What data goes into a retail site selection brief

Collecting demographics and competition numbers gets you a fact sheet. Turning that into a brief means using the numbers to recommend one address over the alternatives, and standing behind the recommendation when a committee starts pulling on it.

Traceability separates the two. When every number names its source and the date it was pulled, a reviewer can challenge any line and get an answer. When the figures float with no provenance, the document reads as a slide deck. That gap matters more as store footprints shrink. Much of the retail space leased today is small-format, which means each site decision rides on a smaller box with less room to recover from a weak read on the trade area.

A three-mile radius is easy to draw and usually wrong. It ignores medians, rivers, one-way grids, and the evening backup on the main arterial, so it counts people who look close on a map but cannot reach the door in a car. A drive-time or walk-time area built from the road network counts who can actually get there in ten or fifteen minutes. State the method in the brief, for example a 10-minute drive-time isochrone at a representative weekday departure, then report the population inside it.

Residential population tells you who sleeps in the trade area. Daytime population tells you who is there at noon. For coffee or quick-service concepts those can diverge by a wide margin, and a brief that reports only residents undercounts a downtown corner and overcounts a bedroom suburb.

• Residential population inside the trade area, the baseline everyone already reports.
• Daytime / workplace population so an office-dense or commuter-heavy corner gets credit for the demand that drives in each morning.
• Household income to gauge whether the area can afford the offer, separate from whether people live there.
• Category spend and demand for your specific concept, so the demographics turn into addressable dollars.

The brief maps who else competes for the trade area and who helps fill it. Direct competitors inside the drive-time polygon show how contested the demand already is. Co-tenancy, meaning the anchors and neighbors at the center, shows whether the location pulls the kind of traffic you want. A grocery anchor or a complementary brand next door can move the forecast as much as the raw headcount does.

This section ages faster than any other. Point-of-interest data churns constantly, so a competitor that closed last spring can still sit in the file as an open store, and a new entrant may not appear yet. Name the POI source and its snapshot date so a reviewer knows how stale the competitive picture might be.

Include passing traffic counts, with state DOT average annual daily traffic as the usual source, plus ingress and egress quality, signal access, parking, and visibility from the road. The mistake is scoring these as more-is-better. Most behave as fit bands. A corner can carry too much traffic to turn into safely, or sit on a road so fast that nobody slows down for it. Report the raw value and the band it lands in, and flag any count that is several years old, since many of them are.

A site that looks strong on its own can simply pull sales from your store two miles away. A brief that scores one location at a time misses this, and the transfer only surfaces later in same-store comps. So the brief has to include your own network: which existing units share a trade area with the candidate, how much demand overlaps, and how much of the forecast is genuinely net-new.

Report demand-weighted overlap rather than the bare share of polygons that intersect, and name the existing store that takes the biggest hit. A small geographic overlap can still hide a large transfer when it sits over the densest demand, so the geometry by itself understates the damage.

Every figure should carry where it came from and when it was current. Census American Community Survey demand often comes from five-year estimates, so a population count can represent a window that closed years before the meeting. Competitor counts depend on a POI snapshot date. Traffic counts can be half a decade old. None of that disqualifies a number. It just has to be on the page, because a committee cannot weigh a figure it cannot date.

The rule is short to state and hard to keep by hand. No figure enters the brief without a source and a vintage, and the final score traces back to those inputs. A brief that holds to that bar can be challenged line by line and still stand.

• Trade area definition (drive-time or walk-time method stated) and population inside it
• Residential population and daytime / workplace population
• Household income and category spend / demand for your concept
• Direct competitors within the trade area, with POI source and snapshot date
• Co-tenancy and anchors at the center
• Traffic counts, accessibility, ingress / egress, and visibility (with fit bands, not raw maximums)
• Cannibalization and demand-weighted overlap against your existing units, naming the worst-affected store
• Net-new vs transferred demand in the forecast
• A component score with visible weights, traceable to the inputs above
• A source and a vintage date on every figure, plus a confidence note where data is thin

Geod produces this brief end to end: drive-time and walk-time trade areas, demographics, competition, cannibalization against your network, and a weighted component score, exported as a PDF where every figure carries its source and snapshot date. The data is most of the work. The part that holds up in the room is that the source-and-vintage rule is enforced automatically, so a brief never reaches the committee with an undated number in it.