Do You Underestimate The Tasks You Like?

I’m sure I do this and I’m wondering whether you think everybody does. When I’m asked to estimate how much time some task is going to take:

  • If it’s something I enjoy doing, like writing, or programming, or driving  on an open road, my estimate is always too low.
  • If it’s something I don’t enjoy, like chores, or long meetings, or driving through city traffic, then my estimate is always too high.

I’m serious. Back in the 1990s when I was actually programming real software product – I did a third of the code in the first Business Plan Pro – my estimates of the time things would take were terrible. Even today, when I’m mostly writing, the time I actually take doing this blog post, to name one example, is about three times more than what I would have thought when I started it.

And that’s a business problem, not just a random thought.  Coordinating and collaborating requires managing time estimates. How do you coordinate marketing implementation and launch, for example, if time estimates are wrong?

It was hell on me when I was getting any of my books finished. It was so much more fun to imagine them, outline and highlights and all, think about them, than to actually write all those words.

I tried to double or triple my own estimates for my own work, but that didn’t work very well either, because every so often I’d get really psyched on a job, get into a zone, and finish it as quickly as I originally thought I would.

So I have no solution to this problem. Maybe it’s a good thing to have jobs you like as much as I’ve liked some of mine. But it’s hard to deal with. Do you have a solution?

(Image: Enrico Fianchini/istockphoto)


  • AviD says:

    Yes, actually – I found a fantastic solution, embodied by Fog Creek’s fantastic project planning tool – Fogbugz.

    This tool implements EBS – evidence based scheduling, which is basically repeated and optimized Monte Carlo simulations on your actual history of estimates, using your time tracking as input.

    So, the first time you might still be stuck, but if you repeatedly make the same error in estimates, Fogbugz will know statistically, based on historical *evidence*, how long it will really take you. Turns out it’s pretty accurate.

    Avi D

    • Tim Berry says:

      @AviD thanks for adding that. Generally I don’t approve comments that are product plugs like yours for FogBugz, but I’m hoping you’re not directly connected with the company, and we use Fogbugz at Palo Alto Software, too. Furthermore, I really liked Joel Spolsky’s blog until he stopped it, but I still read his stuff when he publishes elsewhere. (Joel is co-founder of Fog Creek Software).

      Ironically, FogBugz wasn’t something I was using in the 1990s when I was actually programming portions of the product, and I don’t use it now because I’m not doing development anymore. Maybe I should though … you have a point.

  • company profiles says:

    Thanks for this – I enjoyed reading it a lot!

    I think it’s certainly fair to say that I’m guilty of under-estimating tasks if I enjoy them. Partly, it may be because I recollect them as being easier (and less time-consuming) than they really are.

    In a business environment, this weakness certainly has dangers associated with it. May be time to try and change my approach.

  • Charles Robinson says:

    I look at my actually delivered time and use that to make estimates. I don’t rely on my gut, I try to make it as fact-based as possible.

Leave a Reply

Your email address will not be published. Required fields are marked *