The 100k Proposition

      Liz Keogh
      April 2011
Meet the Team
Assumed Costs per Day
£700



              £350
The Team’s Budget
   £800k
    over
   1 year
The Team’s Scope
   4 features
 every 2 weeks

 (80 features)
Quality




75% of features         10% of time
   are used             is bugfixing
Quality

A bug takes ½ day on average

         Each week,
      4 devs fix 8 bugs

     Over the project,
     devs fix 320 bugs
The Goal


To show how Agile training
      for a new team
     can pay for itself
   by reducing the cost
    of an £800k project
         by £100k.
Descoping a feature

          Hmm! They
          won’t really
          need that.
Descoping a feature


         Each feature takes
         1 dev 10 days
         and 1 tester 2 ½ days

         Saving:
         12 ½ x £350 = £4375
Descoping a feature


         If we get 50% of
         unused features out



         Saving:
         12.5% of £800k = £100k
Done!

That was
 easy!

 Can we do it
   without
 descoping?
The Goal

To show how Agile training
      for a new team
     can pay for itself
   by reducing the cost
       of the project
     without reducing
        the scope.
Finding a bug early
        Oops! We
         found a
           bug!

          Oops! Let
           me fix
           that…
Finding a bug early

   Now:                   In 2 months:

½ hour = £25             ½ day = £175


        Saving: £150 per bug
Finding a bug early




If we do it for ½ the bugs,
  saving over the project
½ x £150 x 320 = £24,000
Increasing code quality by 25%
           This is
            much
           better!
           Do I get to do
             something
             new then?
Increasing code quality by 25%
4 dev days per week -> 3
½ tester day per week -> ¼
(fewer bugs to write up too)

Saving: £436.50 per week

If this happens ½ way through
the project:

Saving: £8730
Increasing code quality by 25%
            Hold on. We
            take ½ hour
             now, not ½
                day!

              For ½ the
               bugs…

           Cheat. You can’t
            count it twice!
Increasing code quality by 25%

4 dev days per week -> 3
(x ½ the bugs)
2 dev hours per week -> 1 ½
(x ½ the bugs)
½ tester day per week -> ¼

Saving for ½ the project: £5,730
Co-location and collaboration

  If you could do the same
    project again with the
     same people and the
   same requirements and
   technologies, how long
        would it take?
Co-location and collaboration


        At least 60% of a
            project is
             learning.
Co-location and collaboration
Increasing the rate of learning by 10%
(really 20% but not counting other things twice)
= 60% x 10% x 800k

Saving: £48,000
Co-location and collaboration
        Co-location is
         impossible
            here!

       That’s a shame.
        Let’s look at
         some other
         practices…
Learning comes mostly from
Co-location             50%       £24,000
Stand-ups               17%       £8,000
Retrospectives          17%       £8,000
Breaking down stories   8%        £4,000
Showcases               8%        £4,000


          Saving without co-location:
                   £24,000
Late Change in Requirements
        Time taken to agree change and
        update documentation
        reduced from 1 day to 1 hour

        10% of analysis needs changing

        6 hours x £50 x 10% x 200 days

        Saving: £6,000
Automated tests
      The cost of
     testing goes
     up linearly as
       the code
        grows.
Automated tests
  Cost of developers helping with
  automated tests at 1 day every
  2 weeks: £7000


  Not needing to hire another
  tester in the last ¼ of the
  project: £17,500

  Saving: £10,500
Motivation and effectiveness
                            More time
                 I get    exploring, less
I get to see   feedback    time writing
  my work      on mine!    bug reports!
  working!
Motivation and Effectiveness
      If motivation makes team
        only 4% more effective

          Saving: £32,000
Practices                 Total earned
Finding bugs early            £24,000
Increasing code quality         £5,730
Learning practices            £24,000
Changes in requirements         £6,000
Automated testing             £10,700
Motivation &                  £32,000
effectiveness

Total                        £102,430
Done!
      The numbers
    provided here are
   fairly conservative,
    and these are just
  some of the benefits
     Agile can offer.
The Real Benefit

  The real benefit of Agile over
  Waterfall happens when we
  showcase or release, and find
  out we did the wrong thing.

  Ideally, we’re only wrong by 2
  weeks, instead of 1 year.
The Real Benefit
  Or, we get to release early, and
  start getting benefit from the
  right thing.

  That means 6 months extra of
  earning money from the early
  release.

