk prevention, architecture and productivity improvement
• Deep X-RAY on targeted critical IT applications (one app analyzed in ~one week)
Where should I start 2

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


CAST Highlight: Fast health check of a portfolio's source code

CAST Highlight identifies software risks, complexity and cloud


readiness in the code:
Excellent CxO talking point

• Fast
– A scan for a normal sized application1 takes 5 min
– Up to 50 apps portfolio1 can be analyzed in a week
– Does not require deep technical knowledge
Objective metrics deliver visibility
into portfolio health • Secure
– All source-code is retained in house
– Analysis output is encrypted
• Non-disruptive
– Simply requires 10 – 15 minutes per application
• Flexible
– Common programming languages2 are covered
– CloudReady to assess PaaS posture
1. Normal sized application is an application with around 350k lines of code, a portfolio consists of 100 applications
2. Languages covered are: ABAP, C, C++,C#, Cobol, Java, JavaScript, Ksh, Nsdk Objective C, PHP, PL/SQL, PL1, TSQL, VB, VB6
Source: http://casthighlight.com/
Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS
Leveraging CAST Highlight improves throughput 100X

Traditional Methods
• Due Diligence or M&A
− Based on SME interviews and
ticket logs: may not provide the
full/ accurate picture • Manual  Automated CAST Highlight
− No code reviews − 1 application: 3 – 5 minutes
• Partial  Comprehensive
− Typically, 1 application takes ½ − Full set of portfolio metrics
- 1 person-day • Subjective  Substantiated
available in 1 – 2 weeks1
• Cloud Readiness Assessment
− Based on SME interviews and
partial, manual code reviews
(usually 20 – 25% of code base)
− Typically, 1 application: 2 – 3
person-days

100 applications
1

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


CAST AIP: Consistent and transparent means to measure performance
and drive innovation across the entire portfolio

1. Industry standard analytics, metrics and benchmarks for clear, unambiguous


measurements

2. Quantify operations risk, maintainability, performance and trending analyses


for proactive assessment and remediation

3. Quality targets that can be matched up to deliverables, to drive a true lower


cost approach

4. Methodology and platform independent

5. Specific guidance on what to fix, how to fix and, validation of has it been
fixed

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


What results do CAST’s clients achieve?
21% Reduction in application development costs “Communicate with clarity.”
50% Improvement in application quality Frederic Veron,
25% Increase in monthly deployments SVP IT, Fannie Mae.
Promoted to CIO
56% Defect Reduction over 4 years
60% Improvement to delivery time “Trade off speed & risk.”
20% Reduction in ADM spend over 3 years Thaddeus Arroyo,
CIO, AT&T.
Now CEO AT&T Business
10% Run Cost Reduction (break fix, tickets,
patches and runtime operation)
“Transparent supplier relationships.”
60% Reduction in developer discovery time Lester Thomas, Head of System Architecture, Vodafone,
3X Acceleration of new release time to market Promoted to Chief IT Systems Architect

12% Reduction in maintenance cost


3X Reduction in defect density “It keeps the business safe!”
Stephen Sepulveda, VP, Application Performance
25% Improvement to MTTR Management, BNY Mellon
Promoted to VP APM
20% Reduction in mainframe cycles
$10,000/month savings “Insight like never before.”
Michael Synn, VP of Enterprise Architecture,
10% Improvement of developer productivity
Broadridge,.
20% Reduction in defects
Promoted to CTO

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Proven use cases to gain foothold or door-opener
• Automated application portfolio assessment in days or weeks
− SaaS model, source artefacts do not leave the client network
− Cloud readiness
− Benchmark applications with industry peers
− Digital transformation roadmap

• Due diligence for M&A or CIO advisory


− What to keep, what to discard, what to repair
− Technical debt

• Software Intelligence as a Service, using Highlight and AIP


− Proactive remediation vs. reactive maintenance
− Where are my ADM $ going?
− Trend analyses
− Quality gates
o Architecture compliance
o Development to Maintenance
o Before production move

Confidential 13 SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Intelligence and Metrics that support real issues
• Operations Risk :The stability and predictability of
software operations based on factors that have been
proven to be issues in operations.

• Maintainability: The ease that the code can be


maintained, adapted and adjusted. Impacts
development costs and project estimation, over-runs
and resource consumption.

