Six Sigma in a software products
environment




Jeff Lee
Six Sigma Team Leader, Investment Services Group at Fiserv

NYC ...
Agenda



  • Fiserv introduction
  • Introduction of Six Sigma
  • How it applies to a software products environment
  • ...
But first a word from my sponsor…




                     3 © 2009 Fiserv. All Rights Reserved.
Providing Scale Advantages And Differentiation




                 Financial Strength Leading Market                     ...
Strategic Framework


Vision     A global leader in transaction-based
           technology solutions



Mission    To pro...
Introduction and History of Six Sigma

                        What Does It Do?

R       Defects
                         ...
Six Sigma is a process

        Define    Identify what is important to the customer and define
                  the proj...
Six Sigma -- What’s it based on?
 Customer
       Anyone Who Receives
       Product, Service or Information


 Opportunit...
Variation -- less is better:




  Less Variation ( x’s). .
  Means fewer defects (y’s)
  (if centered)




              ...
Standard deviation:
     A metric that displays variation from it’s “target”.




                            1 Std. Dev.
...
If we can squeeze six standard deviations in
      between our target and the customer’s
               requirements . . ....
Six Sigma... the analytical goal:

              (DPMO Distribution Shifted ± 1.5σ)

                                     ...
Isn’t 99% Good Enough?

 A 99% level of performance would mean:
 • 20,000 lost articles of mail per hour by the United Sta...
What are some applications to software products?


•Focus on maximizing value-added time in a Systems Development Life
Cyc...
Example 1: Improve turn-over to a QA group

                          New builds are often delivered from the development
...
Cause & Effect Diagram
                             Methods                                               People

Changes ...
CTQs
  CTQ Tree           Drivers
                                                                 Document delivered to Q...
Example 2: Theory of Constraints – 3 Projects

  Project 1          A B C D E F G H I J                                   ...
Scenario 1 – Multi-tasking
A           B           C D         E                  F                     G H             I ...
Example 3: Statistics and the Big Picture


Initial Data Set




                      20 © 2009 Fiserv. All Rights Reserv...
The analysis – Pearson product moment correlation
 coefficient ( a.k.a. Pearson coefficient)

Correlations: Group A, Group...
A more intelligent review – Regression analysis with a
graphic
                                              Fitted Line P...
Re-do without the outlier
Correlations: Group A_1, Group B_1

Pearson correlation of Group A_1 and Group B_1 = 0.868
P-Val...
The regression analysis shows this as well
                                            Fitted Line Plot
                  ...
Conclusion

  • Six Sigma applied to software products should fit form to
   function (not all manufacturing best practice...
References


What is Lean Six Sigma – Mike George, Dave Rowlands, & Bill Kastle
iSixSigma http://www.isixsigma.com/; Softw...
Thank You!


Questions?



  27 © 2009 Fiserv. All Rights Reserved.
Upcoming SlideShare
Loading in...5
×

Presentation title Arial 30pt Second line

282

Published on

0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
282
On Slideshare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
3
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Transcript of "Presentation title Arial 30pt Second line"

