DTC loyalty program economics
A loyalty program is an economic system, not a signup target. Judge it on reward liability, contribution margin per member order, incremental repeat behaviour against a comparable non-member cohort, and the data it captures that nothing else in the stack can. Enrolment counts prove nothing on their own.
A loyalty program is a financial product with a brand face. It only works if the economics compound, not if the discount budget leaks.
The economics checklist
- What is the program's reward liability at any given moment?
- How does the program change contribution margin per member order?
- What share of incremental repeat behaviour is caused by the program versus would have happened anyway?
- What data does the program capture that the rest of the stack cannot capture otherwise?
- What is the cost of operating the program monthly, including tooling, staff and rewards?
Points vs tiers vs member pricing
Points are flexible but create a liability. Tiers create status but require a clear progression. Member pricing is simple but can train the whole base to wait for the discount. Each mechanic has a different margin and data profile.
Reward liability
Outstanding points or rewards are a balance-sheet item. We model the breakage rate, the expected redemption curve and the cash impact if redemption suddenly spikes.
Discount dependency
If the program's main lever is a discount, the brand may be teaching customers to buy only when rewarded. We check the share of member revenue that is discount-driven and compare it to non-member cohorts.
Contribution margin impact
Incremental revenue is not enough. We subtract COGS, payment fees, reward cost and incremental service cost to see if the program actually makes the business more profitable.
Data capture value
Loyalty is also a capture mechanism. The value of a known profile, with consent, across channels and over time, can exceed the direct margin of the program. We model it, not assume it.
Decision matrix
| Question | Launch now | Fix first | Do not launch |
|---|---|---|---|
| Repeat purchase rate under 20%? | Fix lifecycle and capture | Loyalty will subsidise one-time buyers | |
| Data layer cannot identify members across channels? | Fix identity | Program will be a silo | |
| Margin can absorb reward cost and still beat baseline? | Good signal | ||
| Program can capture zero-party data you cannot get elsewhere? | Good signal |
Reducing discount dependency
Discount dependency is usually a timing and mechanics problem rather than a pricing one. If the only lever available at the second-order moment is money, margin will keep leaking.
- Replace blanket codes with member pricing or access, so the benefit is conditional on identity.
- Move the incentive earlier, to capture, and later, to reactivation, rather than sitting on every send.
- Give the team non-price levers: replenishment timing, bundles, education, early access, service.
- Track discount depth per cohort next to contribution margin, so the trade is visible monthly.
- Withdraw an incentive in a comparable cohort before assuming it is what drives the repeat order.
Points, tiers or member pricing
| Mechanic | Works when | Main risk |
|---|---|---|
| Points | High purchase frequency, low basket variance | Liability grows quietly; becomes a delayed discount |
| Tiers | Clear spend distribution and a status motive | Top tier unreachable, so most members disengage |
| Member pricing | Strong brand and repeat consumables | Trains customers never to buy at full price |
| Access and services | Category with community or scarcity | Perceived value is hard to sustain |
Model the economics of your program
We build and run lifecycle, loyalty and customer-data systems inside your existing stack. Tool-agnostic. Outcome-owned.
Request a call