• Supported by volume and technical size


measurements

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


14
How clients are integrating quality metrics
Four basic patterns of structural metrics in Vendor Management across CAST customer base

1 Scorecards 2 Policy 3 SLAs 4 Acceptance Criteria

Measurement and discussion Ask vendors to provide specified Treat software enhancements and Measure and demand
in governance committees metrics on their own (e.g., provide maintenance as a service; track minimal set of acceptance
goes a long way to setting CISQ scores/certificate for each levels, penalties, credits but criteria for any new
behavior if given enough release) …… which can be extended through CAST, to cover development or release
attention unreliable and somewhat operational and maintenance metrics
inconsistent
Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS
15
Supported by standard metrics and reports
1 Scorecards

CAST provides an overall “portfolio” Extracted from Client


view

− Selectable by Vendor/ Subcontractor


− Common SLAs and trends
− Industry Standard metrics
− Online and in standard reporting format

… as well as an “application” view


– Common SLAs and trends
– Industry Standard metrics
– Focus on actionable plans
– Integration with JIRA and other ticket management
systems

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


16
CAST products provide a consistent and transparent means to measure performance
and drive innovation across the entire portfolio

1. Industry standard analytics, metrics and benchmarks for clear, unambiguous


measurements

2. Quantify operations risk, maintainability, performance and trending analyses for proactive
assessment and remediation

3. Quality targets that can be matched up to deliverables, to drive a true lower cost approach

4. Methodology and platform independent

5. Specific guidance on what to fix, how to fix and, validation of has it been fixed

Confidential 17 SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


17
Illustrative Examples

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Analyze commonly used technologies

Web IBM SQL -


VB .NET J2EE C/C++
Languages PSM
“ CAST tells us why to invest in one
Microsoft area versus an other, something the
Microsoft
EGL Visual Tibco
T - SQL business can see and better appreciate.”
Basic
- F. Veron. SVP, Head of Application
Oracle Sybase T - Development
PHP PL/SQL Mainframe C# SQL

SQL – PSM
DB2 Business
Universal Fortran PL1 Objects
Database
Oracle
Power ABAP
Flex Forms/ RPG Builder Objects
Reports
“To drive behavior and make the right
PeopleSoft Siebel Shell ABAP tradeoffs between speed, business risk and
technical debt.”
- Thaddeus Arroyo, CIO
HTML5 JQuery Angular JS Node.js JavaScript

Confidential 19 SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Industry Benchmark Results (Appmarq)

Confidential 20 SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


CAST-generated interactive application technical documentation
What are all the artifacts linked to a From GUI to Database, what are What artifacts are included in the API
complex object? the artifacts involved in a module and how do they interact with
transaction? Applications A & B?

View of a database structure and


related data access modules
View of all artifacts and components used
for a GUI screen

Confidential 21 SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Architecture Checker - Validate against desired model

• Check the architecture model and the application's compliance, interactively


• Generate and enforce quality rules

Confidential 22 SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Transaction Risk Index: Highlights the riskiest transaction
streams/ flows

UI Layer
Logic
Data Layer

Riskiest transaction streams with high number of Robustness, Performance or Security violations

Confidential 23 SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Propagated Risk Index: Impact of violations across the portfolio

UI Layer
Logic
Data Layer

Violations with the largest impact on the rest of the portfolio, on Maintainability, Robustness, Performance or Security

Confidential 24 SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Transition use case (17 – 20% benefits)
Traditional transition models rely on
‘guestimates’ and reactive discovery; not
reflective of required effort
 Accelerate, optimize, de-risk transition
 Identify application hot-spots, problem areas and
structural vulnerabilities up front
Ticket data is not always an indicator of
application quality, complexity, performance or  Automate knowledge discovery and
stability
documentation
 Focus on application pain points and high-risk
Knowledge transfer is a top-down approach; areas rather than a boiler plate approach
critically dependent on outgoing SMEs for
success  Reduce learning curve
 Reduce transition cost and team induction effort
Variations and knowledge gaps are impossible to
justify and always impact profitability ,
customer satisfaction

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Continuous Improvement and Innovation use case
What is the basis for productivity benefits?
Can they be substantiated?
 10% Run-the-Business cost reduction
How do I baseline my portfolio so I can measure  7% - 10% of installed codebase: Identify dead code
continuous improvement and value realization
over time?  10% Improvement in developer productivity
 60% Improvement in delivery time

How can I demonstrate innovation?  14% - 32% Reduced system technical debt

 50% - 70% Reduction in defects


How do I improve delivery efficiency, increase  30% Reduction in production defects
profit, customer satisfaction?
 100x Reduction in Function Point counting effort

How does my performance compare in the


market place and in the industry?

