Skip to content

Why Collaborative Filtering Failed for Silk & Grace Intimates (And What Worked)

/images/Silk&GraceIntimates-hero.jpg?x-oss-process=image/resize,m_fixed,m_lfit,w_300

The Limits of Manual Curation in Intimate Apparel

Context: For Silk & Grace Intimates, scaling to $2.4M monthly GMV brought a critical challenge: High return rate of 38% due to size sensitivity and fit issues, creating significant operational overhead and customer dissatisfaction.

At this stage, standard rules failed. The Fashion-conscious women aged 25-45 seeking premium, comfortable lingerie and loungewear demanded relevance. Collaborative filtering algorithms, reliant on broad purchase patterns, couldn’t decode the nuanced fit requirements of intimate apparel. When a customer bought a 34C bra, the system recommended popular 34C panties – ignoring critical factors like cut preference, fabric stretch, and regional sizing variations. This one-size-fits-all approach exacerbated returns and damaged trust.

From Rules to Deep Learning: A Phased Evolution

To solve this, Silk & Grace Intimates deployed WooRec. Note: Depending on their scale, they leveraged WooRec Private Deployment for the perfect balance of speed and control.

The transformation wasn’t instant. We architected a phased evolution of their recommendation engine:

Phase 1: Expanding the Candidate Pool with Vector Retrieval

Powered by WooRec Strategy Module

We needed to move beyond simple keyword matching. We implemented a Hybrid Recall strategy:

  • Foundation: Real-time Intent (U2I):
    • We started by mapping user browsing behavior to item embeddings, solving the cold-start problem for new collections.
  • Advanced: Vector Retrieval (Embedding):
    • The Logic: We mapped users and items into a high-dimensional vector space using FAISS, encoding features like fabric composition, support level, and style attributes.
    • The Result: This allowed us to capture “latent fit preferences”—finding items that matched a user’s unspoken requirements beyond size tags.

Phase 2: Precision Ranking with DeepFM

Powered by WooRec Model Serving

This is the core of the engine.

  • The Iteration: Initially, the system relied on Logistic Regression. However, this failed to capture complex feature interactions between fit, style, and purchase history.
  • The Upgrade: We upgraded the ranking model to DeepFM.
    • Why this model?: By combining wide linear models (for memorization of size/rule patterns) with deep neural networks (for generalization of style/fit preferences), WooRec could now decode non-linear relationships like “customers who buy underwire bras in lace prefer Brazilian-cut panties in microfiber.”
  • The Goal: A precise score for p(CTR) and p(CVR), directly targeting Bra-Panty Matching Ratio.

Phase 3: Business Logic & Diversity

Powered by WooRec Rule Engine

Raw scores aren’t enough. We applied final adjustments to align with business KPIs:

  • Diversity (MMR): We enabled Maximal Marginal Relevance to prevent category fatigue, ensuring users saw complementary items (e.g., bralettes with thongs and boyshorts).
  • Business Weighting: We boosted items based on Inventory Depth, prioritizing complete sets to maximize AOV and reduce dead stock.

The Seamless Frontend Experience

Here is how these intelligent recommendations appear on the Silk & Grace Intimates storefront:

dashboard *Figure 1: The result of WooRec's engine—hyper-relevant product recommendations displayed to the user.*

The Impact: 80% Growth in Set Purchases

The speed of deployment meant faster results. By toggling on these strategies, Silk & Grace Intimates achieved:

> *Interactive Chart: The rapid growth curve following WooRec configuration.*
  • Bra-Panty Matching Ratio: Increased by 80% (25% → 45%).
  • Return Rate: Reduced by 40% (38% → 22.8%).
  • Average Order Value: Increased by 40% ($80 → $112).
  • Customer Lifetime Value: Improved by 30% ($320 → $416).

Customer Voice

“Moving from manual rules to DeepFM was a turning point. The system now understands user intent in real-time. The 80% lift in Bra-Panty Matching Ratio speaks for itself.” — Elena Rodriguez, Head of Data Analytics at Silk & Grace Intimates

Ready to Configure Your Growth?

You don’t need a team of data scientists to build a world-class recommendation engine. With WooRec, it’s just a matter of configuration.

Launch Your Strategy with WooRec