t+1 - Latest Comments in Don’t negotiate on your estimates

Re: Don’t negotiate on your estimates

cheap condoms — Mon, 21 Sep 2009 06:22:21 -0000

The flip side of that, though, is that we need to be true to that by *revising* an estimate as requirements change, and more importantly as we get a better understanding of the facts that affect our estimate.

Re: Don’t negotiate on your estimates

Ben Finney — Tue, 09 Jun 2009 00:04:32 -0000

The point, that estimates are not an appropriate target for negotiation, is good and needs to be spread widely. Estimates are like weather forecasts: they're an expression of one's understanding of truth, not a promise or a commitment.

We also need to treat estimates more like forecasting: as testable predictions, that can be improved by examining the track record and learning how to estimate better. If we want the recipients of these estimates to trust us, we need to commit to improving our skill at estimating.

Re: Don’t negotiate on your estimates

Aaron Oliver — Sun, 07 Jun 2009 19:38:26 -0000

THANK YOU for acknowledging that it may be necessary for us to get along with People Not Like Us. Too often we brand entire classes as "dumb" or "ignorant", when really they're just wired to do a different job than us.

Re: Don’t negotiate on your estimates

Brian Wojcik — Sun, 07 Jun 2009 17:46:17 -0000

Interesting take... I'd say that the software managers whom I've worked with and respect most were more than happy to enter into negotiations with sales and marketing. Like you suggest they didn't allow the discussion to be about how long features would take to implement. They instead took the opportunity to negotiate over feature sets and release dates.

They would approach things like "Whatever your experience at {BIGGER COMPANY} doesn't matter, my team puts their estimate on that feature at eight points. If you want it to go in the next release, we'll have to slip the date by a month."

Of course this only works well if both sides have some level of good faith about realizing that the engineering team has finite limits of people and time to throw at any problem and that ultimately, success for the organization is about engineering delivering a saleable product in a time frame that the sales team can use. Otherwise, it just becomes an ugly and destructive game and nobody wins.