← Back to portfolio Email me →
Strategy · Summit Rx

Knowing exactly what a customer is worth.

Every growth decision comes back to one question: how much can we afford to pay for a customer? I built a customer-level unit-economics model, and a simple view of it, so the CEO can answer that in seconds and see which levers move it most.

In 20 seconds
  • A per-customer model: every cost a subscriber carries against everything they pay, across billing cycles.
  • A break-even grid that shows at a glance which combinations of acquisition cost and retention work.
  • Confidence intervals on campaign results, so budget moves follow real differences. Nine of 11 campaigns turned out to be statistically alike.
  • A 90-day plan built on two KPIs, focused on the highest-leverage levers.
RoleSolo: model, analysis, plan
TimelineSep 2026
AudienceCEO
Output90-day plan, 2 KPIs

The question

Ad platforms tell you what a customer costs. They don’t tell you what a customer is worth. For a subscription business the difference is everything, because value arrives month by month. The CEO needed one clear answer he could plan around.

How I answered it

Method 01 · Contribution model

Price the whole customer, not the first order.

For a subscriber I modelled product cost, shipping, payment fees, offer discounts, and acquisition cost, against the payments they make over their lifetime.

So: One honest number per customer, contribution after acquisition, that every other decision hangs off.

Method 02 · Scenarios

Typical customer and best-case customer, side by side.

Modelling both showed how much of the outcome depends on how long customers stay versus how cheaply they’re acquired.

So: It made clear that retention and offer design were the biggest levers, bigger than day-to-day ad tweaks.

Method 03 · Break-even grid

Show the trade-offs in one picture.

A grid of break-even retention across acquisition cost and product cost. The CEO can point at a cell and see what has to be true for it to work.

So: Planning conversations became "which two numbers do we move, and by how much?"

Method 04 · Signal vs noise

Don’t steer on noise.

With confidence intervals on cost per order, 9 of 11 campaigns were statistically indistinguishable, so reshuffling budget between them wouldn’t have changed much.

So: Consolidate, and put energy where it moves the result.

What changed

The model became a 90-day plan run on two KPIs, with actions grouped by lever:

1
customer-level model every decision hangs off
11
campaigns tested for signal vs noise
2
KPIs in a 90-day plan

Business figures (revenue, spend, conversion and retention rates) stay private at my employer’s request. Happy to walk through the real work on a call.

If I did it again

I'd build this model before the first ad dollar. Unit economics should set the acquisition budget, not explain it afterwards.