86.7% Cross-Sell Surge: How Evergreen Gardens Built a Private AI Engine

The Limits of Manual Curation and Data Silos
Context: For Evergreen Gardens Direct, scaling to $1.8M monthly GMV brought a critical challenge: High return rates due to plant death during shipping or from improper care by customers.
At this stage, standard rules failed. The Home gardening enthusiasts, landscape professionals, and eco-conscious consumers aged 28-55 with above-average disposable income demanded relevance. Generic third-party solutions were untenable due to data security concerns, while their self-developed system couldn’t factor in regional climate zones or seasonal planting schedules. The result? A 22% return rate and stagnant cross-selling at 15% – directly impacting perishable inventory and customer lifetime value.
From Rules to Deep Learning
To solve this, Evergreen Gardens Direct 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: We started by ensuring popular items were visible via Hot Retrieval, solving the Cold Start problem.
- Advanced: Vector Retrieval (Embedding)
- The Logic: We mapped users and items into a high-dimensional vector space using FAISS/Graph Embedding, capturing semantic relationships between plant varieties, care requirements, and regional climates.
- The Result: This allowed us to capture “latent interests”—finding items that are semantically related, not just textually similar. For instance, customers in arid zones received drought-tolerant succulents alongside complementary soil mixes.
Phase 2: The Model Evolution (LR to ESSM)
Powered by WooRec Model Serving
This is the core of the engine. To achieve the target Cross-selling Rate, we iterated through three stages:
- The Baseline (Logistic Regression): Initially, we used linear models. While fast, they failed to capture complex feature interactions like how humidity levels interact with plant care difficulty.
- The Upgrade (DeepFM): We introduced Deep Factorization Machines to learn high-order feature interactions, significantly improving accuracy on sparse data like seasonal planting schedules.
- The Final State (ESSM):
- Why this model?: To solve the CVR estimation bias inherent in traditional CTR models, we deployed the Entire Space Multi-Task Model (ESSM). This jointly optimizes click-through and conversion rates, ensuring recommendations don’t just attract clicks but lead to actual purchases of complementary gardening tools.
Phase 3: Traffic Control & Business Logic
Powered by WooRec Rule Engine
Raw scores are just probability predictions. To align with business goals, we applied a Traffic Control Layer:
- Diversity (Scatter/Shuffle): We implemented a sliding window rule—no more than 2 items from the same category in a row—to prevent visual fatigue in product listings.
- Business Injection (Hard Insertion): Specific slots (e.g., Position 4 and 10) were reserved for
Evergreen Gardens Direct’s strategic partners or high-margin house brands like premium organic fertilizers. - Dynamic Weighting: We boosted items based on Inventory Depth, prioritizing plants nearing optimal shipping maturity to minimize transit mortality and reduce returns.
The Seamless Frontend Experience
Here is how these intelligent recommendations appear on the Evergreen Gardens Direct storefront:
*Figure 1: The result of WooRec's engine—hyper-relevant product recommendations displayed to the user.*The Impact: 86.7% Cross-Sell Growth
The speed of deployment meant faster results. By toggling on these strategies, Evergreen Gardens Direct achieved:
- Cross-selling Rate: Increased by 86.7% (from 15% to 28%).
- Return Rate: Reduced by 45.5% (from 22% to 12%).
- Average Order Value: Grew 36% to $102.
- Customer Lifetime Value: Surged 56.9% to $510.
Customer Voice
“Moving from manual rules to ESSM was a turning point. The system now balances user intent with our perishable inventory logic perfectly. The 86.7% lift in Cross-selling Rate speaks for itself.” — Alex Rivera, Head of Data Science at
Evergreen Gardens Direct
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.