Boosting Retail ROI: Data-Driven Promotional Pricing Optimisation

QuantSpark built a promotional analytics platform for a PE-backed European retailer with €1.3bn in annual turnover, replacing promotional decisions made on experience and precedent with a standardised, data-driven view of what was actually working. The headline output was a modelled opportunity of €3.4m in promotional margin enhancement, identified through optimised strategy and the systematic elimination of unprofitable promotions. That figure is an indicative finding from the analysis rather than an audited, realised saving, but it shows the scale of value that was hidden inside a promotional calendar the retailer previously could not see clearly.

The problem ran deeper than reporting. The retailer had no visibility or tracking of promotional performance across its product range, so patterns and trends went unnoticed and the business could not tell which promotions generated genuine incremental sales versus which simply shifted the timing of purchases customers would have made anyway. There was no standardised framework for comparing different promotion types, price cuts, multi-buy offers and cashback schemes, against each other, so the business could not say with confidence which mechanism worked best for a given product or category. Most critically, it had limited visibility into cross-product effects: cannibalisation of other lines and halo uplift on complementary items sat entirely outside its data. Without that visibility, a promotion that looked successful in isolation could be quietly loss-making once its true cost, including the drag it created elsewhere in the range, was properly accounted for.

QuantSpark's response, delivered working embedded within the retailer's team, started with the data foundation: a comprehensive promotional dataset built at transaction level across stores, capturing promotion type, discount depth, product category, store characteristics and baseline sales and margin. Feature engineering specifically captured cannibalisation and halo effects, the two cross-product dynamics the retailer had never been able to measure. On top of that dataset sat a four-KPI analytics framework: Sales Uplift % (incremental volume driven by the promotion), Margin Uplift % (net margin impact after promotional cost), ROI (a return calculation that nets incremental margin against cost and adjusts for halo and cannibalisation), and Customer Base Penetration (whether a promotion won new customers or simply increased basket frequency among existing ones). QuantSpark then cut this analysis by store size, region and competitive effects to surface patterns that had previously been invisible to the business, and packaged the whole framework into an interactive PowerBI dashboard that let commercial and buying teams explore performance by time period, product category and store segment, separating price effects from volume effects and immediate promotional impact from longer-term customer behaviour.

The workflow ran from raw transaction data to a testable business case in four stages: ingest and engineer the promotional dataset with cannibalisation and halo features; analyse it through the four-KPI framework cut by store, region and competitor context; visualise it in an interactive PowerBI dashboard built for self-serve exploration; then use that visibility to design and run structured tests. That last stage mattered as much as the analytics itself. The framework enabled systematic A/B testing across three variables the retailer had never tested in a controlled way: promotion mechanism (1+1 offers versus cashback versus straight price cuts), product category and supplier brand combinations, and leaflet composition (how many SKUs to feature to maximise engagement and basket penetration).

Applying that ROI calculation, incremental margin net of promotional cost, adjusted for halo and cannibalisation, across the existing promotional calendar and flagging consistently unprofitable promotions for elimination is what produced the €3.4m modelled margin-enhancement opportunity. It should be read as the size of the prize the analytics uncovered, not a confirmed year-one saving; no realised, audited figure or timeframe for capture was disclosed alongside it.

The foundation this created goes beyond the initial number. With visibility, standardisation and cross-product measurement now in place, QuantSpark and the retailer are scoping a predictive promotional costing tool so traders can forecast ROI before a promotion runs rather than after, and expanded halo and cannibalisation logic for more sophisticated cross-product insight, moving the business from measuring promotional performance towards actively engineering it.