Frequently Bought Together Analytics: Boost Cross-Selling

TL;DR: Learn how to mine Amazon's "Frequently Bought Together" data, identify high-value product pairings, and use bundles and recommendation widgets to increase average order value and conversion rates.

Key Takeaways

Table of Contents

Note on marketplaces: This guide is specifically optimized for the US market.

What Is Cross-Selling Analytics?

Cross-selling analytics is the practice of mining customer purchase data to identify products that are frequently bought together. On Amazon, the most visible output is the "Frequently Bought Together" (FBT) widget on product detail pages. That widget isn't random: it's generated from the real co-purchase patterns of millions of shoppers. For sellers, the same data can be used to design bundles, improve product recommendations, and increase average order value.

At its core, cross-selling analytics relies on market basket analysis. Retailers look at each order as a "basket" of items and calculate how often product A appears with product B. The resulting metrics—support, confidence, and lift—tell you which combinations are statistically meaningful. Instead of asking "What products should I recommend?", you ask "What are customers already buying together?" and then make those pairings easier to discover.

Why "Frequently Bought Together" Matters for Amazon Sellers

Amazon's "Frequently Bought Together" module is one of the most powerful cross-sell tools on the marketplace. When shoppers see a recommended add-on right next to a product they are considering, they are more likely to add both to their cart. For sellers, this means higher average order value (AOV), better inventory rotation, and more opportunities to build customer loyalty. Yet most sellers treat FBT as a black box. They see which products are paired with their listings, but they never dig into the analytics behind those pairings.

Understanding FBT analytics allows you to see which of your products are natural complements in the eyes of real customers. You can spot gaps in your catalog, uncover demand for a product you don't sell, and even predict what shoppers will want next. If you are new to using data to grow your Amazon business, start with our Amazon seller analytics guide to build a broader measurement framework. This article focuses specifically on the cross-selling layer.

How to Use Frequently Bought Together Analytics to Increase Sales

To turn FBT data into revenue, you need a repeatable process. Here is a five-step framework you can apply whether you sell on Amazon or on your own eCommerce site.

Step 1: Collect and Segment Purchase Data

Start with your order data. If you sell on Amazon, export your orders from Seller Central or use a reporting tool. If you run an independent store, export from your eCommerce platform. Group transactions by order ID and create a clean table of product-SKU combinations. The goal is to answer one question: which products appear together in the same order?

Step 2: Run Market Basket Analysis

Use a spreadsheet, SQL, or a dedicated analytics tool. Create a co-occurrence matrix: count how many times each product pairs with every other product. Many analytics platforms have market basket analysis built in, but even a pivot table can work for catalogs with a few hundred SKUs.

Step 3: Score Pairings with Support, Confidence, and Lift

Not every pair is worth pursuing. Calculate the three core metrics for each product combination. Filter for pairings that appear in a meaningful number of orders (support) and have a lift above 1. A high lift tells you that customers are far more likely to buy B when they buy A than they would be otherwise.

Step 4: Apply Insights Across Channels

On Amazon, use your high-lift pairs to create virtual bundles or physical bundles, update A+ Content with a comparison table, and structure your PPC campaigns around companion products. On your own site, display a "Frequently Bought Together" widget on product pages and in the cart, or send post-purchase follow-up emails.

Step 5: Measure and Optimize

Track the attachment rate (how often the recommended item is added to the cart), average order value, conversion rate, and return rate. Run A/B tests to compare different placement, price, and bundle combinations. Use the winners to build a permanent cross-sell engine.

Key Metrics in Cross-Selling Analytics

Not all product pairings are created equal. To decide which "Frequently Bought Together" opportunities to act on, you need to understand three classic market basket metrics and a few business-level metrics.

Support

Support = (orders containing A and B) / (total orders). It tells you how common a combination is. A high-support pair may seem attractive, but it can simply reflect two best-selling products that happen to ship together, not a true cross-sell relationship.

Confidence

Confidence = (orders containing A and B) / (orders containing A). It is the conditional probability that B appears when A appears. This metric is practical for recommendation logic: if someone has A in their cart, confidence tells you the chance they will also buy B.