  Either way, that’s a lot of cash.
Agile.
Why wouldn’t you do it?

The 100k proposition

  • 1.
    The 100k Proposition Liz Keogh April 2011
  • 2.
  • 3.
    Assumed Costs perDay £700 £350
  • 4.
    The Team’s Budget £800k over 1 year
  • 5.
    The Team’s Scope 4 features every 2 weeks (80 features)
  • 6.
    Quality 75% of features 10% of time are used is bugfixing
  • 7.
    Quality A bug takes½ day on average Each week, 4 devs fix 8 bugs Over the project, devs fix 320 bugs
  • 8.
    The Goal To showhow Agile training for a new team can pay for itself by reducing the cost of an £800k project by £100k.
  • 9.
    Descoping a feature Hmm! They won’t really need that.
  • 10.
    Descoping a feature Each feature takes 1 dev 10 days and 1 tester 2 ½ days Saving: 12 ½ x £350 = £4375
  • 11.
    Descoping a feature If we get 50% of unused features out Saving: 12.5% of £800k = £100k
  • 12.
    Done! That was easy! Can we do it without descoping?
  • 13.
    The Goal To showhow Agile training for a new team can pay for itself by reducing the cost of the project without reducing the scope.
  • 14.
    Finding a bugearly Oops! We found a bug! Oops! Let me fix that…
  • 15.
    Finding a bugearly Now: In 2 months: ½ hour = £25 ½ day = £175 Saving: £150 per bug
  • 16.
    Finding a bugearly If we do it for ½ the bugs, saving over the project ½ x £150 x 320 = £24,000
  • 17.
    Increasing code qualityby 25% This is much better! Do I get to do something new then?
  • 18.
    Increasing code qualityby 25% 4 dev days per week -> 3 ½ tester day per week -> ¼ (fewer bugs to write up too) Saving: £436.50 per week If this happens ½ way through the project: Saving: £8730
  • 19.
    Increasing code qualityby 25% Hold on. We take ½ hour now, not ½ day! For ½ the bugs… Cheat. You can’t count it twice!
  • 20.
    Increasing code qualityby 25% 4 dev days per week -> 3 (x ½ the bugs) 2 dev hours per week -> 1 ½ (x ½ the bugs) ½ tester day per week -> ¼ Saving for ½ the project: £5,730
  • 21.
    Co-location and collaboration If you could do the same project again with the same people and the same requirements and technologies, how long would it take?
  • 22.
    Co-location and collaboration At least 60% of a project is learning.
  • 23.
    Co-location and collaboration Increasingthe rate of learning by 10% (really 20% but not counting other things twice) = 60% x 10% x 800k Saving: £48,000
  • 24.
    Co-location and collaboration Co-location is impossible here! That’s a shame. Let’s look at some other practices…
  • 25.
    Learning comes mostlyfrom Co-location 50% £24,000 Stand-ups 17% £8,000 Retrospectives 17% £8,000 Breaking down stories 8% £4,000 Showcases 8% £4,000 Saving without co-location: £24,000
  • 26.
    Late Change inRequirements Time taken to agree change and update documentation reduced from 1 day to 1 hour 10% of analysis needs changing 6 hours x £50 x 10% x 200 days Saving: £6,000
  • 27.
    Automated tests The cost of testing goes up linearly as the code grows.
  • 28.
    Automated tests Cost of developers helping with automated tests at 1 day every 2 weeks: £7000 Not needing to hire another tester in the last ¼ of the project: £17,500 Saving: £10,500
  • 29.
    Motivation and effectiveness More time I get exploring, less I get to see feedback time writing my work on mine! bug reports! working!
  • 30.
    Motivation and Effectiveness If motivation makes team only 4% more effective Saving: £32,000
  • 31.
    Practices Total earned Finding bugs early £24,000 Increasing code quality £5,730 Learning practices £24,000 Changes in requirements £6,000 Automated testing £10,700 Motivation & £32,000 effectiveness Total £102,430
  • 32.
    Done! The numbers provided here are fairly conservative, and these are just some of the benefits Agile can offer.
  • 33.
    The Real Benefit The real benefit of Agile over Waterfall happens when we showcase or release, and find out we did the wrong thing. Ideally, we’re only wrong by 2 weeks, instead of 1 year.
  • 34.
    The Real Benefit Or, we get to release early, and start getting benefit from the right thing. That means 6 months extra of earning money from the early release. Either way, that’s a lot of cash.
  • 35.