Grocery site selection: trade areas, delivery, leakage, and cannibalization

Grocery trips are frequent and habitual. A household visits a primary store once or twice a week, so the catchment is shaped by routine and convenience more than by a single destination trip. That catchment is usually larger than a quick-service restaurant or a coffee shop, and it comes in tiers. A primary catchment, often within a short drive, accounts for the bulk of sales. A secondary ring beyond it contributes a thinner stream of trips from shoppers who pass the store on the way home or come in for one item. Treat both as a single flat circle and you hide where the volume actually sits.

Online pickup and home delivery have stretched the old picture. Pickup keeps the store at the center but widens who is willing to drive to it, because a planned order removes the time spent in the aisles. Delivery decouples demand from the storefront, so a store can serve households outside its walk-in catchment, bounded by delivery range and fulfillment cost rather than by drive time. For planning, that means a single trade-area ring understates total demand. You have to decide how much pickup and delivery volume a candidate can realistically capture before you trust the forecast.

An area can look underserved on paper while its residents already spend their grocery dollars somewhere else, driving past closer options to reach a store they prefer. That outflow is leakage, and it is the clearest signal of demand a new or better-matched store could win back. The reverse case matters too. A market can look saturated by store count and still leak sales to a format it lacks. Mapping where current residents shop, and how much of that spend leaves the area, tells you whether a new store would capture genuinely new trips or redistribute trips that already happen nearby.

Two grocery stores a mile apart can barely compete if one is a full-size supermarket and the other is a hard discounter or an ethnic specialty market. Format sets the basket, the price point, and the shopper. A full-size store carries breadth and fresh departments and earns the weekly stock-up trip. A discount banner wins on price and a tighter assortment. A specialty or ethnic grocer serves a basket the majors do not stock well. Counting competitors without reading format overstates pressure in some trade areas and misses it in others. A same-format rival three miles out can weigh more than a different-format store next door.

A grocery store is often the anchor itself, the tenant that pulls steady weekly traffic the smaller neighbors depend on. When you are the anchor, the question is what the surrounding center adds: complementary co-tenants that lengthen a trip, or a vacancy and tired signage that drag on it. When the grocery is a tenant in a larger center, the draw of the other anchors and the visibility of the site shape how many of those trips convert. Either way, the co-tenancy mix and the physical site, access, parking, and ingress, move the forecast as much as the demographics around it.

Grocery chains tend to build out in clusters, and that density is where cannibalization bites. A new store in a strong corridor can post good numbers while quietly pulling a share of its volume from the two existing units that already serve the same households. Scored on its own, the site looks like a clear win. Netted against the network, part of its sales is just demand you moved across the street. The useful measure is net new volume: how much of a candidate's forecast is incremental to the chain, and how much comes out of stores you already operate.

Geod scores grocery candidates on the geography that matters for the format. It builds drive-time trade areas from the real road network instead of a flat radius, and you can read a primary catchment and a secondary ring as nested drive-time bands. It pulls demographics and demand inside each trade area, maps competing stores so you can read them by location and format, and estimates cannibalization against your own units to show net-new volume rather than gross. Every score is explainable, tracing back to the weighted factors you set, and you can export a brief for a real estate committee with the trade areas and sources attached.