Limited shelf, unlimited candidate SKUs. Which products go on the list and how much space does each get? The classical retail OR problem: assortment + shelf-space.
In plain words
Sound familiar?
- The store has 3000 SKUs but only 1800 slots; the buyer renegotiates 'out' and 'in' each month.
- An e-commerce category page lists 200 items; the top 20 drive revenue, the bottom 50 get 0-1 orders per month yet the panel changes weekly.
- A supplier brings a new product; the category manager picks the SKU to displace by intuition.
- Brand A asks for more shelf, brand B says 'my visibility dropped'; both offer listing fees.
- We say 'different stores need different lists' but maintaining 50 different lists for 50 stores is impossible.
- Season changeover means swapping 200 products; we decide between carrying over inventory or selling at discount.
- A customer doesn't find the wanted product on the shelf and switches to the adjacent brand (substitution); we can't measure this loss.
Why it matters
How it's solved
Technical depth
How it's solved
Technical depthOne-liner: When a customer can’t find what they want, they often switch to a nearby product (substitution) — so a dropped slow-mover doesn’t lose just its own sales, it can shrink the whole chain. Conversely, widening the best-seller’s shelf share lifts the entire category’s margin.
The operations research (a discipline that uses math and computing to solve business-decision problems) literature studies this under two headings: Assortment Planning Problem (APP) — which products are on the list; Shelf-Space Allocation Problem (SSAP) — how much space each listed product gets. Modern formulations solve them jointly (Joint Assortment + Shelf-Space). Solution in three stages:
1. Modeling. Per SKU: base demand (expected sales if shelved), space elasticity (sales-space relationship), substitution matrix (probability of switching from X to Y), margin, shelf volume (units per facing), supplier constraints (minimum order, contract). Per store/category: total shelf area, minimum SKU count, brand-variety rules.
2. Solver-driven decision. Single store / single category (50-500 candidate SKUs): MIP or CP formulation solves in reasonable time. Multi-store / multi-category: Lagrangean decomposition or heuristics (greedy + local search). Assortment selection maximizing expected revenue under a multinomial logit (MNL — a method that models the probability of which product a customer picks among several) demand model. Output: SKU list + per-SKU facing count.
3. Field integration. Output becomes planograms (shelf layout drawings) and assortment lists delivered to the store manager and category manager. If store-level variation is supported, separate output per store. POS data closes the feedback loop weekly — actual vs. predicted, model re-calibrated. On e-commerce, listing order and category page layout auto-update."
Alternatives
Manual + spreadsheet + category manager's experience
FreeFree
Who it fits: Single store or 2-3 categories, small SKU count
- + Zero cost
- + Local sensitivity — manager knows the store
- + Fast, flexible decisions
- − Unmanageable beyond 5 stores and 500 SKUs
- − No substitution modeling
- − Supplier pressure (slotting fee) distorts decisions
- − Knowledge in the individual, lost on turnover
Local category management software
Enterprise25,000–100,000 TRY license + 5,000–15,000 TRY/year maintenance (TR market observation)
Who it fits: Mid-size retailer (10-50 stores), planogram-focused
- + Planogram visualization + listing tools
- + Local-language support, POS integration
- + Quick implementation
- − No real optimum algorithm, rule-based
- − Weak substitution / space elasticity support
- − Store-level variation hard
International category management + assortment optimization software
Enterprise30,000–250,000 EUR license + 10,000–40,000 EUR/year
Who it fits: Large retail chain, multi-category e-commerce
- + MNL demand model, MIP/CP optimization
- + Store clustering + localized listing
- + A/B test integration
- − Expensive
- − 6-12 month deployment
- − Depends on POS data quality
Open-source solvers + custom application
Open SourceLicense free; in-house build 12-20 weeks or 350K-1M TRY consulting
Who it fits: E-commerce / retailer with a technology team and custom category rules
- + No license, full customization
- + Direct POS data integration
- + MNL and choice-modeling libraries are open-source
- − In-house data scientist + OR capacity required
- − Very dependent on data quality
- − Ongoing maintenance
Recommendation
Ask in the meeting
- Do you model the substitution effect (probability of switching from X to Y if X is missing)? Do you use MNL or a similar choice model?
- Is space elasticity (space → sales relationship) configurable?
- Do you cluster stores and propose a different assortment per cluster?
- Is the planogram delivered to the store via screen/printer?
- Do you support A/B testing? How are pilot store groups selected?
- Can supplier agreements (minimum SKU, slotting fee, exclusivity) be entered as constraints?
- How does the POS integration work? Is weekly data flow automatic?
- If we stop working with you, how do we get SKU master data, assortment history, and planogram data back?
Technical details
Editor’s note
Colloquially this is called “category management”, “assortment decision”, or “planogram”. Academically there are two distinct problems: Assortment Planning Problem (APP) and Shelf-Space Allocation Problem (SSAP). Modern formulations solve them jointly. Without knowing they are distinct, you’d accept “we do planograms” as enough — what you actually need is substitution-aware assortment optimization.
The most-skipped point in the industry: the substitution effect. When you remove a product, not all its customers are lost — some switch to the neighbor brand, some leave. Does the software model this transition matrix or just use “average sales”? The gap is typically 30%+ in forecast error.
Step-by-step path — for the SMB
Stage 1 — Measure first, plan after. At least 12 weeks of POS data: SKU-level daily/weekly sales, category revenue, per-store product distribution, stockout duration. If loyalty data exists, basket analysis (what’s bought together).
