Retail void analysis finds the categories a trade area wants but does not have. Landlords run it to build a tenant prospect list for a vacancy. Retailers run it to rank markets worth entering. Two different calculations answer to the name, one counting stores and one counting dollars, and they do not always agree.
A leasing team has 4,200 square feet going dark in November and needs a list of chains to call. An expansion team has budget for six openings next year and a map with no obvious next market. Both teams call what they're doing void analysis. There's a real chance they're running different math and don't even know it.
What is retail void analysis?
Retail void analysis compares the mix of businesses in one trade area against a benchmark area to find which categories are missing. What you get back is a list of absent business types, weighted by how well the local demand base would support them. Landlords use that list to target tenants. Retailers use it to target markets.
The idea is old. The term is not. Mapping software is what turned it into a named method. Esri's tutorial on the technique sets up a reference area, one you already understand because you've got a store there or you know it performs, and an analysis area you're considering. Line up the business categories in both, normalize for population, and whatever's present in one and missing in the other, those are your voids.
That's the version most analysts learn. It's not the only version being sold.
Two calculations wear the same name
The count method is void analysis proper. It works off business presence. Does this trade area have a pet supply store, an urgent care, a quick-service breakfast concept? The unit is stores. The answer's a yes or a no per category, normalized against households so you can compare a small town to a metro submarket.
The dollar method is retail gap analysis, also called leakage and surplus analysis. It estimates how much households inside the trade area spend in a category, estimates how much nearby stores in that category actually sell, and subtracts one from the other. Esri publishes the result as a leakage surplus factor running from positive 100 to negative 100. Positive means demand outruns local supply and the money's leaving for somewhere else. Negative means supply outruns local demand and the area is pulling shoppers in from outside.
The distinction matters because the two methods fail in opposite directions. The count method can't see a category that's present but badly underserved, because one struggling store registers the same as five healthy ones. The dollar method can see that, but it depends on two layers of estimation the count method doesn't need. Neither one can see a category that's absent for a good reason.
If a broker or a vendor hands you a void report, the first thing to ask is which calculation is behind it.
Where the numbers actually come from
The demand side usually comes down to household counts inside the trade area, multiplied by category spending per household. That spending figure traces back to the Bureau of Labor Statistics Consumer Expenditure Survey. The most recent complete annual dataset is 2024, released in December 2025.
The supply side is business location and sales data. Public establishment counts come from the Census Bureau's County Business Patterns, most recently the 2023 vintage released in June 2025. Commercial vendors layer their own sales estimates on top of that, because Census doesn't publish store-level sales.
Two things follow from those dates. The government inputs run one to two years behind the market you're looking at, which is fine for a structural read and not great for a fast-moving corridor. And the sales figures in most void reports are estimates built on other estimates.
That second point has a name attached. The retail analyst N. David Milder argued in 2016 that leakage analyses deserve real caution, citing error rates of 20 to 40 percent in business listing databases, especially for small independents, plus annual retail churn of 5 to 30 percent that outdates listings fast. He also made the point that's aged into the biggest one: consumer expenditure data records what households spend in a category, not whether they spent it in a store or online.
None of that makes the method useless. It makes it a screening tool rather than an underwriting input, which is roughly how the good teams already treat it.
Landlords fill a space. Retailers find one.
The mechanical difference comes down to which half of the problem you hold still.
A leasing team, a landlord, or a developer holds the property fixed. One trade area, every category. The sweep runs across a retailer database, and what comes out is a list of named brands that are absent from the trade area and whose typical site profile matches what the center offers. Vendors in this lane score the fit and, in some cases, hand over the broker contact for each chain. That list is a call sheet.
A retailer, a franchisee, or an expansion team holds the category fixed. One category, yours, every trade area. The sweep runs across markets, comparing category density per household in each candidate market against a benchmark market where your format already works. What comes out is a ranking of markets.
That's the part teams get wrong. Tango Analytics states the limit plainly in its own guidance: void analysis identifies market-level opportunities, not specific locations. A ranked market list is where the site work starts, not where it ends. You still have to draw a trade area for each candidate address, check the overlap against your existing stores, and score the parcel itself.
Skip that step and a good market read turns into a bad lease.
Four tests a void has to survive
Is the demand estimate local, or a national average dressed up? Category spending per household varies with income, age mix, and household size. A model that applies a national table to a trade area whose demographics diverge from it will invent demand. Check whether the vendor adjusts for local characteristics or just multiplies.
Is the spending leaving the area, or leaving the map? This is the failure mode that grew teeth. When a household orders the category online or has it delivered, the expenditure data still counts that as spending in the category, and the local store base still shows zero sales. The model reads a void. What's there is a customer already being served, with no storefront involved.
Has somebody already tried this here? A closed unit in your category on the same corridor is the single cheapest piece of evidence you'll find, and it's not in the model. Pull it before you get excited. A gap where an operator failed two years ago is a scar, not an opening.
