Over-discounting to high share-of-wallet customers eliminated
Operationally actionable segments adopted across marketing, commercial, and operations
Graduated reactivation ladders outperformed single deep-discount offers
Post-campaign ROI measurement introduced for the first time
A customer value segmentation and offer framework replaced broad, margin-destroying promotions for an African grocery retailer with segment-specific discount depths, stretch targets, and reactivation ladders calibrated to share-of-wallet. The work eliminated over-discounting to already-loyal customers, introduced post-campaign ROI measurement, and established a repeatable framework the business could run campaigns from quarter after quarter.
The Problem
An African grocery retailer was running promotions that routinely destroyed margin. There was no structured view of customer value: the business had transactional data but no way to estimate how much a customer could spend (total wallet), how much they currently spent, or where the gap sat. Without this, every promotional decision was made in the dark. A 15% discount to a customer already spending 80% of their grocery budget with the retailer is a very different proposition from the same discount to one spending 20%.
The Solution
We built a customer value segmentation and offer framework that gave the retailer, for the first time, a structured basis for deciding who to target, at what discount depth, and with what spend target. The goal was not a one-off analysis but a repeatable framework the business could run campaigns from, quarter after quarter.
The work began with wallet estimation: using basket composition signals, frequency patterns, and demographic benchmarks to approximate each customer's total grocery spend and the share the retailer currently captured. This share-of-wallet metric became the backbone of the segmentation. We then layered in behavioural dimensions (trajectory, shopping pattern, promotional sensitivity) to create segments that were both analytically meaningful and operationally usable. Each segment was given a plain-language label, a one-paragraph profile, and a recommended offer strategy, socialised with marketing, commercial, and operations teams before finalising.
With segments defined, we built offer rule tables specifying discount depth, stretch targets, and incentive ladders per segment. High share-of-wallet customers received shallower discounts (the business was not paying to reward behaviour already happening). Low share-of-wallet customers with growth potential received deeper incentives tied to stretch targets calibrated to their current spend. Lapsed customers received graduated reactivation ladders: a modest first incentive to trigger a return visit, a second offer conditional on repeat, and a third only at target spend. Every rule included margin guardrails to prevent edge-case losses. Post-campaign ROI measurement closed the loop, assessing incremental spend, margin impact, segment migration, and reactivation durability.
Results and Impact
| Metric | Outcome |
|---|---|
| Promo profitability | Improved by aligning discount depth and stretch targets to customer value |
| Margin leakage | Reduced: over-discounting to high share-of-wallet customers eliminated |
| Reactivation effectiveness | Graduated ladders outperformed single deep-discount offers |
| Segment adoption | Segments socialised and adopted across marketing, commercial, and operations |
| Campaign evaluation | Post-campaign ROI measurement introduced for the first time |
| Discount depth | Frequently lower than prior practice, once margin and habit effects were made visible |
The biggest single change was not a technical one. It was the visibility that came from linking discount decisions to customer value. When teams could see that a blanket discount was subsidising already-loyal customers (and doing nothing for those with genuine growth potential), the conversation about promotional strategy shifted permanently.
Key Takeaways
-
A segmentation is only useful if teams can run campaigns from it without extra translation. Each segment had a label, a profile, and a recommended offer. Marketing teams could brief campaigns from the documentation without needing an analyst in the room. Segmentations that require interpretation are segmentations that get ignored.
-
The "right" discount was often lower than people expected. Once the framework made margin and habit effects visible, many promotions turned out to be deeper than necessary. Well-targeted moderate discounts tied to stretch targets outperformed blanket deep cuts that rewarded existing behaviour.
-
Reactivation requires patience, not just generosity. Single deep-discount offers to lapsed customers produced short-term visits that rarely sustained. The graduated incentive ladder generated lower initial response but significantly better long-term reactivation rates.