1. 2

I have ranted often in this and other forums that problem with project estimation is, that nobody actually cares about the answers.

Project Estimation is merely a data intensive exercise in bayesian modeling. ie. Just work. It can be done scientifically, it the answers will be probabilistic.

So I’m going to challenge my own assertion here….

Maybe I am wrong.

Maybe there are Project Managers out there that really care about project estimation, and are looking for a tool that handles the probability distributions correctly.

If I created a Software As A Service site…

  • that possibly integrated with Jira,
  • that allowed a team to properly estimate,
  • to calibrate their estimates,
  • and update them as new information arises
  • and output probability distributions for milestones of interest.

Possibly with some facilities to query and “What If” around adjusting parameters like resources and scope.

Possibly with some sensitivity analysis to indicate which parameters have most effect on various outputs.

Question for Lobsters:

  • Would any project manager pay for such a service?

  • Would any team do the data entry to accurately track time / story points / available man hours / …?


  2. 1

    If you read the tweet-fights between the #NoEstimates folks and folks such as Glen B. Alleman‏ (@galleman), you’ll find out who might be interested. (Half-joking)

    Seriously, though, have you looked at Troy Magennis’ tools and writing (http://focusedobjective.com/)? He has tools that use Monte Carlo simulation tools to come up with eerily accurate estimates with probabilities. Lots of free resources and a book.

    As for who would buy such a service, I don’t know. I don’t think a PjM would buy it (they just “manage” the project, they’re not typically on the line for poor estimates). My cynical view is that the folks who might be able to really use this don’t really care about accurate estimates, but only want adherence to a predetermined schedule.