Estimates and Yesterday’s Weather

Martin Olson, veteran software development Agilist/consultant is an active member of Agilehood KC and also a founding member of AgileKC

One of his Agilehood KC partners, Tom Westervelt, recently posted http://kc.agilehood.org/estimates-can-be-dangerous which is a great read and reminds us of the challenges estimating software development efforts.

It reminds me of Martin Fowler’s mention of Yesterday’s Weather when predicting software development productivity and Mike Cohn’s Agile Estimating work.

As Kent Beck once said, “Inaccurate estimates are a failure of information, not of values or principles. If the numbers are wrong, fix the numbers and communicate the consequences.”
Kent Beck, Extreme Programming Explained: Embrace Change

Want to know how long something will take? Work on it and measure time and if things you will work on in the future look like the type of work you’ve already done utilizing the same team, then you should be able to somewhat reliably predict/forecast time about to be spent.  I like to remind folks that estimates are always wrong. It’s merely a question of how wrong–less wrong being better.

It’s hard to resist the temptation of a project stakeholder asking for “just a ballpark guestimate” that “I won’t hold you to…” but often that’s the job and there is no one solution but it all makes for some great stories.