How can I leverage structural analytics to


differentiate from competition?

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Client case study: CAST benefits during transformation

Aggregated enterprise-level metrics


• Structural Quality improved 38% (on average)
• Productivity improved 28%: reduced release
cycle time
• Total Quality (TQI) improvement
− 2015 baseline: 46th percentile (Yellow/
Red)
− Presently: 70th percentile (mostly Green)

• Periodic scans resulted in 24% improvement in


TQI
Critical violations (Defects)

Application changes per release

Source: Major financial services provider, 2015 - 2017

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


ADM organization score carding
Major consumer services institution CAST Context:
• Three strategic ADM partners, $2B spend • Used for quality measurement and vendor management at four
major IT towers
• 15,000 FTEs in ADM • 300 applications, ~1 trillion LOC analyzed
• IT viewed as important business enabler • Effort data collected from custom PPM tool

CAST
CASTQuality
Quality Defect
DefectDensity
Density Mean
MeanTime
Timeto
to Repair
Repair
Technical
TechnicalCode
CodeQuality
Quality Quality
Quality Quality
Quality

Average
AverageTQI
TQI
Pre
PreProduction
ProductionDefect
February
DefectDensity
February2010
Density
2010––June
June2012
2012
Pre
PreProduction
ProductionMean
MeanTime
February
Timeto
February2010
toRepair
Repair
2010––June
June2012
2012
• Benchmark development team
February
February2010
2010––June
June2012
2012
Best
Bestin
Class
in
2011.02
2011.02 2011.06
2011.06 2011.02
2011.02 2011.06
2011.06 productivity, quality and hit rates
3.31
3.31 Class 0.86
0.86 30.18
30.18
3.01
3.01 3.02
3.02 2.96
2.77 2.96 Good 0.70
Good
• Measure against other towers, and trending
2.77 2.66 0.70
2.66
Average
Average
16.75
16.75 14.32 14.30

over time
12.09 14.32
11.80 14.30
0.23 12.09 11.80
0.23 0.18 0.17
0.17
0.18
0.06 0.07 4.29
0.06 0.07 ---- 2.85 4.29
2.85
Grp
Grp11 Grp
Grp22 Grp
Grp33 Grp
Grp44 Grp
Grp55 Grp
Grp66

Release hit rate up by 10%


Grp
Grp11 Grp
Grp22 Grp
Grp33 Grp
Grp44

Grp
Grp11 Grp
Grp22 Grp
Grp33 Grp
Grp44

Release
ReleaseHit
HitRate
Rate Productivity
Productivity Productivity
Productivity
TThroughput
hroughput Cost
CostEffectiveness
Effectiveness Cost
CostEffectiveness
Effectiveness

February
Release
ReleaseData
February2010
Data
2010––June
June2012
2012
Cost
Costper
perFunction
FunctionPoint
Point||Enhancement
February
Enhancement
February2010
2010––June
June2012
2012
Cost
Costper
perFunction
FunctionPoint
Point||Maintained
February
February2010
Maintained
2010––June
June2012
2012
• Maintenance cost down by 5 –12%
Feb
Feb11 June
June11 Q1
11 11 2011.02 2011.06 Q12011
2011 Q2
Q22011
2011
87%
87% 2011.02 2011.06

• Defect densities down by 3x


85% $95
$95
84% 85% $3,320
$3,320
84% $3,132
$3,132 $78
$78

76% $2,603
$2,603 $58
$58 $55
75% 76% $1,989 $1,901 $2,193 $55
74% 75% $1,989 $1,901 $2,193
74%

$957 $1238
$957 $1238
$9
$9 $8
$8 $6 $6
• MTTR down by >25%
$6 $6
Grp
Grp11 Grp
Grp22 Grp
Grp33
Grp
Grp11 Grp
Grp22 Grp
Grp33 Grp
Grp44 Grp
Grp11 Grp
Grp22 Grp
Grp33 Grp
Grp44

