Learn · Should you run a loyalty program at all
Does your business need a loyalty program
A loyalty program can pay for itself only when customers buy often enough to keep earning, margins are wide enough to fund a visible reward, transactions can be tied to an identity, and you do not already hold most of each customer's category spend. Fail any one of those and the economics cannot close.
A loyalty program can pay for itself only under conditions that are checkable before you build anything. Customers must buy frequently enough for earning to compound, margins must be wide enough to fund a reward a member can feel, transactions must be attributable to an identified person, and there must be spend left in the category for changed behaviour to capture. This lesson is the checklist; the next one turns two of the checks into arithmetic.
The underlying identity is blunt. A program earns its keep when the extra margin from behaviour it actually changed exceeds what the rewards and the operation cost. Each condition below is a way one term of that inequality goes to zero.
## Earning has to compound, and frequency is what compounds it
The mechanism of points is accumulation toward something wanted. A customer who buys weekly watches a balance grow and forms the habit the whole model depends on. A customer who buys a mattress this year and another seven years from now earns once, then forgets the program exists long before the second earning event arrives. Long-cycle categories can still invest in service quality and a good CRM; a points ledger adds nothing to a relationship built on two transactions a decade.
## The reward comes out of gross margin
Identical generosity costs different businesses wildly different amounts. Return 1 percent of spend in a business with 20 percent gross margins and rewards consume a twentieth of gross margin, a defensible experiment. Return the same 1 percent at a 5 percent grocery margin and the program takes a fifth of gross margin, which demands a large and measurable change in behaviour before it breaks even. Thin-margin businesses do run programs successfully, and they manage it with partner funding, or with rewards whose cost to the operator sits well below their face value. What they never manage is cash-equivalent generosity paid from their own margin.
## No identity, no ledger
A program is an account, and an account needs every qualifying transaction attached to a person. If your sales flow through anonymous channels with no natural capture point, the program's first job becomes changing how people check out, a far bigger intervention than the rewards themselves. The identification requirement also reveals the program's second product: the purchase history, which for some retailers ends up worth more than the retention effect it was built to buy.
## Loyalty moves share, and some share has to be missing
A program redirects category spend the customer was already making somewhere. It cannot make the category bigger. If you already capture nearly all of your customers' spending, rewards subsidise purchases that were coming anyway, and the program is a discount wearing a membership card. The customers worth paying for are the ones splitting their wallet between you and a competitor, because theirs is the behaviour with room to change.
Run the four checks honestly and accept that the outcome may be no. That result is this module working as intended, and later lessons cover what to do instead. Leave with two numbers written down, your median repurchase interval and your gross margin percentage. The next lesson stress-tests both.