Evidence-Based Stage-Gate Review of a Retail Loyalty Programme
Halcyon Retail operates 480 stores and an e-commerce site. It has spent €12 million on a loyalty app and omnichannel fulfilment. The app has 1.8 million downloads and 34% monthly active users, but repeat purchase rate has fallen two percentage points and net promoter score is flat. Executive dashboards report downloads, social-media likes, and app-store rating; there is no benefits register. A 60-store pilot of a clienteling tool showed a 6% basket uplift where staff used it, but only 18% of associates adopted the tool. Store managers say the tool adds 12 minutes per shift and duplicates the CRM. The CFO questions the programme's value. A stage-gate review is scheduled in six weeks, and the transformation office must decide whether to scale, redesign, or stop the initiative.
1. Design a KPI and benefits framework for the loyalty and omnichannel programme, separating leading indicators from lagging outcomes and identifying vanity metrics to retire.
Read the worked answer
The framework should start from the intended outcomes: repeat purchase, customer lifetime value, basket size, and net promoter score. These are lagging and should be measured by cohort, channel, and store cluster, with a baseline and a counterfactual. Leading indicators should predict those outcomes and be actionable: monthly active users alone is insufficient; add 30-day repeat app usage, personalised offer redemption, clienteling interactions per associate, and omnichannel fulfilment accuracy. Operational indicators should include staff time per interaction and CRM data quality. Downloads, social-media likes, and app-store rating should be retired from executive dashboards because they do not demonstrate value and can rise while repeat purchase falls. A benefits register should assign each benefit an owner, baseline, target, measurement method, and realisation date. The framework should also include guardrails, such as margin impact and staff workload, so the programme does not optimise basket uplift by discounting or by overloading associates.
2. Assess the pilot evidence and recommend whether to scale, redesign, or stop the clienteling initiative, showing the reasoning behind the decision.
Read the worked answer
The pilot shows a credible signal but not a scalable case. A 6% basket uplift where the tool was used is meaningful, but 18% adoption means the observed effect applies to a small and possibly self-selected group. The tool adds 12 minutes per shift and duplicates the CRM, which explains low adoption and threatens sustainability. Scaling now would multiply a weak operating model; stopping would discard a real signal. The appropriate decision is redesign with a bounded scale test. Redesign should remove CRM duplication, embed clienteling in the existing workflow, reduce interaction time, and target associates with the highest customer-facing potential. The next gate should require at least 60% adoption in a 60-store test, a basket uplift of at least 3% at the store level including non-users, no increase in average handling time, and a positive margin after incentives. The CFO should see a costed benefits case with a stop rule if adoption or margin thresholds are missed.
3. Define the monitoring cadence and corrective-action triggers the transformation office should use after the stage-gate decision.
Read the worked answer
Monitoring should operate at three levels. Weekly, the delivery team tracks adoption, tool usage, data quality, and staff time; this is where corrective action is fastest. Monthly, a benefits review compares realised basket uplift, repeat purchase, and margin against the business case by cohort and store cluster; variances above an agreed threshold trigger a root-cause review. Quarterly, the transformation steering committee reviews the initiative against portfolio priorities and can reallocate funding. Corrective-action triggers should be explicit: adoption below 40% at month two, basket uplift below 2% at month three, or handling time above baseline by more than five minutes. Each trigger should have a named owner and a pre-approved response, such as workflow simplification, training, or pause. The office should also maintain a benefits register and a risk log, and report leading and lagging indicators side by side so the committee does not mistake activity for value. This cadence makes the next stage gate evidence-based rather than narrative.