Loyalty That Actually Captures

The default loyalty program is a margin leak

Spend €100, earn 100 points, redeem 500 points for €5 off. This is the mechanic we found in 71% of the DTC brands we audited. It is also the single most reliable predictor of falling gross margin over 24 months. The program does not create retention. It gives away margin to customers who would have repurchased anyway.

A loyalty program is not a rewards catalog. It is a mechanism that produces first-party data at every interaction and turns it into higher-margin repeat revenue. If the mechanic does not produce data, and the data does not change the next order, the program is a tax on gross margin.

Four archetypes that actually work

Archetype 1. Tiered access

Status unlocks utility, not points. Early access to drops, priority support, member-only sizes, invitation-only events. The mechanic captures preference and intent, and the brand can price above market because access is the reward.

Archetype 2. Community pricing

Members see different prices, not discounts. The difference is subtle and structural. A discount signals that the sticker price is negotiable. A member price signals that the brand has a distinct commercial relationship with a defined audience. Margin is preserved because the anchor moves.

Archetype 3. Progress-based

The program tracks a journey, not transactions. Skincare brands, coffee brands, and supplement brands map particularly well. The customer is rewarded for completing a protocol, not for spending. Data captured is behavioural and predictive, and the program feeds directly into replenishment.

Archetype 4. Hybrid retail

Staff-assisted enrollment, online activation. Retail becomes the acquisition channel for the digital program, and the digital program becomes the retention engine for retail. The mechanic works because both surfaces contribute data and both surfaces benefit.

The anti-pattern to retire

Points-per-euro with a redemption ladder. It is universally the worst-performing mechanic on both retention and margin. If you have one live, the first change is to instrument it against a holdout and prove it does not lift. In every audit we have run, it did not.

How Loiale designs loyalty

We start with the data the program should generate, and work backwards to the mechanic. Then we build the mechanic inside our own loyalty platform, or as an ad-hoc module inside the brand's existing stack. The point is that the mechanic serves the data model, not the other way round. That is the entire trick.