How can I get better at estimating long-term projects?

Question

I’m okay at planning out short-term projects, but I have trouble with longer time horizons. As a new engineer here at Mercari, I'm currently working on a 90 day plan and I've been thinking about how to estimate projects. I want to figure out timelines in particular, so I can properly manage expectations. A lot of the time, I’m over-optimistic about what I can achieve, so I want to build a better muscle there.

For reference, I consider <1 month to be short-term and 2+ months to be long-term.

Alex Chiou · Accepted Answer

In a vacuum, just add 50% to your initial estimates. So 4 weeks becomes 6 weeks. Literally every earlier in career engineer I've seen underestimates work (similar to yourself), and 50% is a great number from my experience (it's sort of uncanny how perfect I've seen it be). It's neither too aggressive (not increasing the time enough) nor too cushy (increasing the time too much). Break things down to the point where they're easy to estimate. Every long-term project is just a bunch of a short-term projects in disguise During retros, make sure to ask yourself how accurate the estimates were. If they missed by a lot, learn from that and bake those learnings into future estimations. Aggressively leverage other engineers to improve your estimates, especially given that you're new. Timelines are very much a team exercise.  Related resources:  How do I make software less overwhelming? I've been given crazy 0 to 1 projects - How can I scope them out and disambiguate them?

How can I get better at estimating long-term projects?

Discussion