  1. 1. Six Sigma in a software products environment Jeff Lee Six Sigma Team Leader, Investment Services Group at Fiserv NYC SPIN, March 11th, 2009
  2. 2. Agenda • Fiserv introduction • Introduction of Six Sigma • How it applies to a software products environment • Examples/Case Studies • References 2 © 2009 Fiserv. All Rights Reserved.
  3. 3. But first a word from my sponsor… 3 © 2009 Fiserv. All Rights Reserved.
  4. 4. Providing Scale Advantages And Differentiation Financial Strength Leading Market By The Numbers Position $3.68 Billion Revenue 25,000 Employees as of 3Q 2008 Managed Accounts 250 Locations $7.75 Billion Market Transaction Around the World Cap as of 3Q 2008 Processing Over 18,000 clients 17.97% Operating Solutions Margin as of 3Q 2008 Over 20 Billion Integrated Solutions Transactions High recurring Processed Annually revenue and strong Core Processing operating margins Serve All Top 100 Internet Banking Banks, with 6,000 Consistent double Bill Payment & Core Relationships digit revenue and Presentment earnings growth Debit/EFT Processing Strong free cash flow generation ($465 Risk Management Million as of 3Q 2008) 4 © 2009 Fiserv. All Rights Reserved.
  5. 5. Strategic Framework Vision A global leader in transaction-based technology solutions Mission To provide integrated technology and services solutions that enable best in class results for our clients Values Clients • Integrity • Respect • Innovation • Teamwork 5 © 2009 Fiserv. All Rights Reserved.
  6. 6. Introduction and History of Six Sigma What Does It Do? R Defects I Customer satisfaction e Mistakes m Quality d Wasted effort p Service u Delays r Cost savings Scrap o Productivity c Time compression e v e Result: Customer expectations met then exceeded 6 © 2009 Fiserv. All Rights Reserved.
  7. 7. Six Sigma is a process Define Identify what is important to the customer and define the project scope. Measure Select key characteristics; Define performance standards; Validate a measurement system Analyze Establish a process capability; Define improvement objectives; Identify variation sources Improve Screen potential causes for variation; Discover variable relationships b/w causes; Establish operating tolerances Validate a measurement system; Determine the ability to Control control vital steps; Implement process controls on vital steps 7 © 2009 Fiserv. All Rights Reserved.
  8. 8. Six Sigma -- What’s it based on? Customer Anyone Who Receives Product, Service or Information Opportunity Every Chance to Do Something Either “Right” or “Wrong” Successes vs. Defects Every Result of an Opportunity Either Meets the Customer Specification or it Doesn’t 8 © 2009 Fiserv. All Rights Reserved.
  9. 9. Variation -- less is better: Less Variation ( x’s). . Means fewer defects (y’s) (if centered) 9 © 2009 Fiserv. All Rights Reserved.
  10. 10. Standard deviation: A metric that displays variation from it’s “target”. 1 Std. Dev. (“Sigma”) One standard deviation around the mean is about 68% of the total “opportunities” for achieving success! 10 © 2009 Fiserv. All Rights Reserved.
  11. 11. If we can squeeze six standard deviations in between our target and the customer’s requirements . . . 6 5 4 3 2 1 1 2 3 4 5 6 then: 99.99966% of our “opportunities” are included! 11 © 2009 Fiserv. All Rights Reserved.
  12. 12. Six Sigma... the analytical goal: (DPMO Distribution Shifted ± 1.5σ) Parts per σ σ Parts per Million Million Sigma is a statistical unit of measure which reflects 2 308,537 process capability. The sigma scale of 3 66,807 measure is perfectly 4 6,210 correlated to such characteristics as 5 233 defects-per-unit, parts-per million 6 3.4 defective, and the probability of a Process Process Defects per failure or error. Capability Capability Million Opportunities 12 © 2009 Fiserv. All Rights Reserved.
  13. 13. Isn’t 99% Good Enough? A 99% level of performance would mean: • 20,000 lost articles of mail per hour by the United States Postal Service • 5,000 incorrect surgical operations per week in the United States • 200,000 wrong drug prescriptions each week in the United States • No electricity for almost 7 hours each month • 12 Newborns will be given to the wrong parents daily • 315 entries in Webster’s Dictionary will be misspelled • 3,065 copies of tomorrow’s Wall Street Journal will be missing one of the three sections From the book by Ed Scannell, Games Trainers Play, 2000 And Garden State Focus October 2001 Edition 13 © 2009 Fiserv. All Rights Reserved.
  14. 14. What are some applications to software products? •Focus on maximizing value-added time in a Systems Development Life Cycle process (time is money) and minimizing re-work and waste •Reduce bugs – either in production or by phase •Reduce document revisions – BRDs, FSDs, TSDs, Test Cases, etc. •Improve business requirements gathering by identifying the critical few items that matter most •Understand operational performance 14 © 2009 Fiserv. All Rights Reserved.
  15. 15. Example 1: Improve turn-over to a QA group New builds are often delivered from the development team to QA without enough information, such as details of changed functionality and installation What is wrong? instructions. This leads to extra hours spent tracking down this information and extends the cycle time of the QA process. The problem occurs at the point in the Software Where does the problem Development Life Cycle when a new build is delivered occur? to the QA team from the Development team. This problem occurs when a new delivery of the code When does the problem is received by the QA team occur? Approximately 90% of new deliveries have this To what extent does the problem problem occur? 15 © 2009 Fiserv. All Rights Reserved.