Stage 2 — Extract the knowledge asset. SKU master data (size, margin, supplier constraints), substitution observations (what got bought during stockout), supplier agreements, category targets (growth, margin, variety).
Stage 3 — Pilot. 8-12 weeks, 1-3 categories and 5-10 stores. Success criteria defined in advance: category margin +3% minimum, stockout duration -20% minimum, dead-stock SKU ratio -40% minimum.
Stage 4 — Roll-out. 3-6 months across all categories and stores. Monthly review — actual vs. predicted, new listing decisions.
Risks — what can go wrong
- Data quality. POS data missing (stockout periods show “zero sales” but real demand exists) → forecasts wrong. Fix: stockout filtering, lost-sales reconstruction.
- Supplier pressure. Brand A asks ’expand my shelf’, offers slotting fee — if the software output can’t withstand this pressure, it becomes meaningless. Supplier contract decisions must be entered as constraints.
- Substitution model miscalibration. MNL parameters need calibration; pilot must be validated with A/B testing.
- Single-vendor lock-in. SKU history and planogram data are critical knowledge assets. Standard-format export must be in the contract.
Solution method — a technical look
| Approach | Typical scale | Solve time | Guaranteed optimum? |
|---|---|---|---|
| MIP (joint assortment + space) | 50-200 candidate SKUs, single store | minutes | Yes (within limit) |
| MNL choice model + greedy assortment | 100-1000 candidate SKUs | minutes | No, 95%+ optimum |
| Lagrangean decomposition (multi-store) | 50-200 stores × 500 SKUs | hours | No, 95-98% |
| Rule-based (rank by margin × velocity) | Most local software | instant | No, 75-85% |
| ML-based (uplift modeling + A/B) | Large e-commerce | continuous | No, experimental |
Objective function choice:
- Objective 1 — Total expected margin: Standard.
- Objective 2 — Maximum revenue: When volume is the priority, not margin.
- Objective 3 — Category-size balancing: For variety.
- Objective 4 — Minimum stockout: When CX is the priority.
For multi-objective, weighted sum or hierarchical. Ideal: A/B testing for real-world measurement.
Academic references
See the sources field in the frontmatter.
Sources
- Corstjens, M. and Doyle, P. (1981). A model for optimizing retail space allocations. Management Science, 27(7), 822–833. Classical paper on shelf-space allocation.
- Kök, A. G., Fisher, M. L. and Vaidyanathan, R. (2008). Assortment planning: Review of literature and industry practice. Retail Supply Chain Management, 99–153. Modern assortment-planning survey.
- Talluri, K. and van Ryzin, G. (2004). Revenue management under a general discrete choice model of consumer behavior. Management Science, 50(1), 15–33. MNL-based assortment theory.
- Turkish Statistical Institute — Retail Sales Index (monthly). Reference for retail revenue trends in TR.
- Turkish CoHE (YÖK) Thesis Center — keywords: ‘assortment’ or ‘shelf’ — 25+ Turkish theses. tez.yok.gov.tr
Glossary
- Assortment Planning
- Selecting which products to carry in a retail store or e-commerce category, given limited shelf or page space.
- Shelf-Space Allocation
- Deciding how much shelf room (how many facings) each listed product gets, balancing demand and margin against space.
Related problems
Dozens of SKUs from the Same Supplier — At What Frequency Do I Order Each So Trucks and Stock Cost Are Minimum Together?
This page is for an SMB wholesaler, importer or manufacturer pulling 50-500 SKUs from the same supplier — typically across 3-15 main suppliers. The weekly question is the same: from this supplier, which products should ship today and which can wait until next week? If every product triggers its own order, the same supplier ends up sending three trucks and three customs filings a week; one truck, one customs filing and one setup cost can cover all of it if the groups are right. Manual planning falls apart past 50 SKUs: some weeks a half-empty truck, others three back-to-back orders — fixed costs come back as 5-15% of unit product cost.
How Much of Each Product Should I Order So I Neither Run Out Nor Pile Up?
Written for retail, e-commerce, or manufacturing SMBs managing 500-5,000 SKUs. If your purchasing team still places orders every month from a spreadsheet using last-three-months average plus ten percent, and yet you see empty shelves and clogged warehouses at the same time, you are in this problem. When the same company carries both stockouts (5-15% lost sales) and dead inventory (10-25% capital tied up), intuition has hit its limit: seasonality, lead-time uncertainty, and slow-moving SKUs cannot be corrected by hand item-by-item. Systematic forecasting paired with an inventory policy typically frees 1.8-5.4M TRY/year of working capital on 50M TRY revenue and lifts the service level from roughly 85% to 95%.
Stock Is Not Clearing as the Season Ends — When and How Deep Should I Mark Down So That Margin Holds and Dead Stock Stays Low?
This page is for you if you run a seasonal-collection fast-fashion retailer, a 30-100 store apparel/shoes/accessories chain, or a fresh-food chain (greengrocer, bakery, butcher). The classic pain: a 200-1,000 SKU collection sells for 4-12 weeks and then the unsold stock becomes a problem — discount too early and you erode the gross margin, discount too late and you sit on a pile of dead stock at the end of the season. The decision: which item, when, and by what percentage to mark down? In a mid-size fashion chain run on gut-feel rules, 20-35% of end-of-season stock ends up as salvage; a proper markdown schedule lifts the gross margin by 5-15%, cuts dead stock by 20-30%, and adds 15-60 million TRY of operating margin a year on a typical revenue base.