Eureka Moments

Finding the Signal in Store Performance

How do you tell a struggling store from an untapped one? For a national specialty retailer, the answer was in the math.

The instinct when sorting underperforming stores is to rank by revenue and point at the bottom. But that's the wrong question. A modest store in a rural county with limited households nearby may be doing exceptional work. A high-revenue store in a dense metro may be leaving most of its market untouched. The right question: given the opportunity available to each store, how well is it performing?

The Methodology

We built a store penetration framework for a national specialty retailer covering 1,467 stores. Each store was evaluated against the household opportunity within a 15-mile radius — calibrated to reflect realistic catchment areas — using American Community Survey data as the household baseline. The output was a sales-per-target-household metric for every store in the network.

The insight isn't "which stores sell the least" — it's "which stores are reaching the smallest share of the households right in front of them."

We segmented stores on two dimensions simultaneously: absolute sales volume and sales-per-household penetration. That combination separates four meaningfully different situations — high volume / high penetration, high volume / low penetration, low volume / high penetration, and stores that are both low-volume and low-penetration. That last group is where the real conversation begins.

1,467
Retail stores
analyzed
15mi
Catchment radius
per store
95%
of stores have a
neighbor within ~30mi
Outlier markets — top performers Lowest-penetration cluster a single regional cluster High volume / high penetration Mixed performance Low volume / low penetration
Underperforming stores clustered heavily in a single region — a regional pattern worth investigating, not just flagging.

Outlier markets ranked top nationally on absolute volume — a function of geography and limited competition. Bottom performers clustered heavily in a single contiguous region. The question worth asking wasn't just that they were low, but why. Regional housing patterns, income distributions, and home ownership rates all play a role — and the framework is designed to surface those questions, not paper over them.

The Framework's Value

What makes a penetration model useful isn't just identifying who's at the bottom. It's creating a consistent, defensible basis for conversation. When a regional manager asks why their stores are flagged, the answer isn't "your numbers are low" — it's "here's the opportunity we calculated, here's what comparable stores are achieving, and here's the gap we're trying to understand."

The 15-mile radius is a parameter, not a constraint — tighter for dense urban networks, wider for rural footprints. The same logic applies at the state level, or against any segmentation criterion the business finds meaningful. The goal was never to produce a list. It was to build a shared language for talking about performance — one grounded in opportunity, not just outcomes.


Interested in building something similar for your retail or sales network? Say hello.