September 9, 2026 · Tupll
How to Spot Cannibalization Risk Before Opening Another Location
Cannibalization is the failure mode nobody sees coming, because the new store does not fail. It opens, it performs respectably, everyone celebrates, and meanwhile two existing locations quietly give up the revenue that made the new store's numbers possible. Portfolio-wide, you paid capital and rent to move sales you already had.
Here is how to catch it before the lease is signed, not in the year-two comp report.
Why cannibalization hides so well
Site evaluations usually judge candidates in isolation: this trade area, this demographic profile, this competition. On those terms, the strongest candidate is often strong precisely because it sits near your proven success, in the same kind of market, drawing on the same kind of customer.
That is the trap. Proximity to your best market usually means overlap with your best stores. The evaluation praises the site for the very thing that will hollow out your existing revenue. And because the damage lands on the old stores' comps rather than the new store's P&L, the postmortem rarely names the real cause.
Model the network, not the site
The fix is structural: score every candidate as an addition to your network, not as a standalone bet. That requires knowing where your current customers and revenue actually come from, which is exactly what a model trained on your own locations and sales history encodes.
When a candidate site's projected demand draws from zones your existing locations already serve, the model can flag the overlap and estimate what the addition does to the portfolio, not just what the new pin earns. Sometimes the honest answer is that a site with a lower standalone score is the better open, because its revenue is incremental rather than transferred.
We flag this directly in our scoring: a candidate can rank below your portfolio median with a cannibalization warning attached, even when its trade-area numbers look healthy on their own.
The ROI math is lopsided
Catching one cannibalizing site typically pays for the entire analytical engagement several times over. Run the numbers on your own portfolio: take a mid-performing store's annual revenue, assume a new nearby site transfers even 15 to 20 percent of it while adding its own rent, build-out, and payroll, and compare that ongoing drag to the one-time cost of modeling the decision properly.
The asymmetry is the argument. Site analysis is cheap insurance against a mistake that compounds every month the doors are open.
Three checks before your next open
First, map overlap explicitly: for each candidate, ask which existing stores draw customers from the same zones, and how much of the candidate's projected demand is already being served. Second, demand a portfolio-level forecast: what does the network earn with this site added, versus without it? Third, be suspicious of candidates that cluster near your winners; they borrow their strength from stores you already own.
Growth that moves revenue sideways is not growth. The whole point of modeling cannibalization before signing is to make sure the next open adds a market instead of splitting one.