Lift

Lift = confidence / (support of B). It measures how much more likely B is to be purchased with A than on its own. A lift greater than 1 means A and B are truly associated; lift close to 1 suggests no real cross-sell opportunity; less than 1 indicates a negative association.

Business Metrics

Beyond the math, track the money: average order value, attachment rate, conversion lift, and return rate. A pair can have high lift but if the bundled product has a high return rate, the pairing may not be profitable.

How to Implement Product Recommendation Analytics in eCommerce

Product recommendation analytics can be implemented at several levels: Amazon product pages, email marketing, and your own website's recommendation engine. Start with the data you already have.

On Amazon: You cannot directly create the "Frequently Bought Together" widget, but Amazon uses customer purchase history to generate it. You can influence it by creating virtual bundles or by using A+ Content to embed a comparison/related products table. Sellers with Brand Registry can also use Brand Tailored Promotions to send cross-sell offers to customers who purchased a related product. To understand the full data landscape, read our comprehensive Amazon seller analytics guide before you build your strategy.

On your own eCommerce store: Use an analytics tool that supports market basket analysis, or write a simple SQL query against your orders table. Then display recommendations on the product page, cart page, and post-purchase confirmation. Many platforms offer "Frequently Bought Together" apps that do the heavy lifting. The key is to train the recommender on real orders, not just manually selected cross-sells.

Measure what matters: Set up A/B tests. Compare a control page with no recommendations versus a test page with a "Frequently Bought Together" block. Monitor add-to-cart rate, AOV, and repeat purchase rate. If the recommendation block lifts AOV by even a small percentage, it can have a major impact across your catalog.

Bundles, Related Products, and Cross-Selling Strategies

Once you have the analytics, the strategy is where the growth happens. Three of the most effective ways to use FBT data are virtual bundles, related product grids, and targeted cross-sell campaigns.

Virtual Bundles

Amazon Brand Registered sellers can create virtual bundles that combine two or more products into one detail page. FBT analytics helps you choose which products to bundle. Instead of guessing, look for pairings with high lift and a clear use case. A virtual bundle lets you control the price, promote it as a set, and drive more units per order. To understand the mechanics, see our detailed guide on Cross-Selling with Virtual Bundles on Your Listing.

Related Products

Use FBT data to decide which products appear in the "Compare with similar items" or "Customers also viewed" sections. While you can't edit Amazon's algorithm directly, you can use your analytics to guide your catalog structure, product groupings, and even PPC targeting.

Targeted Promotions and Email

If you collect customer emails (via Amazon's Subscribe & Save or your own store), send post-purchase follow-ups that recommend a complement. For example, if a shopper bought a coffee maker, recommend filters or beans. Your market basket analysis will give you the list of high-lift partners.

Best Practices and Common Pitfalls

Key Benefits

Cross-selling analytics delivers higher average order value, better inventory turnover, a more personalized shopping experience, and data you can use for product development. It also helps you build a stronger product ecosystem around your bestsellers.

Limitations and Risks

Correlation is not causation. Two products may be purchased together because of seasonality, promotions, or sheer popularity, not because they genuinely complement each other. On Amazon, the FBT widget is controlled by the algorithm, and it may recommend competitor products alongside yours. Over-bundling can also cannibalize sales or increase return rates if customers feel forced into a set.

Common Pitfalls to Avoid

Don't optimize for support alone. Don't ignore lift. Don't create bundles without validating profit margins. Don't assume a recommendation that worked on your own website will work on Amazon. And always watch for inventory availability—promoting a product that is out of stock is a fast way to hurt customer trust.

Conclusion

Cross-selling analytics turns observable customer behavior into a repeatable growth playbook. By decoding "Frequently Bought Together" data, you can identify the pairings that matter, create bundles that feel natural, and place recommendations where they drive the most revenue. The key is not to rely on guesswork. Use the metrics, test your ideas, and let your customers show you what they want to buy together.

FAQ

Next Steps

References

• SellerSprite — Amazon Seller Analytics Guide View
• SellerSprite — Cross-Selling with Virtual Bundles on Your Listing View