You can’t advertise your way out of a problem you haven’t found yet.
Growth stalls in one place at a time. Acquisition, activation, conversion, retention, expansion: the brief almost always blames the first one. We go into the data and find out which it really is.
We start in the data, not the deck
Most revenue advice gets assembled out of interviews. Talk to the exec team, collect the theories, arrange them into slides. We do the interviews too. But interviews produce hypotheses, not evidence. The evidence is sitting in your data, and we go query it ourselves.
Then we cohort it, because averages hide the thing you’re looking for. A conversion rate that looks flat is usually one segment collapsing while another grows. Stable churn can be a healthy cohort covering for one that never activated at all. Reporting built to show a total will never show you either of those.
So the answer comes back specific enough to act on. Not “improve retention.” More like: the accounts from this one channel never reach the point where the product is worth paying for, and they’re gone by month two. Then you get a plan you can run with the team already on payroll.
There are five places it can be going wrong.
-
Acquisition
Are the right people arriving at all, what does each one cost, and which channels bring the ones who stay?
-
Activation
Do new users ever reach the moment the product becomes worth paying for? This is the one most often mistaken for an acquisition problem.
-
Conversion
Where does intent die between the first visit and the first payment, and is it the funnel, the pricing, or the promise?
-
Retention
Who leaves, when, and what did they have in common on the day they arrived?
-
Expansion
Does an account get more valuable over time, or is the day they signed up the best it ever gets?
Underneath all five sits a sixth question: whether your instrumentation can answer any of them. Often the first finding is that it can’t, and that fixing it costs less than another quarter of guessing.
Three shapes, depending on how deep you need us
Diagnostic
4–6 weeksWe get access to the data, work the five stages, and come back with which one is costing you the number and why. You get one prioritized set of recommendations scored on impact and effort, so it survives contact with your actual capacity. Scope is written down at the start; additions go through a change order, not a re-quote.
Advisory retainer
~5 hours a weekA standing seat beside the exec team. Reading the numbers with you each month, pressure-testing the plan before it reaches the board, and calling the hiring and agency decisions. For companies that have the operators and need the judgment.
Embedded
~10 hours a weekHands on the work. Building the cohort reporting and metric definitions the decisions depend on, running the experiments, sitting in the standups. For companies that know what to do and have nobody with the time to do it.
Four things worth knowing before you call
Findings are about the system
Never about people. Whoever owns the work we’re examining is a collaborator in week one, not a subject of the report. It’s also the only way findings survive after we leave.
Everything is scored on effort
A recommendation you can’t staff is not a recommendation. Impact and effort both get a number, so you can start on Monday.
You get the queries, not just the conclusions
The SQL, the metric definitions, and the cohort views leave with your team. If we found it once, you can watch it every month without us.
Small by design
We take a limited number of engagements at a time, and every one is run by a principal. There are no juniors to hand it down to and no bench to keep busy.
Four reasons to call someone else
- 01You’re pre-launch, with no customers and no data yet. We work from evidence. If there isn’t any, we’re the wrong spend. Come back when there is.
- 02You want an agency to run the campaigns. We’ll tell you what to brief them on; we don’t execute the media.
- 03You need someone full-time. This is deliberately part-time, senior work. If the role is a headcount, hire the headcount.
- 04You want validation for a plan that’s already been decided. We’re expensive for that, and you won’t like the invoice.
If you can’t name the stage costing you the number, that’s the conversation.
Tell us what the number is doing and what you think is causing it. Thirty minutes is usually enough to know whether there’s an engagement here.
hello@cohortstrategy.com