  16. 16. Cause & Effect Diagram Methods People Changes not tracked Development team Members of QA team manager not enforcing not getting enough agreed procedures info, having to “chase” Changes not documented as Dev team they are coded Not all Dev team adhering Procedures not followed to procedures, or providing information early enough Rushed deliveries made to QA The delivery of new builds from Development to QA takes too long to An installation guide is not achieve and results Release notes do not have always provided in higher than enough info re: bug fixes, acceptable defect contents of release levels. Team Track not always updated with latest status of CCR’s delivered in build Materials 16 © 2009 Fiserv. All Rights Reserved.
  17. 17. CTQs CTQ Tree Drivers Document delivered to QA at time of release, for every new build Release Notes Complete up to date list of all listing all changes changes provided every time TeamTrack updated correctly prior to QA commencing testing for every delivery Need A list of all elements that need to be installed for every build Smoother process for Installation Directions where items need delivering instructions to be installed, every delivery new builds to Directions on how to install all QA parts of every delivery A list of known issues and fix, every delivery Limitations / known issues Potential problems listed in each delivery Any missing deliverables and their expected delivery date, every delivery 17 © 2009 Fiserv. All Rights Reserved.
  18. 18. Example 2: Theory of Constraints – 3 Projects Project 1 A B C D E F G H I J 20 Weeks Project 2 A B C D E F G H I J 20 Weeks Project 3 A B C D E F G H I J 20 Weeks Resource 1 Resource 2 Resource 3 From the book by Kai Yang, Design for Six Sigma for Service 18 © 2009 Fiserv. All Rights Reserved.
  19. 19. Scenario 1 – Multi-tasking A B C D E F G H I J 48 Wks A B C D E F G H I J 50 Wks A B C D E F G H I J 52 Wks Scenario 2 – No Multi-tasking A B C D E F G H I J 20 Wks A B C D E F G H I J 28 Wks A B C D E F G H I J 36 Wks The scenario without multi-tasking had the longest project still finish before the first project in Scenario 1 19 © 2009 Fiserv. All Rights Reserved.
  20. 20. Example 3: Statistics and the Big Picture Initial Data Set 20 © 2009 Fiserv. All Rights Reserved.
  21. 21. The analysis – Pearson product moment correlation coefficient ( a.k.a. Pearson coefficient) Correlations: Group A, Group B Shows very little correlation Pearson correlation of Group A and Group B = 0.143 P-Value = 0.658 Implies no statistical significance But we should always seek to understand the data graphically and not just mathematically – otherwise we can’t see the forest for the trees 21 © 2009 Fiserv. All Rights Reserved.
  22. 22. A more intelligent review – Regression analysis with a graphic Fitted Line Plot Group B = 1.85E+08 + 0.564 Group A 1400000000 S 357490387 R-Sq 2.0% 1200000000 R-Sq(adj) 0.0% 1000000000 800000000 Group B 600000000 Again, the numbers 400000000 don’t imply any relationship 200000000 0 0 0 0 0 0 0 000 0 00 000 000 0 00 00 00 00 00 00 50 10 0 15 0 20 0 25 0 But the graph shows Group A an outlier 22 © 2009 Fiserv. All Rights Reserved.
  23. 23. Re-do without the outlier Correlations: Group A_1, Group B_1 Pearson correlation of Group A_1 and Group B_1 = 0.868 P-Value = 0.001 Implies statistical Shows high significance positive correlation 23 © 2009 Fiserv. All Rights Reserved.
  24. 24. The regression analysis shows this as well Fitted Line Plot Group B_1 = 41434037 + 1.142 Group A_1 400000000 S 62228124 R-Sq 75.4% R-Sq(adj) 72.7% 300000000 Group B_1 200000000 A different story 100000000 0 0 00 00 00 00 00 000 000 000 000 000 00 00 00 00 00 5 10 15 20 25 No other outliers Group A_1 present Worth examining further if this correlation is not desired 24 © 2009 Fiserv. All Rights Reserved.
  25. 25. Conclusion • Six Sigma applied to software products should fit form to function (not all manufacturing best practices apply) • Apply the methodology and toolset when it adds value to the business outcome and client experience • Like all tools, Six Sigma exists to assist the business not the other way around - use common sense • True fundamentals of Six Sigma should be honored: • Top down leadership • Define quality in the eyes of the client – what are they willing to pay for? • Define the problem before solutions • Data driven decision making – but not paralysis by analysis • Quality is not a separate function but a building block of all functions 25 © 2009 Fiserv. All Rights Reserved.
  26. 26. References What is Lean Six Sigma – Mike George, Dave Rowlands, & Bill Kastle iSixSigma http://www.isixsigma.com/; Software and IT channels • David L. Hallowell • Bruce J. Hayes SEI @ Carnegie-Mellon University http://www.sei.cmu.edu/ (type Six Sigma under search) ASQ – http://www.asq.org/six-sigma/index.html 26 © 2009 Fiserv. All Rights Reserved.
  27. 27. Thank You! Questions? 27 © 2009 Fiserv. All Rights Reserved.
  1. A particular slide catching your eye?

    Clipping is a handy way to collect important slides you want to go back to later.

×