Re:How to make Product give a shit about your architecture p...
by ahoef on Thursday October 10, 2024 @12:58AM
(#41795666)
Indeed. Things I've used before are: "product speed is declining, see this chart. We must make these optimization to stay acceptable." And for example "We must migrate to this hosting model to keep compute costs down."
[ Parent
| on HN ]
Re:How to make Product give a shit about your architecture p...
by dns_snek on Thursday October 10, 2024 @08:13AM
(#41798058)
> product speed is declining, see this chart.
How do you measure product speed in a useful manner?
I'd disqualify metrics such as number of tickets, lines of code, and "story points", because they either have an undefined relationship with product velocity (e.g. LOC per week), can vary in size too much to be useful (e.g. number of tickets per week), or they're tautological (e.g. story points per week).
What's left?
Re:How to make Product give a shit about your architecture p...
by hiatus on Thursday October 10, 2024 @09:25AM
(#41798550)
I took it to mean the speed of the product, not developer velocity.
Re:How to make Product give a shit about your architecture p...
by actinium226 on Thursday October 10, 2024 @11:23AM
(#41799811)
I've noticed no one seems to look at number of PRs per week or average length of time and MR is open. Those seems like they should be good metrics for both how well a team is collaborating and how fast they're moving (perhaps normalized by LOC in the MR, but then again smaller MRs tend to go through faster so maybe LOC is already reflected in the metric).