The problem
At roughly 750,000 SKUs, with new collections every 4–6 weeks, every launch triggered a scramble. Supplier feeds arrived inconsistent. Attributes were missing. Category mappings varied from vendor to vendor. Merchandising and ops spent the launch window cleaning data instead of merchandising the drop.
The cost was not only overtime. It was delayed assortment live dates, incomplete product pages at launch, and a recurring sense that the catalog would never be “ready enough” before the next cycle started.
The approach
Perspiq was introduced as a validation and enrichment layer inside the existing ingestion pipeline—not as a replacement for PIM, search, or ecommerce stack. New SKUs were checked, structured, and routed before publish.
- Attribute verification at ingestion — required fields and mappings checked as products entered the pipeline
- Structured enrichment — missing or inconsistent fashion attributes filled to a consistent schema
- Confidence-based routing — high-confidence records progressed automatically; exceptions surfaced for review
- Pre-publish readiness check — assortment only went live when catalog quality cleared launch criteria
The outcome
Launch readiness stopped depending on heroic cleanup sprints. Collections could move from feed to live with a predictable quality bar, and the team reclaimed the weeks between drops for selling and merchandising instead of firefighting attributes.
Catalog quality stopped being something the team had to achieve before every launch. It became something the system delivered automatically.