Team rituals that stuck at one year
The three team rituals we kept after a year, the two we killed, and the one we reintroduced after thinking we'd retired it. A short field note on what a small company's process actually looks like.
We started the company with nine rituals on the calendar — standups, retros, reviews, all the textbook things. A year later, three stuck, two got killed, and one I was wrong to retire and had to reintroduce. This is the short version.
The three we kept
Monday morning triage (30 minutes). Every Monday at 9 AM Eastern, every person with an assigned pursuit or engineering initiative reads one sentence about their week and one about the biggest blocker. No demos, no updates-for-updates-sake. Ends at 9:30 sharp. Described in more detail in the Monday standup post.
This one stuck because it does exactly one thing: it surfaces misalignment before the week starts. When someone says “I’m blocked on X,” the rest of the call is two minutes of unblocking them or calendaring a separate conversation. Not thirty minutes of watching someone demo their in-progress work.
Saturday retro (20 minutes, async in a thread). Every Saturday morning, the team posts a three-part message: what went well, what was annoying, and one thing to try next week. No meeting. The proposal manager aggregates themes on Sunday evening and the ones that show up two weeks running become calendared action items. Here’s the long version.
This stuck because it’s async and it compounds. The action items that survive two-week repetition are the real systemic issues; the ones that appear once and vanish are usually situational and self-resolving. Async means nobody has to schedule around it and nobody performs for an audience.
Monthly “what are we wrong about” (60 minutes). Once a month, the founding team spends an hour explicitly trying to name what we’re currently wrong about. Not “what’s hard” — what we’re wrong about. Product positioning, pricing, engineering decisions, team design. The output is a list; the inputs are the things anyone has been noticing but hadn’t formalized.
This one stuck because a small company can drift confidently toward being wrong about something for six months before anyone surfaces it. The calendared 60 minutes makes disagreement low-friction.
The two we killed
Weekly all-hands. We ran a Friday all-hands for the first year. It started as a 30-minute update and ended as a 75-minute showcase. Demos got longer, context got lost, and the meeting became a thing to prepare for rather than a thing to learn from. We killed it at month 14. What replaced it: a written weekly update in the company channel, five bullet points, everyone reads it in three minutes.
Nobody has complained. A few people said it felt weird for the first month and then said it felt right.
Daily standup (for engineering specifically). We tried a daily engineering standup for the first six months. It was 15 minutes that consistently became 25 and then 35 as topics leaked. The actual coordination happened in async threads between standups. We killed the meeting at month six. Engineering runs on async updates in the dev channel, a weekly 30-minute sync for genuinely cross-cutting decisions, and ad-hoc pair calls when something’s actually blocked.
The result was more productive time and no loss of coordination.
The one we reintroduced
The Friday review of outgoing pursuits. I retired this at month 10 because I thought the Monday triage covered it and a separate Friday meeting was redundant.
I was wrong. The Friday review does a different thing than the Monday triage. Monday is look forward; Friday is decide the last 90 hours of the week. A pursuit that was “on track” on Monday and is “still on track” on Wednesday might be off track on Friday, and the Friday review catches it with enough weekend runway to fix.
I reintroduced it at month 14. It’s 15 minutes on Friday at 3 PM. Three questions per pursuit: will we submit Monday, what’s the blocker if not, do we need help over the weekend. Nothing else. No presentations, no status reports, no tangents. Described in the Saturday review field note, although that post is about the adjacent Saturday check, not this Friday one.
The pattern
Looking at the five surviving rituals (three kept, one reintroduced, plus the Monthly wrong-about call), they share a shape:
- Time-boxed hard. Every one ends when the clock says it does.
- A specific decision or surfacing job. Not a status-update broadcast.
- Low prep. You can show up with what’s in your head.
The ones we killed all violated at least one of those: the all-hands had high prep and no decision, the daily standup had vague surfacing and no hard time-box.
The takeaway
Process at a small company compounds if you ruthlessly kill the rituals that don’t do work, and if you reintroduce the ones you were wrong to retire. The calendar is a statement of what you think is load-bearing. It should update.