Is retention actually your growth constraint?

Every consultant in this category will tell you retention is undervalued. It usually is. But “undervalued” is not the same as “your binding constraint”, and the difference decides where the next two quarters of budget should go.

A constraint is the thing that, if you relaxed it, would let the system produce more. Most consumer businesses have three candidates — acquisition volume, conversion, and retention — and only one of them is actually binding at any given time. Working on the other two feels productive and moves nothing.

Here are the three tests I run in the first fortnight of an engagement, in the order I run them.

Test one: does the cohort curve ever flatten?

Take twelve months of customer-level data. Group customers by the month they were acquired. For each cohort, plot the share still active in month one, two, three, and so on.

You are looking for one thing: whether the curve flattens.

A curve that flattens has a floor. Some fraction of every cohort becomes durable, and everything above the floor is churn you might be able to influence. A curve that never flattens — one that keeps sloping toward zero at roughly the same rate in month 18 as in month 3 — is telling you something much more serious. It says you do not have a retention problem. You have a product-market fit problem wearing a retention costume, and no lifecycle program will fix it.

This distinction is the single most useful hour you can spend, and it is astonishing how rarely it has been done. I have walked into companies with a full CRM team, a six-figure messaging platform, and a churn-reduction OKR, where nobody had ever drawn this chart.

If the curve does not flatten, stop reading. Go and fix the product.

Test two: what does one point of churn cost?

If the curve flattens, price the problem.

Take your current monthly churn rate. Model what happens to your customer base over twelve months if it were one percentage point lower, holding acquisition constant. Then work out how many additional new customers you would have had to acquire to end up in the same place, and multiply by your fully loaded CAC.

That number is what one point of retention is worth to you this year. Write it on one line, in currency, and take it to your CFO.

Two things usually happen when you do this. The first is that the number is much larger than anyone in the marketing organisation expected, because compounding is unintuitive and nobody has ever run it. The second is that the conversation changes register — you stop asking for budget for a lifecycle program and start proposing a way to reduce the cost of growth, which is a thing finance already knows how to fund.

The arithmetic, run once

Illustrative. The three inputs below are invented to show the shape of the calculation — yours will be different, the method will not.

Take 100,000 customers, 3.0% monthly churn and a fully loaded CAC of €40, and hold acquisition constant.

Twelve months at 3.0% leaves 0.97¹² of the base, or 69,384 customers. One point lower, at 2.0%, leaves 0.98¹² — 78,472. The gap is 9,088 customers, and that gap is what you would have had to buy back to finish the year in the same place. At €40 each: €363,520.

That is the conservative reading. Replacements bought during the year churn too, and at an even monthly intake each one is only exposed to about half the year, so the gross intake needed is nearer 10,900 customers and the honest figure is nearer €436,000.

One percentage point. The reason to run it yourself, rather than take the shape of the argument on trust, is that the inputs are yours and the objections will be too.

If the number comes back small — and sometimes it genuinely does, in businesses with very cheap acquisition and short intended lifetimes — then retention is a margin problem, not a growth constraint. Treat it accordingly and spend the money on acquisition.

Test three: where does the marginal euro do more?

The first two tests establish that retention is worth something. The third establishes whether it is worth more than the alternative, which is the only comparison that matters.

Take the next hundred thousand of budget. Estimate the contribution it produces spent on acquisition at current CAC and current retention. Then estimate it spent on retention, using a deliberately conservative assumption about how much churn a well-run program can actually influence — I use a third of addressable churn as a working number, and I would rather be argued down than up.

Whichever produces more contribution is where the money goes. It is a crude comparison and it will not survive contact with a rigorous analyst, but it is dramatically better than the usual method, which is whichever function argued more confidently in the planning meeting.

Two things to be careful about here.

• Do not compare against your best-ever acquisition performance. Compare against what the next hundred thousand actually buys, which in most channels is worse than the average, because you have already spent the cheap money.
• Do not credit retention work with revenue it did not cause. The customers who respond best to a save flow are disproportionately the ones who were never going to leave. This is why holdouts are non-negotiable, and why any program that cannot be measured against a control should be treated as a hypothesis.

When the answer is yes

If all three tests point the same way, you have established something genuinely useful: not that retention matters — everyone agrees retention matters — but that it is the binding constraint right now, with a number attached and a comparison behind it.

That number is what buys you the mandate to do the structural work: contact strategy with a real budget, LTV tiering your finance team recognises, survival models that give you a time-to-event rather than a risk score, and a decision layer above your channels instead of five teams optimising their own open rates.

None of that is glamorous, all of it is downstream of having answered this question honestly first, and most of it needs someone whose job it actually is rather than a line on a marketing plan.

When the answer is no

Say so, loudly, and go and work on the actual constraint.

The most valuable thing a retention specialist can do in their first month is occasionally conclude that retention is not the problem — which is also the third of the three cases where hiring one is the wrong call. It costs a quarter of revenue to say it, and it is the reason the advice is worth anything the rest of the time.