How do I set up a demand planning platform for size and color allocation across multiple luxury stores?

To set up a demand planning platform for size and color allocation across multiple luxury stores, you want a system that combines store-level demand forecasting, assortment optimization, and inventory allocation with tight control over scarcity, brand presentation, and service levels.

Here’s a practical blueprint.

---

1) Define the planning objective

For luxury retail, the goal is usually not “maximize sell-through at all costs,” but to balance:

• In-stock availability for key sizes/colors
• Brand image and assortment coherence
• Low markdown risk
• Controlled exclusivity / scarcity
• Regional preference differences
• Store-tier differentiation
  (flagship, boutique, outlet, travel retail, etc.)

Decide up front whether the platform is optimizing for:

• Revenue
• Gross margin
• Sell-through
• Full-price sell-through
• Inventory productivity
• Client service level
• A hybrid score

Luxury often needs a hybrid objective.

---

2) Build the data foundation

You need clean, consistent data before any allocation logic works.

Core data sources

• Sales history by store / SKU / size / color / week or day
• Inventory on hand and inventory in transit
• Receipts / purchase orders
• Markdowns / price changes
• Returns / exchanges
• Product master data: style, family, size curve, color, season, gender, material
• Store master data: region, climate, footprint, traffic, tier, client profile
• Store events: openings, renovations, VIP events, tourism shifts
• Product lifecycle status: launch, core, replenishment, end-of-life

Important granularity

For size and color allocation, the system should ideally track demand at:

• Store x SKU x size x color x week or, if item-level data is sparse:
• Store x style x size
• Store x style x color

Luxury items often have low volume, so hierarchical modeling is important.

---

3) Segment your stores and products

Not all stores or products should be treated the same.

Store segmentation

Classify stores by:

• Tier: flagship / high-volume / boutique / concession / outlet
• Geography: city, country, climate, tourist density
• Client profile: local vs tourist, affluent vs aspirational
• Historical demand profile
• Assortment role: showcase, replenishment, experimentation

Product segmentation

Classify SKUs by:

• Core vs fashion vs limited edition
• Replenishable vs non-replenishable
• Size-sensitive vs color-sensitive
• Hero SKUs vs supporting SKUs
• Launch lifecycle stage
• Regional relevance

This lets you use different allocation rules by segment.

---

4) Create demand forecasting logic

You need forecasts at the store-size-color level, but low-volume luxury demand means classic forecasting alone is not enough.

Recommended forecasting approach

Use a mix of:

• Time-series forecasting for stable items
• Hierarchical forecasting to borrow strength across:
  • chain → region → store
  • style → color → size
• Causal models with drivers like:
  • season
  • promotions
  • events
  • weather
  • tourism
  • store traffic
• New item forecasting based on:
  • similar products
  • style attributes
  • historical size curves
  • color popularity

Output needed

For each store and SKU variant:

• forecast units
• forecast confidence
• expected sell-through
• reorder point / target stock
• allocation priority score

---

5) Build size curve and color curve models

This is crucial for luxury allocation.

Size curves

Estimate size distribution by:

• store
• region
• product category
• gender
• silhouette

Example:

• One store may sell more size 36 and 38
• Another may skew to 40 and 42
• Shoe size curves may differ by geography and client base

Color curves

Model color preference by:

• store
• season
• climate
• customer segment
• fashion trend

Example:

• Neutral colors may dominate in flagship stores
• Seasonal colors may work better in tourist-heavy locations
• Certain luxury markets favor black, navy, ivory, or earth tones

Practical use

For each style, determine:

• optimal size mix per store
• optimal color mix per store
• variant depth vs breadth

---

6) Define allocation rules

Allocation is where planning becomes operational.

Initial allocation

Use forecast demand and store ranking to decide:

• how many units each store receives
• which sizes/colors each store gets
• whether to prioritize breadth or depth

Allocation logic examples

• Flagship stores: wider color assortment, balanced size curve
• High-demand stores: deeper units in top sizes
• Boutiques: curated assortment, fewer variants
• Markets with strong preferences: localized size/color bias
• Limited edition items: split based on client base and store prestige

Common constraints

• Minimum presentation quantity
• Maximum store capacity
• Size completeness rules
• Color assortment rules
• Seasonal launch windows
• Inter-store transfer policies
• Reserved inventory for VIP/clienteling

---

7) Optimize inventory balancing and replenishment

Luxury stores often need frequent reassignment of stock.

