Overlapping Shopper (Cross Shop) Views
What Is It?
The Cross-Shopper Model, or “By Overlap” view, quantifies how much overlap exists between the customer bases of two brands.
“Do people who shop at Brand A also shop at Brand B?”
This model assigns a score from -1 to +1 representing the statistical affinity between two brands. It enables analysts to uncover shared audiences, competitor adjacency, and potential co-marketing opportunities.
Why It Matters
Understanding customer overlap unlocks insights for:
- Competitive Benchmarking — identify direct and indirect rivals based on audience overlap
- Co-Marketing Strategy — find potential partners that share loyal customers
- Churn Analysis — detect where customers are drifting
- Audience Expansion — identify adjacent brands your customers already trust
How It Works
Carbon Arc’s Cross-Shopper Model uses Normalized Pointwise Mutual Information (nPMI) to evaluate brand-to-brand customer overlap each month.
Affinity Score Breakdown
| Affinity Score | Meaning |
|---|---|
| +1 | Perfect overlap — identical customer sets |
| > 0 | Statistically significant positive overlap (shared audience) |
| 0 | No measurable affinity |
| < 0 | Shared customers less than expected by chance (anti-affinity) |
| -1 | No overlap at all |
Even if two brands share customers, a negative score may result if that overlap isn’t statistically significant.
Example Use Cases
| Use Case | Value |
|---|---|
| Competitive Analysis | Determine core vs. fringe competitors by overlap strength |
| Brand Partnerships | Co-market with brands sharing your top user segments |
| Customer Retention | Spot potential churn to overlapping competitors |
| Product Strategy | Understand which adjacent categories your shoppers engage with |
Model Logic
The model computes overlap monthly using:
- Unique customer counts for each brand
- Total customers in the ecosystem that month
- Count of customers shared between both brands
The result is normalized using PMI to account for brand size and total universe.
Output Format
| Brand A | Brand B | Month | Affinity Score |
|---|---|---|---|
| Nike | Adidas | 2024-03 | 0.45 |
| Sephora | Ulta | 2024-03 | 0.83 |
| Costco | Whole Foods | 2024-03 | -0.21 |
- Only includes brand pairs with ≥10 users per brand per month
- Optionally excludes brand pairs with zero overlap (
score = -1)
Key Metrics
- Shared Customers: Unique users transacting with both Brand A and Brand B in a given month
- nPMI Score: Normalized Pointwise Mutual Information score between -1 and +1
Availability
- Delivered monthly as brand-to-brand affinity tables
- Integrated into dashboards, brand trackers, and LTV measurement pipelines