Originally published on MyRevify.com — republished here for the Pricing Lever audience.
Product affinity analytics uncovered $10MM of hidden cross-sell opportunity at a mid-market manufacturer using machine-learning association rules.
Overview: Product Affinity Analytics
This case study shows how product affinity analytics uncovered $10MM in hidden cross-sell opportunity at a mid-market manufacturer by applying machine-learning association rules across the full transaction history. Product affinity was the concept leadership believed in directionally but could not operationalize; the analysis surfaced statistically significant co-purchase patterns, then ran customer-level product affinity gap analysis to flag every missing SKU in every basket. Product affinity analytics turns cross-sell from a gut call into a sized, prioritized commercial motion. See the full case study below, or read our related case study on Turning Purchase Patterns Into Pipeline.
The product affinity signal is refined by strict Support, Confidence, and Lift thresholds so only statistically robust rules survive. Every rule is then translated into a dollar-sized cross-sell opportunity at the customer level, making product affinity actionable for commercial leaders rather than abstract for data teams. Because the product affinity refresh is continuous, the cross-sell backlog behaves like a living pipeline rather than a one-time report, and the sales force always has a current list of the highest-value gaps to work.
Client Situation
The manufacturer had thousands of customers and a broad, technical product catalog — but no systematic way to identify which customers were buying part of a logical bundle while missing obvious adjacent SKUs.
Product affinity was a concept the leadership team believed in directionally but couldn’t operationalize. Without quantified gap analysis, sales conversations defaulted to ‘what the customer asked for’ rather than ‘what the customer probably needs’.
The Revify Approach
Discover — Machine Learning Association Rule Mining
- Analyzed 12 months of transactions using Machine Learning algorithms at both Product Line and SKU grain to surface statistically significant co-purchase patterns.
- Filtered to rules that passed strict Support, Confidence, and Lift thresholds to eliminate spurious associations.
Operationalize — Customer-Level Gap Analysis
- For every customer, compared their purchased basket against high-confidence rules and flagged each missing SKU as a specific cross-sell opportunity.
- Attached Revenue Upside and Gross Margin Upside to every flagged gap, giving the sales force a ranked, dollar-weighted list instead of a generic recommendation.
Concentrate — The 80/20 Within the 80/20
- Identified three product lines — as the cross-sell concentration points.
- Surfaced that $1MM of the upside was concentrated across just six customers (ranging from $76K to $225K of opportunity per account), enabling a targeted key-account motion instead of a broad-brush campaign.
Key Findings & Results
Identified a total $10MM of incremental sales upside available from cross-selling to the existing customer base — a figure the manufacturer was materially under-capturing and did not know was quantifiable.
Each sales rep and territory manager received their customized list of which customers was likely to buy which product, providing an easily executable opportunity roadmap.
The concentration pattern (three product lines; six top customers for the first $1MM) turned what could have been a multi-quarter sales program into a high-focus, fast-return motion.
| IMPACT DIMENSION | QUANTIFIED BENEFIT |
| Total cross-sell sales upside | $10MM |
| Incremental units | 150K |
| Incremental gross margin | $5.1MM |
| Concentration — top 6 customers | $1MM upside ($76K–$225K each) |
| Customer acquisition cost required | Zero — existing customers |
Why This Matters
| A ‘breadth and depth’ distribution problem looks intimidating until you see the concentration pattern. Three product lines and six customers accounted for the first million dollars — which made the next move obvious. |
Conclusion
By converting raw transaction history into statistically grounded cross-sell recommendations — each sized in revenue and margin dollars — the manufacturer unlocked a $10MM commercial motion with no new-customer acquisition required.
Because the analysis refreshes on a continuous cadence, the cross-sell backlog is a living pipeline rather than a one-time report.
Related Case Studies
- Turning Purchase Patterns Into Pipeline: $1MM+ Cross-Sell Opportunity Uncovered at an Existing Customer Base
- Protecting $1.6MM of Margin: Price-Cost Discipline and Category Strategy at a Mid-Market Manufacturer
Further reading
For broader industry perspective on revenue growth management and pricing analytics, see McKinsey’s Growth, Marketing & Sales insights.
