Comments on: Software estimation nightmare

By: Jagdish

Jagdish — Sat, 28 Nov 2009 21:58:06 +0000

Very good inputs on estimates. This is exact situation I faced recently in a project.

By: CertPal

CertPal — Mon, 12 Oct 2009 17:00:01 +0000

@fsilber On the contrary there are times when the opposite is also true. There are project that wrap up quickly before schedule, although rare. Things do not always go wrong but it is safe to add buffers to estimates in case something pops up out of the blue.

By: fsilber

fsilber — Mon, 12 Oct 2009 10:50:56 +0000

What is the difference between overestimating versus recognizing the fact that there will _always_ be many tasks in the project that you didn’t think of when making your estimate?

By: CertPal

CertPal — Fri, 09 Oct 2009 15:09:51 +0000

@David Clark
I agree that the last point sounds too irrational, but it has saved me more than once. I should probably not have said ‘Always overestimate’ which makes it sound like some thing will go wrong every time. I do over estimate and add buffers. Beefing up beyond that is also something I do, but not all the time.

I have suffered from the Sandbagging before too Thanks for leaving your thoughts

By: David Clark

David Clark — Fri, 09 Oct 2009 14:23:01 +0000

I have seen a lot of what you mention. It goes on at every company I have ever been at. And I agree with most of your lessons — except the last. “Sandbagging” your estimates just leads to a never-ending spiral of deception. Once your manager figures out that you are over estimating, they will just mentally reduce any estimate you give them. You then figure out that your boss is mentally adjusting your figures and compensate again. And so on. And you lose credibility at the same time you lose integrity.

Always just give the best estimate you can come up with in the time your allowed to research the problem. You retain your self-respect and as your estimates get better and better over time the trust your colleagues and boss place in you will grow.