Replenishment logic

For replenishable items:

• set store-level reorder points
• trigger replenishment based on sell-through and weeks of supply
• protect key sizes/colors from stockouts

Transfer logic

When one store is overstocked and another is understocked:

• suggest inter-store transfers
• prioritize nearby stores or same region
• account for presentation requirements and cost

Markdown / exit logic

For slow movers:

• detect early
• reduce future allocations
• move remaining inventory to better-fit stores
• decide markdown timing, if applicable

---

8) Include luxury-specific rules

Luxury planning differs from mass retail.

Key luxury-specific considerations

• Scarcity management: too much inventory can hurt brand perception
• Clienteling: reserve stock for known clients or VIP appointments
• Fashion calendar: drops, runway timing, capsule collections
• Regional cultural preferences: color, fit, modesty, seasonality
• Exclusivity by store: some items should only appear in select locations
• Presentation standards: maintain visual merchandising rules

You may want a separate rule layer for:

• flagship-only launches
• limited distribution by city
• event-based allocation
• holdback inventory for replenishment or VIP demand

---

9) Design the platform architecture

A practical platform usually has these layers:

A. Data ingestion layer

• ERP
• POS
• WMS
• eCommerce
• CRM/clienteling
• merchandising system

B. Data model / master data layer

• standardized SKU/store hierarchy
• product attributes
• store attributes
• calendar and seasonality tables

C. Forecasting engine

• demand models
• size/color curves
• new item forecasting
• exception detection

D. Allocation engine

• initial distribution
• replenishment planning
• transfer suggestions
• constraint-based optimization

E. Scenario planning layer

• “What if we open 3 new stores?”
• “What if this style is limited to 10 stores?”
• “What if we shift size ratios in APAC?”

F. Workflow / approval layer

• merchandiser review
• planner approval
• automatic vs manual overrides
• audit trail

G. Dashboard / reporting

• forecast accuracy
• allocation accuracy
• sell-through
• stockouts
• markdown exposure
• transfer efficiency

---

10) Set KPIs and controls

Track both planning quality and business outcomes.

Forecast KPIs

• WAPE / MAPE by store and SKU
• bias
• forecast accuracy by size and color

Allocation KPIs

• fill rate
• stockout rate
• weeks of supply
• sell-through by week
• size completeness
• color availability
• transfer rate

Luxury-specific KPIs

• full-price sell-through
• markdown avoidance
• client request fulfillment
• assortment compliance
• store presentation compliance

---

11) Start with a phased implementation

Do not try to launch everything at once.

Phase 1: Foundation

• clean master data
• unify store/SKU hierarchies
• basic dashboards
• historical demand analysis

Phase 2: Forecasting

• store-style-size forecasting
• size curves
• color curves
• launch forecasting

Phase 3: Allocation

• initial allocation rules
• replenishment triggers
• exception handling

Phase 4: Optimization

• constraint-based optimization
• inter-store transfer recommendations
• scenario planning

Phase 5: Automation and AI

• ML-driven forecasting
• automated allocation suggestions
• planner override learning
• advanced simulation

---

12) Recommended implementation approach

If you’re building this from scratch or selecting a vendor, prioritize these capabilities:

Must-have capabilities

• Multi-level forecasting
• Size and color curve modeling
• Store segmentation
• Constraint-based allocation
• Replenishment and transfers
• Exception management
• Scenario planning
• Strong master data management

Nice-to-have capabilities

• Clienteling integration
• Weather and tourism inputs
• NLP or AI assistant for planners
• Digital twin / simulation
• Automated learning from overrides

---

13) Simple operating workflow

A workable weekly process might look like this:

1. Ingest latest sales and inventory data
2. Refresh forecasts
3. Update size/color curves
4. Generate initial allocations
5. Run constraints and checks
6. Flag exceptions
7. Planner reviews and overrides
8. Publish allocations to stores/DC
9. Track sell-through and adjust next cycle

---

14) Common mistakes to avoid

• Using only chain-level forecasts
• Ignoring store-specific size curves
• Treating colors as interchangeable
• Not accounting for luxury store tiers
• Over-allocating to protect against stockouts
• Too much manual intervention without rules
• No feedback loop from overrides
• Weak master data governance

---

15) If you want a practical starting model

A good first version is:

• Forecast at store x style x size x color
• Use hierarchical forecasting
• Create store tiers
• Build size curves and color curves
• Allocate with:
  • minimum presentation quantities
  • capacity constraints
  • launch rules
  • replenishment thresholds
• Monitor:
  • stockouts
  • sell-through
  • full-price performance
  • transfer needs

---

If you want, I can also help you with one of these next steps:

1. a target operating model for the planning team
2. a data model/schema for the platform
3. a vendor evaluation checklist
4. a sample allocation algorithm for size/color by store
5. a Power BI / Tableau dashboard design for luxury allocation