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Measured decrease in rework across 350 systems

Ticket Volume Average Fix Effort


4 4
3.5 3.5
3 3
2.5 2.5
2 2
1.5 1.5
1 1
0.5 0.5
0 0
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

Log of Sum of Tickets Log of Average Effort


Average TQI Average TQI

n = 350 applications; 60 million LOC

High degree of correlation between CAST TQI (Total Quality Index) and fix/ maintenance effort

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Dramatic drop in break-fix as Maintainability increases
Study of maintenance effort across 20 customers shows tight correlation between maintenance fix
tickets and CAST Maintainability

Increase of CAST Maintainability Index (*) by 0.24 = decrease in maintenance activity by 50% thus
improving margin ratio for relevant Maintenance contracts.
(*) Aggregate of Transferability Index and Changeability Index

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


AM: Media and entertainment – Continuous improvement

Context/ Objective
• Sony Pictures Entertainment, Inc. (SPE) is the television and film production/distribution unit of
Japanese multinational technology and media conglomerate Sony.
• Performance and Stability Issues observed on business critical app used across 23 territories. The
app was constantly being enhanced since going live in 2003 and had become very complex with
1.5+ million lines of code.

Scores Color Risk level

1 to 1.99 Red Very high risk

Solution 2 to 2.99

3 to 3.99
Orange

Yellow
High risk

Moderate risk

The Analysis was guided by frequent slow response issues on online booking screen in the J2EE 4 Green Low risk
WebLogic solution
• Code-violations along entire Transaction flow (JSP-Controller-Business Object-Hibernate Call-
Stored Proc) were detected using CAST.
• 972 violations across 15+ modules were identified to be fixed in a phased manner.

Benefits
• CAST Analysis saved tens of hours of manual investigation effort.
• Systematic Code Remediation resulted in appreciable reduction in slow response issues.

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Transition and Due Diligence – Reduction in maintenance cost

Cost reduction program focused on Run costs, including break fix, tickets, patches and
runtime operation

Results of Cost Reduction Program

Applications managed w/o CAST


↓10% - 26 systems
- Cost reduction of 8% from baseline

Applications managed with CAST


- 4 systems
- Additional 10% cost reduction

n = 30 applications ; 26 million LOC

The teams using CAST for a year improved their cost profile by an additional 10%

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Improve ADM productivity by >8% with Shift Left
Structural Quality Drives 30% of non functional Driving the Right ADM DeliveryTeam Behavioural with measurement and continuous
defects improvement reduces ~20% of the non functional defects

Drastic reduction in newly introduced code level defects

Defect Origin % Severity 1 or 2


Defects
1. Design defects 17.00%
2. Code defects 15.00%
3. Structural defects 13.00%
4. Data defects 11.00%
5. Requirements creep defects 10.00%
6. Requirements defects 9.00%
Correlation btw code quality and production defects:.
7. Web site defects 8.00%
8. Security defects 7.00%
9. Bad fix defects 4.00%
10. Test case defects 2.00%
11. Document defects 2.00%
12. Architecture Defects 2.00%

TOTAL DEFECTS 100.00%


Severity 1 = total stoppage
Severity 2 = major disruption

Source: Capers Jones. Data collected from 1984 through 2014 Source: Major Telecommunication Operator, from 2009 to 2014. 900 developers working on 4 mission critical
About 675 companies and 13,500 total projects; from 24 countries. systems under active development & maintenance

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Structural Quality Gate: Find defects earlier in the SDLC (Shift Left)
Quality must be designed in, not tested in Empirical evidence at a major retail bank
- an emerging realization in the IT industry

150X
Relative Cost of Fixing
Defects

50X

20X
5X 10X
1X
Req. Design Code Dev. Accept. Prod.
Testing Testing

Using CAST to detect errors early in the development


process reduces cost by
• Avoiding rework (catching defects after development)
• Streamlining the QA processes (more efficient and
comprehensive
code reviews; less work for QA/QC team)

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


AM: Health, pension and insurance company
Health, pension and life insurance company Business Challenges
• €16.6B managed, 9M people insured • €2B of annual revenue supported by the main app of the
“savings” portfolio
• Complex and changing portfolio 
• 200 partners connected to the app to sell the savings offer
• 1MLOC, 65,000 person-days of development since 1998
• CEO/ CIO/ Business questioning maintenance costs and system
• 1500 person-days of evolution per year
reliability

