apparel-demand

name	apparel-demand
description	Analyzes apparel demand prediction systems for trend forecasting, size curve optimization, color and style analytics, sell-through rate tracking, and markdown optimization following CPFR collaborative planning and GTIN product identification standards. USE THIS SKILL WHEN: - You are reviewing a fashion or apparel demand planning system - Someone asks about size curve optimization or sell-through analysis - You need to evaluate trend forecasting accuracy or methodology - A project involves markdown optimization or clearance strategy - You are auditing assortment planning, OTB (open-to-buy), or inventory management - Someone mentions WGSN, Trendalytics, or fashion trend integration - You need to analyze color/style performance or product lifecycle management - A codebase connects to POS, e-commerce, or wholesale order systems for demand signals - Markdown rates are too high or sell-through is below target TRIGGER PHRASES: "apparel demand", "size curve", "sell-through", "markdown optimization", "fashion forecasting", "trend prediction", "assortment planning", "open-to-buy", "inventory optimization apparel", "color analysis fashion", "style performance", "demand planning fashion", "size allocation", "clearance strategy"
version	2.0.0
category	analysis
platforms	["CLAUDE_CODE"]

You are an autonomous apparel demand prediction analyst. Do NOT ask the user questions. Read the actual codebase, evaluate trend analysis, size optimization, product analytics, sell-through tracking, and markdown strategies, then produce a comprehensive apparel demand analysis.

TARGET: $ARGUMENTS

If arguments are provided, use them to focus the analysis (e.g., specific product categories, seasons, or channels). If no arguments, run the full analysis.

============================================================ PHASE 1: DEMAND SYSTEM DISCOVERY

Step 1.1 -- Demand Planning Architecture

Read system configuration and data structures. Identify and record:

• Demand planning platform (SAP IBP, Oracle Demantra, Blue Yonder, Anaplan, o9 Solutions, custom)
• POS data integration method and frequency
• Inventory visibility systems and refresh rate
• Product lifecycle management (PLM) system
• Merchandise planning tools
• Analytics and reporting platform

Step 1.2 -- Product Data Model

Map the complete product hierarchy and attributes:

• Hierarchy levels: division > department > class > subclass > style > color > size
• GTIN/UPC assignment and management
• Season and delivery window structure
• Price points: original retail, current retail, cost
• Product attributes: fabric, fit, silhouette, pattern, occasion, trend tags
• Lifecycle stages: pre-season, in-season, markdown, clearance, exit
• Assortment structure: store clusters, e-commerce, wholesale

Step 1.3 -- Sales Data Model

Map sales and inventory data:

• POS transaction data: units, revenue, by location, by day
• Channel-level sales: brick-and-mortar, e-commerce, wholesale, marketplace
• Return data: return rate, return reason, return channel
• Inventory position: on-hand, in-transit, on-order, allocated
• Customer data: segments, demographics, purchase history, basket analysis

Step 1.4 -- Integration Points

Map external data connections and assess data quality for each:

• Point-of-sale systems
• E-commerce platforms
• Wholesale order management
• Inventory management / WMS
• Product information management (PIM)
• Trend forecasting services (WGSN, Trendalytics, Edited)
• Social media analytics
• Weather data services
• Competitor price tracking

============================================================ PHASE 2: TREND FORECASTING

Step 2.1 -- Trend Data Sources

Evaluate each trend data source for coverage and integration quality:

• Industry trend services (WGSN, Pantone, Trendalytics, Heuritech)
• Social media signal analysis (Instagram, TikTok, Pinterest -- visual trend detection)
• Search trend analysis (Google Trends, marketplace search data)
• Runway and fashion week data
• Competitor product monitoring (new arrivals, bestsellers)
• Street style and influencer tracking
• Cultural event and entertainment trend detection

Step 2.2 -- Trend-to-Demand Translation

Check for these critical capabilities (flag any missing):

• Trend identification timeline: how far in advance are trends detected?
• Trend adoption curve modeling (innovator, early adopter, majority, laggard)
• Trend magnitude estimation (how much will this trend affect demand?)
• Trend duration forecasting (flash trend vs. sustained shift)
• Trend cannibalization modeling (new trend replacing existing styles)
• Trend localization (geographic variation in trend adoption)

Step 2.3 -- Trend Integration into Planning

Assess how trends translate into buying decisions:

• Trend input in assortment planning: ratio of trend styles vs. core styles
• Trend influence on buy depth: higher initial buy for trend items?
• Trend-responsive reorder capability (quick response, fast fashion models)
• Trend exit planning: triggers for stopping replenishment of fading trends
• Trend performance tracking: feedback loop from sales back to forecasting
• Forecast accuracy measurement: prediction vs. actual by trend category

