Strategies define how the plugin calculates product relevancy. Each strategy is a set of weighted ranking signals that determines which products are recommended together.
Default strategy
The plugin ships with a Balanced Baseline strategy that uses all five ranking signals with balanced weights:
- Category (30%) — How closely products share categories
- Attributes (25%) — How many product attributes overlap
- Price (20%) — How similar the price points are
- Sales (15%) — How well products sell (best-sellers get a boost)
- Recency (10%) — How recently the product was published
Adjusting weights
Drag the weight sliders to change how much each signal influences the recommendations. Weights are normalized automatically — they always sum to 100%.
Examples of when to adjust weights:
- Fashion/apparel stores — Increase the Attributes weight so products with matching colors, sizes, or materials score higher
- Stores with frequent new arrivals — Increase the Recency weight to keep recommendations fresh
- Stores focused on best-sellers — Increase the Sales weight to surface popular products
Custom taxonomy signals
If the store uses custom product taxonomies — such as brands, collections, or product lines — these can be added as additional ranking signals. Products that share the same taxonomy terms score higher.
Enable custom taxonomy signals in Settings, then select which taxonomies to include. Each selected taxonomy appears as a new weight slider in the strategy editor.
Strategy filters
Filters narrow which products can appear as recommendations for a given strategy. Available filter rules:
- Category — Matches current product category, is a specific category, or is not a specific category
- Stock Status — Only recommend products that are in stock (or exclude out-of-stock products)
- On Sale — Only recommend products that are on sale, or exclude sale products
- Price — Only recommend products above or below a price threshold
- Custom Taxonomy — Filter by any registered product taxonomy
Multiple filter rules can be combined. All rules must match for a product to appear as a recommendation.
Creating additional strategies
Create multiple strategies to use different recommendation logic in different placements. For example, a “Best Sellers” strategy with high sales weight for the shop page, and a “Similar Products” strategy with high category and attribute weights for product pages.
Each placement selects which strategy to use, so the same product catalog can produce different recommendations depending on where they are displayed.
Regenerating scores
After changing strategy weights or filters, click Sync on the Dashboard to regenerate relevancy scores. The sync runs in the background and processes products in batches. A progress indicator shows the current status.