ECOMMERCE MARKETING CONSULTING
Case Study: Brewing Better Bundles with Data
How a Tea Company Uncovered Hidden Product Pairings Beyond Shopify & Google Analytics
During Black Friday, a direct-to-consumer tea company wanted to do more than just track sales and revenue. They wanted to understand customer buying behavior at a deeper level, particularly which teas were being purchased together, and whether that information could be used to improve bundling, product recommendations, and targeted marketing.
Instead of relying on platforms like Shopify or Google Analytics, we extracted and analyzed raw transaction data, using our advanced data layering techniques to generate a product pairing heatmap and rank their top co-purchased items.
While Shopify and Google Analytics provided basic metrics like:
- Best-selling products
- Conversion rates
- Average order values
Let’s dive into our methodologies, starting with the raw data.
Product Affinity Clustering (Basket Analysis)
Using raw order-level data from Shopify (exported into a clean database), we built a co-occurrence matrix (i.e. a product affinity matrix). Each cell in the matrix represented how many times two specific tea products appeared in the same order.
This method allowed the team to surface co-purchase patterns that are often invisible in Shopify or Google Analytics.
“Chamomile” appeared in multiple top pairs, suggesting it’s a universal complementary flavor. Ideal for cross-sells or bundles.
Strategic Insights
Create new bundles: Pairs like Oolong + White Peony and Jasmine + Peppermint represent natural bundles the company wasn’t promoting.
- Revise product recommendation engine: Instead of recommending top sellers, suggest items based on co-purchase history
- Boose Average Order Value (AOY) by suggesting known pairings at checkout.
Impact
The tea company redesigned its Q1 2026 product bundles based on these findings and saw:
- Increase in AOY from bundle-based promotions
- Higher conversion rate on personalized “You Might Also Like” suggestions
- More efficient inventory planning, based on expected demand for paired items
Conclusion
While Shopify and Google Analytics are powerful for top-level performance monitoring, they fall short when it comes to uncovering nuanced customer behavior like product pairings and natural bundling opportunities.
By exporting raw order data and applying basket analysis, this tea brand gained strategic, bottom-line-enhancing insights that were otherwise hidden from view
This is what Deep Strata Layered Data really looks like: moving from “what happened” to why it happened and how to act on it.
Projects
Check Other Case Studies
Ready To Stop Guessing And Start Moving With Confidence?
If you’re spending real money on digital ads but not sure what’s actually working—or where the next growth pocket is hiding, it’s time to put your market on the map.
Marketing Consultants