Developing a Recommendation System Module for 1C-Bitrix
"Customers who bought X also bought Y" increases average check. Recommendation system analyzes purchase history and suggests related products.
Algorithm
Collaborative filtering — find similar customers, recommend their purchases:
$similarCustomers = getSimilarCustomers($userId);
$theirOrders = getOrderedProducts($similarCustomers);
$recommendations = array_diff($theirOrders, $userOrders);
Recommendation Types
- Similar products — by category, attributes
- Often bought together — association rules
- View-to-buy — products viewed with high purchase rate
- Trending — popular this week
Implementation
Widget on product page and cart shows 3-5 recommendations. Admin configures:
- Algorithm type
- Display limit
- Minimum confidence threshold
- Exclusions (out of stock, low rating)
Development Timeline
| Stage | Duration |
|---|---|
| Basic (similar products) | 5–7 days |
| Algorithm tuning, A/B testing | 10–14 days |
| Full system with analytics | 16–20 days |