13%
• CAST AIP implemented
as part of the SDLC Reduced Maintenance Budget by
54%
• Application audits performed Total Incidents Down by
at each major release 70%

Code Related Incidentis Down by


• CAST Health Factors integrated 45%
in SLA with service centers Number of New Tickets Down by
88%
• CAST analysis extended
Unfixed Defects After Rework Down by
to “pensions” and “health” applications

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


AM: Insurance company
Major Insurance Institution CAST Context:
• Reduce maintenance costs • Statically analyze application quality and stabilize
• 700,000 LOC insurance management system maintainability score

• 10 million customers and processes • Provide reports to help development teams remediate problems
faster
• 8 million claims a year

Early Defect Detection


56% reduction in
defects in 4 years • Maintainability stabilized despite
40% increase in code over 4 years
Defects / KLOC

• Reduced delivery time by 60%


• Reduced costs 20% over 3 years

Year 1 Year 2 Year 3 Year 4

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


AM: Consumer finance company
Major consumer finance institution Business Challenges:

• Four strategic ADM partners • Reduce system outages, break / fix cost and production incidents due
to change requests
• Total of 5,000 FTEs in ADM
• CAST AIP implemented as part of an Application Quality Management
• Establish a measurement-driven culture for full collaborative initiative - deployed across 20 teams and over 2,000 applications
transparency

Production Outage Impact

• Early detection of technical defects


• Management visibility into ADM output
30% • Significant cost avoidance (downtime, break/
fix)

CAST detects 30% of software related


production issues.

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


AM: Telco
Global telecommunications company CAST Context
• Global initiative to enhance quality and Service • Reduced total number of violations from
(IT and Business) 385,000 to 11,000 in 12 months
• 900+ developers. 7 systems, on 12 months, • Identified a strong correlation between violation
4 major releases, 10 intermediate releases and production incidents

Critical Violation Impact Production Defect Impact

70% drop in
production
defects

11,000
385,000 575 165
Reduced Violations
“The reduction of CAST violations and the drop
Over in production incidents is no coincidence. Our own analysis shows that there is strong correlation
12 Months
that as you reduce violation density you experience less production outages.” - Vice President.

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


AD: Manufacturing and distribution
Manufacturing and Distribution CAST Context:
• Heavily customized SAP solution with over 1000+ ABAP • CAST AIP deployed during the build phase
custom objects developed every release. as part of transport order release management
• Goal to aid developers with programming best practices,
mitigating performance threats and optimizing memory
utilization

Maintenance Cost Performance Impact

System performance improvement


50% Before CAST
Fewer tickets per
quarter After CAST

20%
Faster

Support Support
Tickets Tickets

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


AM: Communications company
Manufacturer of broadcasting technology CAST Context:
• Experienced performance and stability issues in critical SAP • ABAP code quality was baselined across 7 modules
systems (MM, SD, CO, FI, PP, WM, CS)
• CAST identified areas of improvement in the code quality of complex
ABAP custom objects.
• Recommendations to improve performance, programming practices &
documentation
were identified and implemented

Overall code quality improved by 8% System Performance Impact

Original Batch Processing Time


Improved Batch Processing Time

30%

SAP Module Baseline Results by Health Factor

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


A CAST success story ……. Global SI
Started CAST as a program in 2015 with these objectives
− Consistent enterprise-wide deployment
− Measure delivery effectiveness and productivity
− Provide competitive advantage through risk-adjusted smarter solutions

82 60M 90 63% 10+ 2


Accounts covered LOC analyzed Major projects Managed services Case studies: Wins: Where
audited accounts impacted Innovation, CAST analytics
Retention, was a
Expansion differentiator

 68% of the applications inherited had a TQI < 3


• Post transition remediation reduces profitability of account and customer satisfaction
 Improving TQI from 2.5 to 3.05  50% reduction in defects
 Application security went up by 35%  increased end-user experience
 Robustness increased by 12%  Priority 1 tickets down to 0
 60% decrease in pre-production defects in Development projects

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Leverage CAST across the sales and delivery cycle