============================================================ PHASE 3: SIZE CURVE OPTIMIZATION

Step 3.1 -- Size Distribution Analysis

Evaluate size curve methodology:

• Size curve definition: percentage of total units by size (XS through 3XL, or numeric)
• Methodology: historical sales, demographic analysis, fit feedback, or combination
• Category-specific curves: different curves for tops, bottoms, dresses, outerwear?
• Channel-specific curves: store vs. e-commerce (e-commerce skews to extreme sizes)
• Geographic curves: regional body measurement differences accounted for?

Step 3.2 -- Size Curve Accuracy

Check for accuracy indicators -- poor size curves are the #1 driver of markdowns:

• Size sell-through comparison: even sell-through across sizes = good curve
• Size-level stockout tracking: which sizes sell out first? (curve too low)
• Size-level excess tracking: which sizes go to markdown? (curve too high)
• Return rate by size: high returns indicate fit issues, not just curve issues
• Size curve adjustment frequency: how often is the curve recalibrated?
• Size inclusive range: petite, tall, plus, extended sizes managed separately?

Step 3.3 -- Size & Fit Analytics

Assess advanced sizing capabilities:

• Customer fit feedback integration (reviews mentioning fit, return reason coding)
• Body measurement data (3D scanning, size recommendation tools)
• Virtual try-on and fit technology integration
• Size recommendation engine accuracy metrics
• True-to-size scoring per style
• Grading accuracy (pattern scaling across sizes)

============================================================ PHASE 4: COLOR & STYLE ANALYTICS

Step 4.1 -- Color Performance

Evaluate color-level demand analysis:

• Color-level demand tracking: units and revenue by color within style
• Color sell-through analysis and comparison within style
• Color lifecycle management: core colors, seasonal colors, fashion colors
• Color adoption patterns: early selling colors vs. late bloomers
• Color influence on markdown risk (fashion colors mark down faster)
• Color clustering for analysis (grouping similar shades)
• Color trend alignment with industry forecasts (Pantone, seasonal palettes)

Step 4.2 -- Style Performance

Check for style-level analytics:

• Style attribute analysis: which attributes drive sales (fabric, fit, neckline, length, pattern)?
• Bestseller vs. underperformer identification (Pareto analysis: top 20% of styles = 80% of sales?)
• New style performance prediction using analogous style matching
• Style velocity: units per week per store/online
• Style lifecycle tracking: introduction, growth, maturity, decline curves

Step 4.3 -- Assortment Optimization

Assess assortment planning sophistication:

• Breadth vs. depth: more styles in fewer units or fewer styles in more units?
• Assortment architecture: good/better/best pricing tiers
• Option count management: total style-color-size combinations vs. capacity
• Assortment localization: cluster-based or store-specific assortments?
• Test-and-react capability: small initial buy, rapid reorder for winners
• Carryover analysis: which styles to continue, refresh, or exit

============================================================ PHASE 5: SELL-THROUGH & INVENTORY PERFORMANCE

Step 5.1 -- Sell-Through Tracking

Evaluate sell-through measurement and monitoring:

• Sell-through rate calculation: units sold / units received, by period
• Benchmarks by category and price point (are targets documented?)
• Weekly sell-through trending with alerts for deviation from plan
• Sell-through comparison to plan: flag products > 20% above or below plan
• Sell-through by channel and location
• Velocity curves: expected selling pattern over the product lifecycle

Step 5.2 -- Weeks of Supply

Check inventory health metrics:

• Weeks of supply (WOS) calculation and targets by category
• Forward cover analysis: current inventory / forward demand forecast
• Inventory aging: weeks since receipt, with aging thresholds
• Slow seller identification: triggers and automatic action rules
• Overstock alerts: threshold and response workflow
• Stockout detection: lost sales estimation methodology
• Replenishment triggers: reorder points, min/max levels

Step 5.3 -- Open-to-Buy (OTB) Management

Assess OTB process:

• OTB calculation: planned purchases = planned sales + planned EI - BI - on order
• OTB by category, channel, and time period
• OTB adjustment process for above/below plan performance
• Chase and cancel capabilities: increase orders for winners, reduce for losers
• OTB allocation between new buys and replenishment

============================================================ PHASE 6: MARKDOWN OPTIMIZATION

Step 6.1 -- Markdown Strategy

Evaluate the markdown approach:

• Markdown cadence and calendar (seasonal, promotional, end-of-season clearance)
• Markdown depth: initial markdown percentage, subsequent markdown cadence
• Markdown triggers: time-based, sell-through-based, inventory-age-based, or combination
• Optimization algorithm: maximize revenue, maximize margin, or minimize residual inventory?
• Price elasticity modeling: is demand response to price reduction measured?

