Retail location analysis is the process of scoring potential store sites against demographics, foot traffic, competitor proximity, co-tenancy, and total occupancy cost to estimate how a location will perform before you sign the lease. Done well, it replaces a gut call with a forecast your finance team can actually check.
Why the Site Decision Is the One That Compounds
The stakes are structural. A lease locks you in for years, so unlike a marketing campaign or an inventory bet, you cannot quietly reverse a bad site. And despite two decades of e-commerce growth, 81% of US retail sales still happened in physical stores in 2025 (US Census Bureau retail trade data). The store is still where the revenue is, which means the site decision is still the one that compounds.
I'm Clyde Christian Anderson, CEO of GrowthFactor.ai and a former retail real estate professional who has personally evaluated hundreds of potential store locations. My experience spans front-line retail operations to investment banking, which is a long way of saying I have watched this decision get made well and badly.
How to Run a Retail Location Analysis: Seven Steps
A retail location analysis follows the same sequence whether you are opening your third store or your three-hundredth. What changes is how much of it a tool does for you.
- Profile your winners first. Pull two to three years of revenue by store and isolate the top quartile. What do they share? Trade-area income, daytime population, co-tenants, road position, square footage. This pattern is your scoring criteria, and it is why two brands on the same street can rationally reach opposite conclusions about the same vacancy.
- Define the trade area. Draw the geography the store will actually draw from — by drive time, not by a radius circle. Rivers, highways, and rail lines cut trade areas in ways a five-mile ring never sees. Every number downstream is built on this boundary, so over-drawing it inflates your forecast.
- Measure demand inside it. Layer demographics, daytime population, and psychographic segments against the customer profile from step one. The question is not "how many people live here" but "how many of my customers live here."
- Count and qualify the traffic. Get foot traffic and vehicle counts, then qualify them. A highway with 40,000 cars a day and no safe right-turn ingress is worth less than a signalized intersection with 12,000.
- Map the competition and the co-tenants. Direct competitors, indirect competitors, and anchors. Density cuts both ways: a competitor next door can signal a proven trade area or a saturated one, and the difference is whether demand is growing.
- Cost it fully. Rent is the number everyone quotes and the smallest part of the answer. Add CAM, taxes, utilities, buildout, and the escalation schedule. A cheaper rent with a $600K buildout is not cheaper.
- Score, forecast, and rank. Combine the layers into a single score and a revenue forecast, a range rather than one brave number, then rank the site against every other site in your pipeline. Ranking is the step teams skip, and it is the one that makes the answer defensible.
The Five Factors That Decide the Score
Every framework in this space is a rearrangement of five things. Getting the weights right for your brand matters more than adding a sixth.
Demographics and psychographics
Age, income, education, household composition, and daytime population set the ceiling on demand. Psychographics tell you whether those people are your people: whether they trade convenience for price, whether they shop weekly or monthly.
One demographic story worth updating: the "millennials are buying suburban homes, follow them" narrative that drove a lot of 2021-era expansion strategy no longer describes the market. First-time buyers fell to 21% of all home purchases in 2026, the lowest share since NAR began tracking in 1981, and millennials are now just 26% of buyers (National Association of Realtors, 2026 Home Buyer and Seller Generational Trends). Household formation is not moving where it was assumed to be moving. If your trade-area model still carries a five-year growth assumption written in 2021, it is worth re-checking against current projections.
Accessibility and foot traffic
Visibility, signage sightlines, parking, ingress and egress, and transit access. This is where foot traffic analysis earns its keep, not as a raw count but as a pattern. When do people come, how long do they stay, where else do they go?
Competition and market saturation
Map direct and indirect competitors, then read density honestly. Restaurant rows and auto malls exist because clustering works; saturated corridors exist because it stops working. The useful output is a whitespace read: where demand exists and competition doesn't. Forever 21's collapse is a case study in what happens when a chain expands into corridors it had already saturated — the lessons from its bankruptcy are mostly location lessons.
Co-tenancy and future development
Anchors pull traffic; complementary neighbors extend the trip. A grocery, pharmacy, and coffee shop in one center means people run three errands and pass your door on two of them. Then look forward: planned residential, permitted construction, and zoning changes can remake a corridor inside a lease term, in either direction. Municipal planning documents are public and almost nobody reads them.
Total occupancy cost
Rent, CAM, taxes, utilities, buildout, escalations. The best trade area in your pipeline is still a bad deal if the occupancy ratio never works.
A Worked Example: 9 New Stores a Year to 27
Frameworks are easy to nod along to, so here is what the sequence above looked like in practice.
Cavender's, a Western-wear retailer, was expanding with a small real estate team and the usual constraint: more candidate sites than analyst hours. Their existing stores were the asset nobody was using. Scoring criteria were configured against their own store performance, what their top locations shared, rather than against generic retail benchmarks, and the score did the job a score should do: qualify and disqualify. It cleared the weak sites out of the pipeline early so analyst hours went to the candidates that deserved them. For the sites that cleared the screen, a custom revenue forecasting model built on Cavender's own store data, through a GrowthFactor Labs engagement, produced the projection each deal was underwritten against.
The result: Cavender's opened 27 new stores in a single year, up from 9 the year before, with every new location performing at or above projection, roughly $2M saved, and about 50% less time spent per site. The throughput mattered less than the hit rate. Tripling your opening pace is easy if you are willing to miss; doing it while every store clears its forecast is the part that requires the analysis to be right. The full Cavender's case study has the detail.
The pattern repeats at different scales. TNT Fireworks opened 153 locations in 6 months, 100% on budget. Across customers surveyed in January 2026, GrowthFactor teams reported roughly 80% fewer underperforming locations.
