Estimation - The Art and Science

Recital Hall

*** Presented at Drupalcon LA 2015 to a full house of estimation enthusiasts.  This session will include parts of that presentation and expand even further upon those concepts.


Why as a software industry are we struggling so often with hitting our estimates?  Does the problems reside in the way we are estimating a project or the way we are executing them?  Through 30+ years of modern software development what have we learned and what are we continuing to struggle with today?  

In this session I would like to discuss not only the problems associated with estimation and how to avoid them, but more importantly how we can plan for them, turning our estimation process into not only an art, but a science. Well cover how to sell your estimate internally, and arm you with the methodologies to support your numbers, and avoid the pitfalls that create projects that go 100% overbudget.


What is the problem with software estimation?

  • The morale, metrics and realities we have learned over time

  • The results of our decades worth of estimation error

Avoiding Risk

  • Project entry point of sale

  • Risk association with point of sale

  • Products in the front, estimations in the back

  • Getting out of jail

  • Other lessons learned

The Elusive Discovery phase

  • How to estimate a discovery

  • How to sell a discovery

Planning for Risk 

  • Estimation types

    • Gut - An art form

    • Comparables - An art/science

    • Factors/formula - A science

  • Contingency

    • Rating systems

    • Formulas

    • Granularity

Business and Open Source