The funnel for a repeat purchase business is a loop

4 days ago
3 min read

Updated: 3 days ago

Introduction

The standard funnel narrows to a point at the first order, which suits a business selling something once. For a café, a supplier, a salon, a garage or anyone whose customers come back, that shape describes a small fraction of the revenue and hides the rest.

In these businesses the first order is not the objective; it is the entry cost. The money is in the second, fifth and twentieth purchase, and the interesting question is not what proportion of enquirers buy but what proportion of first-time buyers return. A model that stops at acquisition will send the whole marketing budget to the least profitable part of the business.

1. The funnel for a repeat purchase business continues past the first order

The structural correction.

First purchase, second purchase, established customer, lapsed. These stages carry most of the value and are absent from the standard model entirely. Drawing them changes where the attention goes, because the diagram stops implying that the work finishes at the sale.

2. First to second purchase is the critical conversion

Where to concentrate.

The drop between the first and second order is almost always the largest in the whole model, and it is the one nobody measures. A customer who buys twice is dramatically more likely to keep buying. Whatever budget exists for improving conversion is usually better spent here than at the top.

3. Measure the repeat rate before anything else

The headline number.

What proportion of customers acquired six months ago have bought again? That single figure tells you more about the health of the business than any acquisition metric. Compute it for three separate cohorts so you can see the direction rather than a single point.

4. Know your natural purchase interval

Required for everything downstream.

Median days between orders, by customer type. Without it you cannot tell a lapsed customer from one who is simply not due, and every re-engagement effort is mistimed.

5. Define lapsed by the interval, not by the calendar

The practical definition.

If customers normally buy every six weeks, one who has not bought in twelve is lapsed. A fixed ninety-day rule applied across different product types will misclassify most of them.

6. Acquisition economics change completely

Where the model pays off.

If a customer is worth five purchases rather than one, you can afford five times more to acquire them. Most businesses in this position are underspending on acquisition because they price it against the first order.

7. Small retention improvements outweigh acquisition gains

The arithmetic that surprises people.

Raising the repeat rate a few points compounds across every cohort and costs far less than a comparable increase in new customers. Run both calculations before allocating the budget.

8. Track cohorts, not totals

The measurement discipline.

Group customers by the month they first bought and follow each group. Total revenue conceals whether new cohorts are behaving better or worse than old ones, which is the thing you most need to know.

9. Watch for the cohort that never returns

The early warning.

If customers acquired since a particular change repeat at a lower rate, something about that change is attracting the wrong people or damaging the first experience. This is only visible cohort by cohort.

Be careful about counting a repeat purchase as loyalty. Some repeat buying is convenience or lack of an alternative, and it disappears the moment either changes. The rate tells you what is happening rather than why.

Conclusion

Extend the model past the first order, because that is where most of the revenue is.

Treat first-to-second purchase as the critical conversion and measure it, establish your repeat rate at six months as the headline figure, work out the natural purchase interval and define lapsed against it rather than a fixed period, recalculate what you can afford to spend on acquisition using lifetime value, compare a few points of repeat rate against the equivalent acquisition increase, track cohorts by first-purchase month, and watch for a cohort that returns at a lower rate than its predecessors.