With the Frequently Bought Together recommendation, you automatically display relevant product suggestions based on the actual purchasing behavior of your customers. By combining advanced market basket analysis with smart corrections, Tweakwise suggests products that add real value to the visitor's current purchase.
In this article, we explain what you need, how the algorithm works, and how to configure a Frequently Bought Together recommendation.
Requirements
To enable Frequently Bought Together, implementing the Purchase Event on your webshop's thank-you page is a requirement.
đ See our documentation for more information on configuring tracking events.
How does the algorithm work?
The Frequently Bought Together algorithm analyzes which items are regularly checked out in the same transaction. To ensure you always display the most relevant suggestions, the formula operates in three steps:
Co-occurrence discovery: It identifies products that frequently appear together in a single order.
Popularity correction: It filters out overly popular, generic bestsellers that end up in the basket purely by chance.
Sample size confidence: It corrects for rare items with insufficient transaction data.
To understand how Tweakwise applies this logic in practice, let's look at the two key corrections happening behind the scenes:
1. Correction for overall popularity
Suppose 10 customers buy apples: 5 of them also buy bananas, and 3 also buy pears.
While bananas might seem like the best match at first glance, it is very likely pure coincidence if 50% of all your webshop visitors buy bananas anyway. If pears are normally bought by only 5% of all visitors, that 30% rate among apple buyers is a much more unique and stronger signal of a genuine product relationship. The algorithm therefore filters out generic bestsellers (bananas) and prioritizes true connections (pears).
2. Correction for low transaction volumes
If an item has only been sold twice in total and once alongside an apple, it appears to be a 50% match on paper. However, a single shared purchase is statistically too weak. Tweakwise corrects for this and places higher confidence in frequent, reliable buying patterns than in rare, one-off coincidences.
How to set up 'Frequently Bought Together'
In the Tweakwise App, navigate to Recommendations.
Click + New.
Select the Frequently Bought Together type.
Enter the Name and specify the maximum number of recommendations to display under Number of results (e.g., 4).
Source item (Optional): Set conditions that the current product on the page must meet (such as a specific category or attribute) to trigger this recommendation rule.
Selection (Optional): Fine-tune which products can appear in the recommended results:
Result category: Limit recommended items to a specific category.
Variation in results: Ensure variety in the output by preventing items from the same category or with identical attribute values from appearing more than once.
Attribute conditions: Add specific filter conditions that recommended products must fulfill.
Priority (Optional): Determine the level of importance for this recommendation rule within a group.
Grouping codes (Optional): Assign one or more grouping codes (or create a new one) to link this rule to other recommendations using the same group key. This is essential for configuring a fallback strategy.
Click Save & return.
đĄ Best Practice: Always configure a Fallback!
Because Frequently Bought Together relies on historical purchase data, newly added products or niche items may not have sufficient transaction data available yet.
Always configure a Fallback rule using the Grouping functionality:
Create a second recommendation (e.g., an Up & Cross Selling rule or a Featured / Popular list).
Assign the same grouping code to both recommendations under the Grouping header.
Assign the highest priority to Frequently Bought Together.
Assign a lower priority to the generic fallback recommendation.
The result: If Tweakwise lacks sufficient purchase data to display Frequently Bought Together, the webshop automatically and seamlessly falls back on the generic recommendation. This ensures visitors never encounter an empty recommendation block on your webshop!