When a retail chain opens a new location near an existing one, some revenue from the existing store will move to the new one. This is cannibalization, and it is essentially unavoidable in any expansion that fills in existing markets. The question is not whether it will happen. It is how much will happen, and whether the economics of the combined two-store situation are better than the single-store situation was before the new opening.
That math is often not done carefully enough before a lease is signed. Teams focus on the new location's projected volume without fully accounting for the revenue that will migrate out of their existing locations. The result can be a new store that hits its individual target while a nearby established location sees a performance decline that the expansion model did not anticipate. Total network revenue may be flat or down despite the addition of a new unit.
Why the Hidden Costs Compound
The most visible cost of cannibalization is the revenue transfer from an existing store to the new one. That transfer is real and should be incorporated into any honest projection. But there are less visible costs that compound on top of the revenue transfer and are often excluded from the pro forma because they are less predictable.
Fixed cost absorption is one. An existing store that loses meaningful volume to a new nearby location now has the same fixed occupancy costs against a lower revenue base. If the existing store was operating at a margin that assumed a certain volume, the cannibalization-driven revenue loss falls almost entirely to the bottom line until costs can be reduced, which often takes longer than a quarterly cycle to accomplish. A store that goes from 100 to 80 units of revenue and has 70 units of fixed cost is in a structurally different position than one at 100 with 70 in fixed costs.
Brand investment dilution is another. Marketing spend and brand awareness efforts in a market have historically benefited a single store. When a second store opens nearby, the same marketing investment now needs to generate returns across two locations. The incremental store effectively gets some of its marketing attribution from the brand presence the first store already established, which is a form of shared overhead that is almost never modeled explicitly.
Staff retention is a third. Experienced employees at the existing store, particularly those who built customer relationships in the trade area, sometimes transfer to the new location either by choice or at the company's request to support the opening. The existing store pays a training and familiarity cost that takes time to rebuild.
When Cannibalization Makes Sense Despite the Cost
None of this is an argument against opening second or third locations in a market you already serve. Cannibalization can be economically justified when the total demand in the market is large enough that two stores cover more of it than one, when the new location captures demand that was previously being served by a competitor, or when network density provides a convenience advantage that improves your overall market share.
A simplified hypothetical: a chain operating a single location in a mid-size metro, capturing an estimated 60 percent of the addressable demand within a 12-minute drive. The remaining 40 percent of market demand is either going to a competitor or going unmet because the drive to the single location is too far for some customer segments. A second location positioned to cover the underserved portion of the market might capture a significant share of that remaining demand while cannibalizing only a portion of the first store's visits. If the total network revenue after cannibalization is higher than before the opening, and the new location's incremental contribution exceeds its total cost, the expansion is economically sound despite the cannibalization.
The test is the net two-store contribution versus the one-store baseline, not the new store's standalone projection. That net calculation requires explicitly quantifying the cannibalization at the first store, not treating it as negligible because it is inconvenient to model.
How to Estimate the Transfer
Estimating cannibalization accurately requires understanding where your existing store's customers come from and what share of them would find the new location meaningfully more convenient. Foot traffic data that includes home origin zone estimates for visitors to your existing store gives you the raw material for this analysis. Customers whose home origin is closer to the new site than to the existing site are likely to shift patronage once the new location opens.
A gravity model approach formalizes this by estimating the probability of customer patronage for each candidate location as a function of distance and store attractiveness. The model's output is a distribution of expected customer allocation across the two-store network, which produces both a total network volume estimate and an estimate of cannibalization rate at the existing store. Gravity models outperform radius-based approaches for this purpose because they account for road network geometry, which affects actual convenience more than straight-line distance.
The output of this analysis is not a precise prediction. It is a range estimate calibrated to the specific geography and customer origin patterns of your existing location. Treating the midpoint of that range as the base case for the pro forma and the upper end as a stress test scenario gives you a more grounded basis for the lease decision than ignoring cannibalization or estimating it based on percentage-of-rule-of-thumb.
The Compounding Effect Over a Network
Individual cannibalization decisions aggregate into a network-level pattern over time. Chains that consistently underestimate cannibalization build portfolios where many locations are modestly underperforming relative to their original pro formas, for reasons that are difficult to diagnose after the fact because the cannibalization happened gradually as the network expanded around each store.
The better position is to treat cannibalization modeling as a standard component of every new-location evaluation, not an optional analysis for cases where the proximity is obviously concerning. Building that discipline into the evaluation process when the network is at 20 to 40 locations, before expansion velocity makes it unwieldy, produces a portfolio with better network economics than one where cannibalization was assessed ad hoc and inconsistently.