The most flattering report in marketing
Loyalty program reporting is almost designed to mislead. Members spend more than non-members, they buy more often, they stay longer, and every one of those statements was true before anyone joined. Programs recruit your best customers, and then the program is credited with the behavior that made them join.
Nothing in that reporting is fabricated. It is simply the wrong comparison, and it is the standard comparison. The result is that a great many loyalty programs have been running for years without anyone knowing whether they generate a return or quietly transfer margin to people who needed no incentive.
Members outspend non-members. So did they, before there was a program. That sentence has ended more loyalty business cases than any analysis I have run.
Holdouts, or you are guessing
The fix is straightforward and unpopular: withhold the program, or a mechanic within it, from a randomly assigned share of eligible customers, and read the result as the difference between the groups.
Random assignment is the part that matters. Comparing members to non-members measures who chose to enroll. Comparing enrolled customers to a randomized holdout measures what enrollment did. The gap between those two numbers is frequently large, and finding out how large is the single highest-value analysis available to most loyalty teams.
This is the same standard I apply to paid media, and there is no principled reason retention should be exempt from it. If a win-back campaign cannot show lift against a control, it is a report of customers who came back, not a program that brought them.
Design that changes a decision, not a price
Once the measurement is honest, program design gets more interesting, because the question becomes what would actually alter behavior.
Accumulated progress. Status, tiers, and stored value create a real cost to leaving. This works when the status is worth something, and becomes an expensive fiction when the benefits are cosmetic.
Personalization that improves with tenure. A relationship that genuinely gets better the longer it runs, better fit, better recommendations, faster service, is a switching cost that no competitor can match on day one.
Access rather than discount. Early access, exclusive assortment, and service-level benefits change the choice without moving the price, which is what protects the margin the program is spending.
Earn-and-burn discounting. The default, the easiest to launch, and the one most likely to be non-incremental. It has a specific long-term risk beyond the margin cost: it teaches a base that was not price sensitive to become price sensitive, and that is very hard to undo.
The choice among these should follow from what the churn decomposition says. A base leaving through never-activated churn does not need a tier structure. It needs an onboarding fix, and a loyalty program layered on top will spend margin on the customers who were already fine.
Win-back is the last resort, not the program
Win-back gets a disproportionate share of lifecycle attention because it is the easiest campaign to conceive: they left, ask them to come back. It is also the lowest-leverage moment in the entire lifecycle, because the decision was made weeks earlier and the only tool left is usually price.
Run it well anyway, for the customers earlier work did not catch:
- Time it to the customer, not the calendar. Lapse should be defined against each customer’s own established interval. A fixed day-60 trigger declares quarterly buyers lapsed while they are still perfectly active, and it reaches monthly buyers a month after they were gone.
- Split by reason. The drifted customer, the one who completed what they came for, and the one who was let down are three audiences. Sending all three the same discount insults the third and wastes margin on the second.
- Lead with something other than money. A real change since they left, a removed friction, an acknowledgment. Discount is the fallback, not the opening.
- Size the offer by expected contribution. A recovered customer who churns again in two cycles cannot justify a deep discount, and a program that recovers customers at negative contribution is buying its own metrics.
And hold out a share of every win-back audience, permanently. A meaningful percentage of lapsed customers return on their own. Without a control, that group is counted as a save every single time.
What the program is allowed to cost
A loyalty program’s cost is not the technology and the team. It is the margin given away, and that number belongs in the same conversation as the incremental revenue it produces. When both are on the table, the discussion changes from whether members like the program to whether the behavior it creates is worth what it spends, which is a question a CFO can engage with and a satisfaction score is not.
How I work this lane
I start by establishing what is actually incremental, which usually means building holdouts that did not previously exist and accepting a quieter number than the one currently being reported. Then the program gets designed around the churn mix rather than around a mechanic chosen in advance, and the margin cost gets stated in the same terms as the return. It is less comfortable than the usual loyalty deck. It is also the version that survives the year the CFO asks what the program is for.
Further reading
- Your churn is a payments problem more often than a loyalty problem
- Subscription churn is a pricing problem in disguise
- Your retention rate is lying if you don’t cohort it
Where this fits
Loyalty and win-back are what you run for the customers upstream work did not hold. Churn reduction is that upstream work, and subscription economics sets the ceiling on what any of it can afford to spend recovering a customer.
- Churn reduction →, decomposition, cohort curves, and the risk window
- Subscription economics →, payback, contribution, and what a recovered customer is worth
- ← Back to Lifecycle & Retention