Home / Case Studies / Retail Optimization

RetailAssortment & space optimisation

Transform shelf space into revenue with data-driven assortment and space optimisation.

Every inch of shelf space shapes what shoppers see and buy. We built a platform that decides what belongs on each store's shelves and how much space each product earns, then generates the planogram automatically.

+8.3%Revenue per store, $125K → $135K
89%Shelf utilisation, up from 71%
350 → 280SKUs, a leaner range per store
30 secPer planogram, down from 16 hours

Pilot results for an urban, high-income store cluster. Estimated annual impact: $2.5M additional revenue and $180K operational savings.

Industry context

Planogram tools show how shelves look. They don't decide what should be on them.

Managing assortment and shelf placement across hundreds of stores involves thousands of SKUs, varying layouts, regional preferences and vendor agreements.

Retailers use planogram and space management software such as NielsenIQ Spaceman, Blue Yonder Space Planning, RELEX, Oracle Retail Category Management, Aptos and SymphonyAI. These platforms offer shelf visualisation, drag-and-drop editing, category dashboards and merchandising rules.

They still rely on people to make the hard decisions: which SKUs each store should carry, how much space each product gets, and which configuration maximises revenue per square foot. So shelf planning runs on habit, manual analysis and trial and error.

Why revenue is left on the shelf

Five reasons existing tools fall short.

  • Manual assortment decisionsCategory managers evaluate thousands of SKUs across stores by hand.
  • No revenue optimisationSystems focus on visualising shelves, not on mathematically optimising outcomes.
  • Product substitution ignoredHow shoppers switch between products when one is missing is rarely modelled.
  • Static plansPlanograms are updated infrequently even as demand patterns change.
  • Human scale limitsLarge retailers manage millions of SKU-store combinations that no team can evaluate manually.
Our solution

An optimisation intelligence layer on top of your planogram system.

Instead of starting with manual assortment choices and then drawing shelves, the platform first works out the optimal assortment and space for each store, then generates planograms for store execution.

  1. Retail dataSales, stores, SKUs
  2. Store intelligenceClusters & segments
  3. Assortment optimisationKeep · Delist · Consult
  4. Space optimisationFacings per SKU
  5. Automated planogramReady for stores
  6. Revenue growthMeasured & iterated
Phase 1

Assortment optimisation

Decides which SKUs belong in which stores.

Stores are clustered by similar demand patterns and customer profiles. Within each cluster, every SKU is scored on sales performance, profitability, substitution relationships and strategic importance.

  • Inputs: store sales, customer demographics, store layout constraints, SKU attributes
KEEP

Strong-performing SKUs

DELIST

Underperforming SKUs

CONSULT

SKUs for business review

Phase 2

Space optimisation

Decides the optimal facings and shelf placement for each SKU.

  • Demand saturation curves
  • Substitution demand transfer
  • Shelf dimension constraints
  • Brand blocking rules
  • Refrigeration and category placement

Outputs: optimised shelf placement, facings per product, revenue projections and ready-to-use planogram layouts.

Pilot results

Store cluster: urban, high income.

MetricBeforeAfterChange
Revenue per store$125K$135K+8.3%
SKU count350280−20%
Shelf utilisation71%89%+18 pts
Stockout rate12%7.8%−4.2 pts
Planogram creation time16 hours30 seconds99.9% faster

5–12%

Expected revenue increase across store networks

15–25%

Improvement in shelf utilisation

70–80%

Reduction in planogram creation time, with 10–20% SKU rationalisation

Implementation

From data warehouse to store shelf in eight steps.

  1. Data ingestion from the retail data warehouse
  2. Store clustering and segmentation
  3. Assortment optimisation
  4. Category manager review of Consult SKUs
  5. Shelf space optimisation
  6. Planogram generation
  7. Store deployment
  8. Performance monitoring and iteration
Conclusion

Every inch of shelf space working for revenue.

By combining store clustering, assortment optimisation and space allocation, the platform turns shelf planning from a manual merchandising task into a data-driven revenue system.

  • Data-driven merchandising decisions
  • Store-specific assortments
  • Scales across large retail networks
  • Transparent decision logic category managers can review
Your business next

What decision could AI optimise for you?

Assortment, pricing, capacity, inventory or scheduling: if a decision has constraints and a measurable outcome, we can optimise it.