Retail site selection analysis: a practical guide

A store's performance is mostly decided before it opens. Four measurable forces explain most of the variance between a strong site and a weak one:

• Demand density — resident and daytime population inside a realistic trade area, weighted by the income and age bands that buy your category.
• Competitive saturation — how many comparable operators already serve that demand, and how close they are.
• Accessibility — road network, transit, parking, and the friction of actually reaching the door.
• Cost to occupy — rent and buildout relative to the revenue the trade area can realistically support.

Most credible analysis draws from public sources. In North America that means the US Census American Community Survey and Statistics Canada census profiles for demographics, OpenStreetMap for competitor and amenity points of interest, and mapping providers for travel-time isochrones. Climate and seasonality data matter more than teams expect for footfall-driven formats.

The hard part is never access — it is joining these sets to a consistent geography, keeping them current, and normalising them so two cities can be compared on the same scale.

1. Define the trade area by drive time, not a round radius.
2. Profile demand inside it: population, income, age mix, household composition.
3. Count and locate direct and adjacent competitors, then compute density per capita.
4. Weight each factor by what your format is sensitive to, and score every candidate identically.
5. Re-run quarterly — trade areas move as competitors open and close.

The trade area is the geography your customers actually come from, and every downstream number inherits it. Model a primary boundary that captures roughly two thirds of visits, using travel time rather than distance so rivers, motorways and one-way systems are respected.

If you already operate stores, calibrate against reality: geocode customer postcodes, plot them, and find the isochrone that captures 65% of them. Apply that shape to comparable formats. Keep one travel mode, one duration and one time-of-day profile across every candidate so scores stay comparable, and state the definition on the report.

More detail: trade area analysis and drive time versus radius.

• Measuring demand in an isochrone while counting competitors in a radius — the saturation metric can flip sign entirely.
• Mixing census vintages inside one comparison so markets are not on the same basis.
• Attaching tract-level demographics to a trade area without area-weighting the partial tracts.
• Approving a pro forma that only clears hurdle rate with zero sales transfer from an existing unit. Model a 10%, 20% and 30% cannibalization case instead.
• Treating the analysis as a one-off rather than a quarterly refresh.

The traditional workflow is a GIS analyst downloading census tables, geocoding competitor lists, buffering, and hand-building a scoring spreadsheet — days of work per market, and a model only its author can maintain. Location intelligence software collapses that: the joins, normalisation, and weighting run continuously, so the analysis becomes a question rather than a project.

The practical test of any tool is whether it shows its work. A score you cannot decompose into demand, competition, and accessibility contributions is not a decision aid.

Georithm asks for a plain-English question — "where should I open a second coffee shop near Ottawa?" — then assembles the trade area, live census demographics, OpenStreetMap competitor counts, and climate context into a transparent Georithm Score with a per-factor breakdown, an interactive map, and an exportable report.

You can save scenarios, compare geographies side by side over time, and set alerts that fire when competitive density in a market crosses your threshold.

What is retail site selection analysis?

It is the process of scoring candidate locations against the demand, competition, accessibility and occupancy-cost factors that drive store revenue, so sites can be compared on one consistent scale.

How large should a trade area be?

Size it by travel time rather than distance. Most formats use a 5-, 10- or 15-minute drive-time isochrone, calibrated against where existing customers actually come from.

How much data do I need before choosing a site?

Demographics inside a drive-time trade area, a consistent competitor count for the same boundary, an accessibility assessment, and occupancy cost. Everything else is refinement.

How often should the analysis be refreshed?

Quarterly for markets you are actively pursuing. Competitor openings and closures reshape trade areas within a single quarter.

• White-space analysis: finding unreached markets
• QSR site selection checklist
• Franchise territory mapping
• Data sources behind a defensible model
• Expanding into Canada

All articles