Who should attend the weekly growth review?

Everyone who owns a number and whoever controls the budget. At FX Replay that meant the channel owners across paid, lifecycle, creative, and affiliates, plus me. Keep it under ten people. Observers who neither own a number nor decide anything turn a working meeting into a presentation.

What belongs on the shared dashboard?

The handful of numbers the business actually runs on: revenue, new customers, cost per acquisition by channel, and the pipeline of experiments in flight. Built on the same tracked events finance recognizes. If a number cannot survive a finance audit, it does not belong on the dashboard.

How is this different from a normal marketing status meeting?

Status is read before the meeting, from the dashboard. The meeting itself spends its minutes on decisions: which experiments concluded, what the rules say happens next, and what ships this week. If nothing was decided, it was a status meeting wearing a different name.

What if we only run one or two experiments at a time?

The rhythm matters more than the volume. A small team reviewing two experiments with honest thresholds compounds knowledge every single week. Velocity grows naturally once concluding tests becomes routine instead of an event.

How do I stop the meeting from relitigating old decisions?

Write decisions down with their reasoning in the experiment log, including what evidence would reopen the question. When the topic resurfaces, the log answers it in thirty seconds. Reopening requires new evidence, and a louder opinion than last time is not evidence.

The weekly growth review: one dashboard

Audits find the problems and fixes solve them, once. The weekly review is what stops everything from quietly breaking again.

Every growth system I have handed off came with the same warning: the fixes decay. Tracking drifts as engineers ship. Winning ads fatigue. A budget someone set in March is still running in August because nobody owns the question.

The difference between teams that compound and teams that repeat themselves is rarely talent. It is rhythm: a fixed weekly moment where the numbers get faced and decisions get made.

This is the format I install after the 30-day audit wraps, and the one that ran the growth machine at FX Replay. Dozens of experiments moved through it.

One dashboard everyone believes

The meeting has a precondition. There is one dashboard, built on tracked events that finance recognizes, and everyone in the room treats it as the score.

Most growth meetings fail here before anyone speaks. Marketing brings platform numbers, finance brings billing numbers, and the first twenty minutes disappear into reconciling them. When the numbers are contested, every decision downstream is contested too.

Getting to one trusted dashboard is mostly a measurement problem, and it is exactly what fixing your tracking buys you. The meeting is where that investment pays out, week after week.

Status lives in the dashboard before the meeting starts. The meeting spends its minutes on decisions.

Block one: numbers against plan

Ten minutes, and only on deltas. Where are we against what we said would happen, and what explains the gaps.

The discipline is having a plan number at all. A dashboard without expectations is scenery: numbers go up and down and everyone nods. Against a plan, the same numbers become questions with owners.

Block two: experiment readouts

Fifteen minutes for tests that concluded this week. Each readout is three sentences: what we bet, what happened, what we do about it. The learning goes into the log before the meeting ends.

Tests still running get a line on the dashboard and no airtime. Narrating half-finished experiments invites deciding on partial data, which is how teams ship losers early and kill winners before they mature.

Block three: kill and scale

Ten minutes to move money. Losers below the threshold stop. Winners above it get more budget. The thresholds were written down when each test launched, so the meeting applies rules rather than negotiating them.

This is the block that separates the format from a status meeting. Attention is cheap. Reallocated budget is what makes the review an operating tool.

Block four: next week's ships

Ten minutes on what launches next: which tests, which creative, which fixes, each with an owner and a date. This block feeds block two, two weeks from now. A review with nothing shipping is a book club.

The two artifacts

Only two documents survive the meeting, and both stay short:

The experiment log: one line per concluded test. What ran, what won, why we think so, and what evidence would reopen the question. It is the team's institutional memory, and it stops old losers from being rebuilt next quarter.
The ship list: what launches this week, who owns it, what could block it. It is next week's accountability, written in advance.

The failure modes

Three patterns kill this meeting, and all three are recoverable:

Status theater: people narrating work instead of deciding. Fix: status moves to the dashboard, read before arrival.
Metric tourism: wandering through interesting numbers with no decision attached. Fix: if no action could follow from a number, it comes off the agenda.
Relitigating: reopening settled decisions because someone feels differently this week. Fix: the log records what evidence reopens a question. Feelings do not qualify.

Start it this Monday

You need less than you think: one dashboard everyone will accept for now, a 45-minute recurring invite, the four blocks, and an empty experiment log. The first few weeks feel thin. By week six the log has entries, the thresholds have teeth, and the meeting starts paying for itself.

Want help installing the dashboard and the rhythm behind it? Let's talk, or email me at karran@karrangupta.com.

The weekly growth review: one dashboard, four blocks, no debates