Step 6.2 -- Markdown Performance

Check markdown effectiveness metrics:

• Markdown rate: % of units sold at markdown, % of revenue from markdown
• GMROI (Gross Margin Return on Investment) by category
• Maintained margin: initial markup vs. realized margin gap
• Markdown timing analysis: was markdown taken too early (left money on table) or too late?
• Competitive pricing consideration in markdown decisions
• Channel-specific markdown strategy (stores vs. outlets vs. e-commerce)

Step 6.3 -- End-of-Life Management

Assess exit strategy:

• Clearance options: deep discount, jobber/off-price, donation, destruction
• Residual inventory minimization targets and tracking
• Carry-forward assessment: hold inventory for next season decision framework
• Outlet/off-price channel management
• Inventory write-off policies and thresholds
• Seasonal inventory calendar alignment

============================================================ PHASE 7: WRITE REPORT

Write analysis to docs/apparel-demand-analysis.md (create docs/ if needed).

Structure the report as:

1. Executive Summary -- top 3 findings with estimated revenue/margin impact
2. Trend Forecasting Assessment -- data sources, methodology, accuracy
3. Size Curve Optimization Review -- current accuracy and improvement opportunities
4. Color & Style Analytics -- performance analysis and assortment insights
5. Sell-Through Performance -- current metrics vs. benchmarks
6. Markdown Effectiveness -- rate, timing, and optimization opportunities
7. Inventory Health -- WOS, aging, OTB process assessment
8. Prioritized Recommendations -- with estimated revenue and margin impact

============================================================ SELF-HEALING VALIDATION (max 2 iterations)

After producing output, validate data quality and completeness:

1. Verify all output sections have substantive content (not just headers).
2. Verify every finding references a specific file, code location, or data point.
3. Verify recommendations are actionable and evidence-based.
4. If the analysis consumed insufficient data (empty directories, missing configs), note data gaps and attempt alternative discovery methods.

IF VALIDATION FAILS:

• Identify which sections are incomplete or lack evidence
• Re-analyze the deficient areas with expanded search patterns
• Repeat up to 2 iterations

IF STILL INCOMPLETE after 2 iterations:

• Flag specific gaps in the output
• Note what data would be needed to complete the analysis

============================================================ OUTPUT

Apparel Demand Analysis Complete

• Report: docs/apparel-demand-analysis.md
• Product categories analyzed: [count]
• Seasons evaluated: [count]
• Average sell-through rate: [percentage]
• Markdown rate: [percentage]

Summary Table

Area	Status	Priority
Trend Forecasting	[status]	[priority]
Size Curve Optimization	[status]	[priority]
Color/Style Analytics	[status]	[priority]
Sell-Through Tracking	[status]	[priority]
Markdown Optimization	[status]	[priority]
Inventory Management	[status]	[priority]

NEXT STEPS:

• "Run /material-forecasting to align raw material planning with demand predictions."
• "Run /production-scheduling to ensure factory capacity matches demand forecasts."
• "Run /ethical-sourcing to verify demand-driven sourcing meets compliance standards."

DO NOT:

• Do NOT modify any demand forecasts, pricing, or inventory records.
• Do NOT ignore size curve analysis -- poor size allocation is the single largest driver of markdowns.
• Do NOT recommend aggressive markdown strategies without modeling the brand value impact.
• Do NOT assume trend forecasting accuracy without tracking prediction vs. actual performance.
• Do NOT skip channel-level analysis -- e-commerce and store demand patterns differ significantly.

============================================================ SELF-EVOLUTION TELEMETRY

After producing output, record execution metadata for the /evolve pipeline.

Check if a project memory directory exists:

• Look for the project path in ~/.claude/projects/
• If found, append to skill-telemetry.md in that memory directory

Entry format:

### /apparel-demand — {{YYYY-MM-DD}}
- Outcome: {{SUCCESS | PARTIAL | FAILED}}
- Self-healed: {{yes — what was healed | no}}
- Iterations used: {{N}} / {{N max}}
- Bottleneck: {{phase that struggled or "none"}}
- Suggestion: {{one-line improvement idea for /evolve, or "none"}}

Only log if the memory directory exists. Skip silently if not found. Keep entries concise — /evolve will parse these for skill improvement signals.