How Should a Benefits Programme Measure Partner-Offer Success?
4 min read
Define who funds a partner offer, what each party can count and which claims need a holdout. A UAE dining example shows why redemptions are not ROI.
Start with two facts: who funds the benefit, and which party can observe each outcome. A programme can count eligible members, views and confirmed redemptions; a merchant can count its bills and visits. Neither can infer incremental sales or retention from redemptions alone. Claims of additional visits, spend or renewals need a comparison group.
What can each partner claim without a test?
For a merchant-funded discount, the merchant bears the markdown while the programme may provide discovery and member communication. For a programme-funded reward, the programme pays the incentive; for a co-funded offer, both contribute under their agreement. The same redemption must not be reported as a full acquisition success by each party.
Measure | Usually countable directly? | What it proves |
|---|---|---|
Eligible members | Yes, from programme rules | Potential reach |
Offer views or opens | Yes, where tracked | Discovery, not use |
Confirmed redemptions | Yes, if the merchant or programme records them | Use, not additional sales |
Cost per redemption | Yes, after funding split is known | Spending on the offer |
Additional visits or spend | Only with a credible comparison | Incremental commercial effect |
Retention lift | Only with a credible comparison and programme data | Possible member value |
This distinction is not pedantry. A Dutch household-panel study found a loyalty programme's estimated effect far smaller after accounting for members who were already more likely to buy. That study is not a UAE benchmark, but it shows why comparing members with non-members can overstate a programme's effect. Partner quality belongs in the scorecard too: research on coalition programmes found that a partner's service failure could affect loyalty to the wider programme (Schumann and colleagues, 2014).
What does a co-funded dining offer cost?
Consider an illustrative, before-VAT Dubai restaurant bill of AED 300. A 20% discount is AED 60. Assume a bank funds AED 30 and the merchant funds AED 30; this is a hypothetical split, not a claim about UAE bank contracts. The customer pays AED 240, the bank contributes AED 30, and the merchant receives AED 270. If the merchant's variable cost is AED 105, a genuinely additional visit contributes AED 165. An existing customer who would have paid full price costs the merchant AED 30 in lost contribution.
With 1,000 redemptions a month, the bank's direct offer spend is AED 30,000. The merchant's result depends on the share of visits the offer actually caused:
Additional share of redemptions | Merchant's monthly change in contribution |
|---|---|
10% | −AED 10,500 |
30% | +AED 28,500 |
50% | +AED 67,500 |
Using unrounded arithmetic, the merchant breaks even when about 15% of redemptions are additional. The bank's break-even cannot be calculated from the merchant's margin: it needs its own card-spend, cost and retention data. Both parties must also check the actual tax treatment of co-funded offers with an adviser.
What would a fair test look like?
A bank can randomly hold back the promotion message from a consented group of eligible members, then compare visits or card spend with those shown it. A merchant can compare total covers across matched weeks or branches and track full-price repeat visits. Agree the measurement window and funding split before launch. Share aggregate results that each party is entitled to use; do not exchange member-level information without the required permission. For smaller partners, show a range because the estimate may be noisy.
DiscGo's current data illustrates the boundary: it measures offer impressions and opens, but its redemption table had no records on 28 September 2026. DiscGo therefore cannot claim redemptions or incremental visits from its catalogue data.
Decision rule: Report reach, discovery, confirmed use and each party's cost honestly. Use “incremental,” “ROI” or “retention lift” only when a pre-agreed comparison supports the claim. An unused merchant-funded offer is not automatically a programme loss; a used offer is not automatically a merchant win.
Help customers find and save your venue
Show your offers while they decide, then make it easy to find you again.