Hmm. Giving up task estimates is a worthy goal but it eliminates the idea of being able to calculate ROI, because there's no...I.
Having managed SWEs in both big & small tech companies, here's the tactic that works equally well:
Never provide an estimate without a confidence level.
Example: that story is probably 3 days of work with a confidence level of 5/10.
You never, ever, provide the estimate without the confidence.
Not in a DM to the PM, not in a slide deck, not even if the CEO is asking. This is where Engineering leaders can instill some discipline. Help us help you. No estimate without confidence level!
When the PM freaks out about the 5/10, they'll ask what it would take to make it an 8/10. That's when you discuss alternatives: Spikes, MVP, or even better....reply with the real question that nobody asks:
"Hey PM/EM, what is the max level of Eng effort where this is still worth doing (has ROI)?"
If they say 10 days, and you have 8/10 confidence it can be done in 10 days or less, then you proceed. Otherwise, you don't.
P.S. If you're having daily 30 min standups, please scream about that and get it changed. Sincerely, everyone
Great post! It seems like your ideas align well with Shape Up which advocates for bets on whether a product feature or idea can be completed over 4-6 weeks.
Zach, thanks for writing this post. I agree with you, the practice of estimation needs revisiting. I wrote this post on the topic and would love to get your thoughts: https://open.substack.com/pub/daliahavens/p/less-estimation-more-iteration
Hmm. Giving up task estimates is a worthy goal but it eliminates the idea of being able to calculate ROI, because there's no...I.
Having managed SWEs in both big & small tech companies, here's the tactic that works equally well:
Never provide an estimate without a confidence level.
Example: that story is probably 3 days of work with a confidence level of 5/10.
You never, ever, provide the estimate without the confidence.
Not in a DM to the PM, not in a slide deck, not even if the CEO is asking. This is where Engineering leaders can instill some discipline. Help us help you. No estimate without confidence level!
When the PM freaks out about the 5/10, they'll ask what it would take to make it an 8/10. That's when you discuss alternatives: Spikes, MVP, or even better....reply with the real question that nobody asks:
"Hey PM/EM, what is the max level of Eng effort where this is still worth doing (has ROI)?"
If they say 10 days, and you have 8/10 confidence it can be done in 10 days or less, then you proceed. Otherwise, you don't.
P.S. If you're having daily 30 min standups, please scream about that and get it changed. Sincerely, everyone
Great post! It seems like your ideas align well with Shape Up which advocates for bets on whether a product feature or idea can be completed over 4-6 weeks.