Estimation

Many folks in the industry champion the use of estimation, but should you always use it? Looking back on the 16 years I've spent building software (and the roughly 200+ estimation sessions I've participated in) I am left with the conclusion that estimation has never materially mattered in any of the projects I've worked on. In fact, you probably don't need estimation for most of your projects either. Here are some things to consider if you're trying to determine whether estimation will help or become a futile time sink.

You Don't Need Estimation if a Specific Vision is Required

If there are a minimum set of features that must be completed before the project is accepted, then you probably don't need estimation. A good way to identify whether this is true is to imagine what will happen if the project runs late. If you suspect that the project running late will result in you being sent away, never to return until "finished" is reached, then estimation is probably not right for your project.

You Don't Need Estimation If Your Teams Are Stable

A rare area where estimation might be useful is when you're constrained on people and plan to shuffle engineers around various projects. An example might be that you have a certain team of engineers work on a project in the first quarter and you are going to dovetail some of those engineers to another project in the second quarter. This also hinges on whether an incomplete project will be acceptable in the first place, otherwise your shuffling will be blocked. It's worth noting that this idea is more popular in discussion but is not as readily done in practice. I find even the most proactive companies tend to stabilize into longer running teams after a while.

Estimation Might Be Useful if an Incomplete Project or Alternative is Deployable

If an incomplete feature set or a more predictable alternative is acceptable, then estimation might be useful to identify whether a certain longer approach can be taken. If the longer approach fails, you can switch to a more predictable alternative and it will still be accepted. It's important to note that by "acceptable" I mean that it will be deployed to production and used. If you will not be allowed to deploy to production, or an alternative/incomplete project will largely go unused, then see above.

Estimation Is Not Prioritizing or Communicating

As a disclaimer, I'd like to point out that while estimation might not be useful most of the time, prioritizing and checking in are still useful exercises.

Alternatives - The Shape Up Method