InoGen

Loyalty Scheme Structuring

Cohort-Level Analysis to Separate Incremental Value from Wasted Discount
Retail
Analytics
Marketing
£6M value delivered

£6M

annual revenue impact identified

Significant discount waste exposed via matched control methodology

4

CLV-based loyalty segments defined for targeted investment

Early lapse patterns detected, triggering programme redesign

Cohort-level incrementality analysis revealed that a significant portion of a retailer's loyalty discounts were subsidising existing demand rather than driving new spend, unlocking £6M in annual revenue impact. Using matched control groups and pre/post enrolment comparison, the work replaced assumption-based reporting with causal evidence and a CLV-based segmentation framework that redirected loyalty investment toward high-return customer segments.

The Problem

An international stationery retailer had been running a loyalty programme for several years. Enrolment numbers and redemption rates looked healthy. But leadership had a growing suspicion: was the scheme actually driving incremental spend, or was it simply discounting purchases that customers would have made anyway?

The Solution

We designed and delivered an analytical framework to isolate the true incremental impact of the loyalty programme at both cohort and individual level.

The foundation was a unified customer view joining transaction data, loyalty programme data, and product-level margin information. Customers were grouped into cohorts by enrolment date, enabling us to track how behaviour evolved over time: how quickly spend ramped up, whether frequency increased, and at what point engagement began to decay. Several cohorts showed a sharp drop in activity within three months, suggesting sign-up incentives were attracting transient interest rather than lasting behavioural change.

For each member, we compared spending before and after enrolment (controlling for individual baselines) and against matched control groups of non-members with similar profiles and purchase histories (controlling for seasonal and market effects). This dual approach separated genuine loyalty-driven uplift from both selection bias and background trends. Critically, we calculated incremental margin, not just revenue, net of all programme costs. This gave the business its first honest answer to whether the scheme was making money or draining it.

The loyalty base was then segmented by customer lifetime value, factoring in purchase frequency, margin per transaction, retention, and redemption behaviour. Four distinct segments emerged: high-CLV customers who rarely redeem (naturally loyal, broad discounts unnecessary), high-CLV active redeemers (justify targeted premium treatment), low-CLV high-redeemers (consuming margin for minimal return), and passive low-CLV members (neither costly nor contributory). This framework gave the business a clear basis for reallocating loyalty spend toward the segments where it would generate genuine uplift.

Loading diagram...

Results and Impact

MetricValue
Annual revenue impact£6,000,000
Key findingSignificant portion of discounts subsidising existing demand rather than driving new spend
Segmentation outcomeFour distinct CLV-based loyalty segments identified
Strategic changeReallocation of loyalty investment toward high-incrementality segments
Margin recoveryReduction in untargeted discounting to naturally loyal, high-CLV customers
Retention insightEarly lapse patterns identified in multiple enrolment cohorts, triggering programme redesign

Key Takeaways

  • Incremental margin is the only metric that matters for loyalty evaluation. Reporting total member spend without controlling for selection bias and programme costs creates a misleading picture. The shift to margin-level incrementality exposed where discounts were being wasted on customers who would have spent anyway.

  • Self-selection is the silent confounder. Customers who join loyalty programmes tend to be higher spenders to begin with. Any comparison of members versus non-members that does not account for this will overstate programme impact. Matched control groups and pre/post analysis are essential to separate causation from correlation.

  • CLV-based segmentation determines where loyalty investment should go. Treating all members identically is inefficient. Segmenting by lifetime value and redemption behaviour allowed the business to concentrate spend where it would generate genuine uplift, leading directly to a reallocation of loyalty budget worth millions annually.