OKR Check-In Templates · 7 min read

A Growth OKR Check-In Template That Tracks Shots on Goal, Not Just the Headline Metric

I've run growth teams for about a decade, and the check-in that works for them looks nothing like the one I use for a finance team. Growth lives and dies on shots on goal. If you only stare at activation rate every week, you'll panic at noise and miss the thing that actually predicts whether the number moves: how many experiments you're shipping and how many win. So this template asks about velocity first and the headline metric second. Here it is, the way I'd hand it to a lead on day one.

By Max Bondarenko · Last updated June 2026

The template, up front

No burying it. A growth check-in is a weekly answer to five questions, and four of them are about your pipeline of experiments, not your topline number. The number is a lagging signal. How many tests you ran, how many won, and what you learned is the leading one. Copy these five and ask them every week.

Why each question is there

Question one is the one most teams skip, and it's the one I'd never cut. Experiment velocity is the single best predictor I've found of whether a growth KR lands. If a team committed to 12 experiments a month and they're shipping 4, the activation number is almost irrelevant this week. You already know it won't move, because they aren't taking enough swings. I'd rather see a team at 10 shipped tests with a flat metric than 3 tests and a lucky bump. The first team is building a machine; the second got handed a coin flip.

Question two, the win-rate split, is where you catch a wrong target early. Say activation is supposed to climb from a 30% baseline to a 50% target in a quarter, and four weeks in your win rate on activation experiments is 1 in 9. That's not a motivation problem. That's the target telling you it was set wrong, or the lever you picked is dead. The whole point of asking this in week 4 is so you revise the target in week 4, not week 12 when it's a postmortem. I once watched a team grind for an entire quarter against a KR nobody believed by week three, purely because the check-in never asked whether the bet was still right. We'd burned eleven weeks defending a number instead of changing it. Now I put that question on the page so it can't be avoided.

Questions four and five protect throughput. A loss that teaches you something is a win you haven't cashed yet, so I make people say out loud what changed in the queue because of it. And blockers compound silently. If instrumentation has been broken for two weeks, your velocity number is a lie and you want to know now, not at the readout.

How to score it

Growth scoring isn't one color on the headline metric. I score the leading indicator (velocity and win rate) and the lagging one (the metric) together, then take the worse of the two. A team can be green on activation and still be at risk, because the gains came from one fluke test and the pipeline behind it is empty. Score the machine, not the lucky week.

How to actually run it

Async by default. Everyone answers the five questions in writing by Monday noon, with numbers, not adjectives. 'Velocity good' isn't an answer; '7 of a planned 12 shipped, 2 won, 1 inconclusive' is. The async write-up is the work. It forces the lead to actually pull the experiment log instead of vibing.

Then a live sync only when a question is red, and only the people who can unblock it. Fifteen minutes, growth lead plus whoever owns the blocker (usually eng or design), no spectators. If every answer is green, skip the meeting entirely and just react in the thread. The week I trust a team most is the week they cancel their own check-in because the doc said everything was fine. The check-in is where growth OKRs live or die, but the meeting is optional; the writing isn't.