Case study
W
Wardrobe Market
Multi-brand fashion marketplace·Shopify Plus

Anticipate stockouts with ML forecasts

From 18% stockout rate to 4% by steering replenishment with Fullmetrix sales forecasts and automatic alerts.

-78%
Stockout rate
+31%
Stock rotation
+12%
Recovered revenue
7 months
Project duration

Context

Wardrobe Market is a French multi-brand fashion marketplace aggregating 85 creators and around 5,200 active SKUs. The Shopify Plus store generates 680,000 euros in monthly revenue and runs on a consignment model: products are stored in the central warehouse and creators are paid on sale.

Fashion is particularly exposed to stock issues: strong seasonality, ephemeral collections, unpredictable trend effects and high cost of end-of-season overstock. The stock team (2 people) manages replenishment via a custom ERP and daily Shopify exports.

The challenge

The stockout rate reaches 18%, meaning one in five product searches ends on an out-of-stock page. The revenue impact is estimated at 75,000 euros per month (lost revenue). Conversely, 22% of the catalog is overstocked and will need end-of-season discounting with negative margin.

Sales forecasts are done manually by the team based on the last 4 weeks of history, without accounting for seasonality, Google Trends, marketing campaign effects or new launches cannibalizing existing products.

The team is overwhelmed: managing 5,200 SKUs manually is impossible and replenishment decisions focus on the 200 obvious best-sellers, leaving the long tail neglected.

The Fullmetrix solution

Wardrobe Market connects Shopify Plus to Fullmetrix and activates the ML Forecasts module. Models are trained on 3 years of history and account for seasonality, trend effects and SKU velocity.

For each SKU, Fullmetrix calculates a 30, 60 and 90-day sales forecast with a confidence interval. Automatic alerts flag products at risk of stockout in the next 14 days, sorted by potential revenue impact. The team can thus prioritize the 50 most critical replenishments each week instead of reacting to stockouts.

A second view identifies low-rotation products at risk of becoming overstock. The team can then activate targeted promotions or bundles to sell through before end of season. The module also integrates margin per product to prioritize marketing efforts on the most profitable SKUs.

The results

Stockout rate divided by 4.5

Stockout rate goes from 18% to 4% in 7 months thanks to predictive alerts and automatic prioritization of critical replenishments.

Revenue recovered +12%

Monthly revenue grows 12% just from stockout reduction, about 81,000 euros of monthly revenue that was previously lost.

Overstock from 22% to 9%

Detection of low-rotation products enables early targeted promotions, dividing end-of-season overstock by 2.4.

Stock team time saved

The stock team saves about 15 hours per week previously spent on exports and manual cross-referencing, freeing time for creator relations.

« We were reacting to stockouts instead of anticipating them. Fullmetrix ML forecasts transformed our stock management: today we know which products to replenish 3 weeks in advance. And we halved our end-of-season overstock. »

L
Laura V.
Head of Supply Chain, Wardrobe Market

This case study is an illustrative example based on results observed with our users. Numbers are realistic but anonymized.

Anticipate your stockouts with ML forecasts

Automatic alerts, revenue-impact prioritization and overstock detection. Forecasts trained on your real history.

14-day free trial. No commitment.