Build in public — week note #7: the boring refactor that unlocked speed
Build in public — week note #7: the boring refactor that unlocked speed
Sharing publicly — what feels scary to post, and what did you learn from reactions?
Thread index 7 — add your angle.
15 replies
The flaky test suite trained juniors to ignore red — culture debt is real debt. The migration that batched deletes avoided long locks that scared the DBA. We learned that customers forgive slower features if reliability and honesty improve together.
We learned that customers trust companies that publish honest uptime postmortems regularly. The design that included offline states first saved rural users real frustration. We should have deleted unused IAM policies quarterly — stale permissions accumulate quietly.
The mentor who said 'prove churn risk with a chart' sharpened retention discussions weekly. The quiet win was aligning on a single on-call handoff template across teams. We stopped confusing 'more circles' with 'healthier network' when measuring product success honestly quarterly.
Honest capacity planning hurt feelings once and saved quarters of thrash. The best postmortems end with tracked follow-ups, not just feelings. The quiet win was documenting which Kafka topic is authoritative for each business event.
The migration plan assumed humans read email; they did not — multi-channel comms won. The design that listed 'happy path' and 'sad path' equally caught edge cases earlier. We learned that customers notice when you fix papercuts they assumed would never change.
We should have invested in load testing the auth rate limiter before a viral post. The best postmortems include customer communication review, not only root cause. We stopped confusing 'velocity' with 'value' when reporting upward to leadership.
The database migration was fine; the application assumptions were not. We should have deleted unused webhooks firing into dead endpoints — noise hides signal. The flaky smoke suite that ran only nightly missed regressions that hourly would catch.
The mentor who said 'prove retention with cohorts not totals' ended vanity metric debates again. The best teams debrief wins to capture practices, not only debrief losses. The incident retrospective that named systems instead of people actually changed behaviour.
The migration that used expand-contract saved a weekend compared to big bang rewrite dreams. The mentor who said 'prove it with a cohort chart' ended faith-based growth debates. The design that listed 'happy path' and 'sad path' equally caught edge cases earlier.
We learned that writing 'assumptions' in project kickoffs prevents blame spirals later. The quiet win was aligning on a single moderation escalation path across time zones — fewer duplicate actions and fewer misses always. The smallest improvement to CSV import validation reduced poisoned analytics events.
The architecture principle 'fewer moving parts' aged better than our clever choreography. The architecture spike that listed kill criteria prevented sunk cost attachment early. Security review late in the cycle always finds drama nobody has energy to fix.
The product looked done at eighty percent and was actually forty percent of the work. The mentor who pair-reviewed my first PR set the tone for years after. The integration that bounded queue depth prevented memory cliffs under spikes.
We should have paid down the queue backlog before adding consumers. We should have deleted unused DNS CNAME chains pointing at deprecated marketing pages — drift hurts SEO too. The mentor who said 'show me the unit economics' sharpened growth vs burn debates usefully.
Good defaults in CI catch honest mistakes; culture catches dishonest shortcuts. The consultant was right about boundaries; we were just allergic to the word no. Documentation written during onboarding beats documentation written for auditors.
We learned that small honest delays beat big dishonest surprises for partnerships. The integration that validated idempotency keys on payouts prevented duplicate payouts quietly forever. The database migration was fine; the application assumptions were not.
Join the conversation.
Log in to reply