More Risk Cost Take Out Higher Growth &


Competitive Mitigation Productivity Expansion

Due diligence & Pricing √ √ √

Transition, KT & Newcomer


√ √ √
induction

Management Dashboard √ √ √

Ticket Reduction √ √ √

Early Detection & Prevention


√ √ √
of Defects

Benchmarking √ √

Speed to Response √ √ √

Differentiation & Innovation √ √

Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS


Why Application Quality has become essential
 Clients have become increasingly Project A in NY Project B in Tulsa
disappointed with delivery quality  20 developers  20 developers
− Hard to defend production issues – $120K per FTE – $120K per FTE
− Velocity is slowing down despite – 3 Function Points per staff month – 4 Function Points per staff month
AGILE and methodology changes – 180 Function Points delivered
– 240 Function Points delivered
$3,333/ Function Point
 Suppliers have become challenged to $2,500/ Function Point
meet SLAs that loosely relate to Project B seems to
output 2 critical violations per FP be 33% more 5 critical violations per FP
Cost for 360 fixes = $163K productive Cost for 1200 fixes = $545K!!!
− SLAs based on “request”, not
throughput $4,200/FP of TCO $4,700/FP of TCO

− Too many SLAs to focus on


Since
Since Project
Project A
A delivers
delivers better
better quality
quality (less
(less defects),
defects), itit is
is in
in fact
fact 13%
13% more
more productive!
productive!
effectively
Critical Violations Changeability
 Need a clear set of SLAs that tie to Transferability Functionality
application operations and
maintainability
Test Test
Plan Dev & QA
Plan Dev & QA

Test/ QA should be a condition of acceptance and payment


Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS
CAST Customer – Example SLA
At Risk Amount and Allocation of Risk
Total Billing Per Release : $1,000,000 10%
10% is
is for
for example
example
Total At Risk Amount (10% of Invoice) : $100,000
Total Risk Pool: 100%

Application Tier 1 Metrics (Critical At Risk At Risk


Name Service Levels) Multiplier Risk Allocation Amount
Amount
Amount service
service provider
provider has
has at
at
risk on this individual Service
risk on this individual Service • SLAs developed by
Total Quality Index 50% $15,000 Level
Level is:
is: strategic sourcing
30%
OMS Critical Violations 30% $9,000 30%
30% ** 50%
50% ** $100K
$100K = = $15,000
$15,000 organization at large telco
Application Pain Violations 20% $6,000
100% $30,000

Total Quality Index 30% $3,000


 Anytime there is a default, the • $$ amounts changed from
CRM Critical Violations 30% 10% $3,000 at-risk amount will be applied original; %’s are
Application Pain Violations 40% $4,000  Incentive is given to service
100% $10,000
preserved
provider equivalent to the at-
Total Quality Index 50% $10,000 risk amount if they exceed the
AMSS Critical Violations 30% 20% $6,000 Expected Service Level by 5%
• Mix between high-level
Application Pain Violations 20% $4,000
of the delta between the then indicator (TQI) and
100% $20,000
current Expected and specific risks (critical
Total Quality Index 50%
20%
$10,000 Perfection violations)
SDP Critical Violations 30% $6,000
Application Pain Violations 20% $4,000
 Credits/ Incentives are settled
100% $20,000 at the Annual Reset
Total Quality Index 50% $10,000
Enabler Critical Violations 30% 20% $6,000
Application Pain Violations 20% $4,000
100% $20,000
Confidential SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS
CAST Reference Video Links
BNY Mellon Reducing Application Outages: https://www.youtube.com/watch?v=PCnyeoVnKAw

Fannie Mae How Software Analytics Enhances DevOps Tools and Processes: https://www.youtube.com/
watch?v=bTpUHJ9F2vshttps://www.youtube.com/watch?v=oBHi-Dbry3g

Fannie Mae Finding and Reducing Software Risk: https://www.youtube.com/watch?v=hClN1AfP_Qo

Highlight Overview: https://www.youtube.com/watch?v=h_na7OoMOUc

IT Rationalization: https://www.youtube.com/watch?v=8MU6AnCZx4s&feature=youtu.be

Digital Transformation: https://www.youtube.com/watch?v=4-eGoM91v9o&feature=youtu.be

IT Due Diligence: https://www.youtube.com/watch?v=vvNXdyWuq5U&feature=youtu.be

Confidential 45 SOFTWARE INTELLIGENCE FOR DIGITAL LEADERS

You might also like