What Changed in Retail Location Analysis for 2026
Two shifts are worth knowing about, because most guides on this topic were written before either happened.
Location data is getting regulated. Oregon's HB 2008 took effect January 1, 2026 and prohibits selling precise geolocation data within a 1,750-foot radius of a consumer's location, with additional restrictions on data from minors under 16. Several other states are weighing comparable rules. If your site model leans entirely on mobile-location panels, that is now a supply-chain risk in a way it was not in 2024. The practical defense is not to abandon foot traffic but to stop treating it as the whole analysis: it is one of five factors, and demographics, co-tenancy, road position, and your own store data are unaffected by any of this.
The market is smaller and more honest than the old projections claimed. Location analytics is now sized at $13.88B in 2026, projected to reach $24.19B by 2031 at an 11.7% CAGR (MarketsandMarkets, July 2026). That is a meaningful step down from the $38.5B-by-2028 figure that circulated for years, including in an earlier version of this article. Category sizing in this space is vendor-driven and noisy; the direction is real, the precision is not.
One more correction worth making, since this article previously carried the looser number. Closing a store does measurably suppress local online sales, but the credible figure is not 50%. ICSC's Halo Effect III study — $848.1B in card transactions across 69 retailers and 2,103 stores — found closures cut local online sales by 11.5% on average, with home goods the hardest hit at 32.2%. Still a real halo, still an argument against treating stores and e-commerce as rivals, just not the number the internet keeps repeating.
Technology, Models, and Where the Judgment Stays
The honest version of the technology story is narrow: modeling handles the volume, and your team keeps the judgment.
What a model does well is process more variables than a person can hold at once — mobility data, demographics, competitor sets, historical performance — and find the lookalikes. If your top store shares a signature with a vacancy three states away, that connection is findable in the data and effectively invisible in a spreadsheet. It also removes the grind: qualification, initial screening, preliminary scoring. Teams using GrowthFactor evaluate roughly 5x more sites as a result, and get a complete PDF report in about 10 seconds.
What a model does badly is explain itself, and that is where most of these tools lose the room. You have probably sat in the meeting: a score lands on the table, finance asks where it came from, and nobody can answer. The score is right or wrong but it is unusable either way, because it cannot be defended.
So the mechanism matters more than the model. In GrowthFactor's Site Scoring Glass Box, clicking a score shows the variables that produced it and lets your team change the weights and re-run. When your strategy changes, the model changes with it. Scoring criteria get configured against your own store performance at the start, which is why the same vacancy can rank differently for two brands and both rankings can be correct.
Does that mean the model makes the call? No. It means the person who makes the call can explain it. For more on the underlying data layers, see site selection data and trade area analysis; for the workflow end, see the retail site selection process and sales forecasting.
Frequently Asked Questions about Retail Location Analysis
What are the first steps in retail location analysis?
Start by profiling your existing stores, not the site you are looking at. Pull two to three years of revenue by location, then find what the top quartile has in common: trade-area income, daytime population, co-tenants, road position, square footage. That pattern becomes your scoring criteria. Teams that skip this step end up scoring every site against generic retail benchmarks rather than against their own brand, which is how a site that looks strong on paper turns into an underperformer.
What data do you need for a retail location analysis?
Five layers cover most decisions: demographics and daytime population for the trade area, foot traffic counts and dwell time, competitor and co-tenant locations, drive-time accessibility and road position, and full occupancy cost including rent, CAM, taxes, and buildout. Census data is free and reliable for demographics. Foot traffic comes from mobile-location panels. The expensive gap is usually your own store performance data, which is what turns generic benchmarks into a model calibrated to your brand.
How long does a retail location analysis take?
With purpose-built retail location analysis software, a preliminary site score and trade-area read takes minutes to hours; GrowthFactor returns a complete PDF report in about 10 seconds. Full due diligence including field visits, lease review, and financial modeling typically runs one to four weeks. Market-planning platforms like Sitewise and Kalibrate offer their own self-service tooling, while consultant-led studies can extend timelines to weeks, which is the part that hurts when a broker needs an answer before Friday.
How does GrowthFactor compare to Buxton for retail location analysis?
Buxton has been doing customer analytics since 1994 and has a strong psychographic methodology when you can feed it rich CRM data. GrowthFactor is self-service: every site score shows the underlying variables and lets your team adjust weights and re-run in real time, rather than waiting on a consultant to return a study. Plenty of teams run both, and GrowthFactor is built to sit alongside an incumbent rather than replace it. Cavender's went from opening 9 new stores in a year to 27: scoring qualified the pipeline, and a custom revenue forecasting model built on their own store data set the projections that every new location has met or beaten.
Does retail location analysis still work if mobile location data gets restricted?
Yes, but the sourcing matters more than it used to. Oregon's HB 2008 took effect January 1, 2026 and bars selling precise geolocation data within a 1,750-foot radius of a consumer's location, and several other states are weighing similar rules. Foot traffic is one input among five, not the whole analysis, so a well-built model degrades gracefully: demographics, co-tenancy, road position, and your own store performance data are unaffected. Ask any vendor where their foot traffic comes from and how they handle state-level opt-outs.
Where to Start
If you take one thing from this guide, make it step one: profile your own stores before you score anyone else's vacancy. Most teams have the answer sitting in their own revenue data and score against industry averages anyway.
From there the sequence holds: trade area, demand, traffic, competition, cost, then a ranked forecast you can put in front of finance without flinching. The tooling only changes how many sites you can push through it.
That is the part we built. See how GrowthFactor scores a site, or compare plans if you already know what you need.