Can you get a site that works? Zoning, parcel size, parking ratios, drive-through restrictions, and rent all sit outside the analysis. A category can be genuinely absent and genuinely unbuildable in the same trade area. Vendors in this space make the point themselves: an empty spot can mean unmet demand, or it can mean the market already tested that spot and the demand wasn't there.
There's a fifth test worth adding for multi-unit operators. If the void sits next to your own stores, some of the revenue you'd capture is revenue you already have. Cannibalization analysis answers that before the site tour, not after.
The trade area boundary decides the answer
Every number above depends on a shape somebody drew. Change the shape and both the demand estimate and the competitor count move with it.
A three-mile ring and a ten-minute drive-time polygon around the same address routinely return populations that differ by a factor of two or more, because the ring counts households the freeway cuts off and the polygon doesn't. Run the void analysis on the ring and you get one answer. Run it on the drive time and you get another. Neither one is wrong in the abstract. They're answering different questions, which is exactly why mixing boundary methods across a candidate list produces a ranking nobody should trust.
Pick one method, match it to how your format actually pulls in customers, and use it consistently across every market you compare. A ranked list built on inconsistent boundaries is sorted by methodology, not opportunity.
What tight vacancy does to a void
Finding a void and getting into it are two different problems, and the second one's gotten harder.
Cushman and Wakefield put national retail vacancy at 6.0 percent in its Q2 2026 MarketBeat, published in July 2026, with average asking rent at $25.65 per square foot, up 2.2 percent year over year. CoStar's May 2026 forecast has vacancy peaking in the mid-4.4 percent range before stabilizing.
Those two figures don't match. The likely reason is inventory scope: Cushman and Wakefield has long reported on a shopping-center universe that leaves out malls and freestanding single-tenant retail, while CoStar tracks a broader base that includes them, and freestanding space under long-term net leases sits close to fully occupied. Call that the conventional explanation rather than a settled one, because none of these firms publishes a side-by-side reconciliation of what each counts. Treat both numbers as directional and anchor to whichever definition matches your own portfolio.
What both agree on is direction. Space is tight and new supply is limited. For a leasing team, that makes a void report more valuable, because a category gap is a reason to call a chain that has options. For an expansion team it cuts the other way: finding the right market is the easy half, and the space in it may not exist at the rent your model assumed. Void analysis tells you where to look. It says nothing about whether you can get in.
Where this fits in a real workflow
Void analysis is a screening step. It narrows a country down to a shortlist of markets, or a category list to a call sheet. The work that follows is where the money gets decided.
That work is trade areas drawn on the road network rather than as circles, candidate sites compared against each other rather than one at a time, overlap measured against the stores you already operate, and a score you can open up and argue with in a committee room. GrowthFactor's market planning handles that part: catchments built on real drive times, population and income read from inside the shape, and a five-site plan that shows what the five sites do to each other. Exploring a market never touches your deal pipeline.
Cavender's used that workflow to move from 9 new stores in a year to 27, evaluating more than 2,000 sites along the way. The void read is what pointed at the markets. The site work is what made the openings hold up.
The tooling does the analysis. Your team still knows the landlord, the local politics, and the deal you can actually get, and that part doesn't come out of a dataset.
If the market read is the piece you're missing, market saturation analysis is the same map read from the opposite end: it answers when a trade area is full rather than when it's empty. Teams that run both stop confusing an empty market with an easy one.
Frequently Asked Questions about Retail Void Analysis
What is the difference between void analysis and retail gap analysis?
Void analysis counts businesses. It compares which categories are present in a benchmark area against the area you are studying and returns the categories that are absent. Retail gap analysis counts dollars. It estimates category spending by households in the trade area, subtracts estimated category sales at nearby stores, and returns the difference. Vendors sell both under the void analysis name, so ask which calculation produced the report.
Does a void in a trade area always mean an opportunity?
No. A category can be absent for a reason that has nothing to do with demand. Zoning may not allow the format, the parcel sizes may not fit your build, an operator may have already tried and closed, or online ordering may already serve the spending the model counted as unmet. Treat every void as a hypothesis and check the supply-side reasons before you tour anything.
What data does a retail void analysis need?
Three inputs. A trade area boundary, which changes the answer more than most teams expect. A business location database covering the categories you care about, tagged by NAICS or a vendor category scheme. And a demand estimate, usually household counts multiplied by category spending per household derived from the Bureau of Labor Statistics Consumer Expenditure Survey. The dollar version also needs a sales estimate for the stores already there.
Can void analysis tell you which specific site to lease?
Not on its own. Void analysis works at the market or trade-area level, so the output is a ranked list of places worth looking at rather than an address. Turning a ranked market into a signed lease takes a separate pass: draw the trade area for each candidate site, check overlap with your existing stores, and score the specific parcel on access, visibility, co-tenants, and rent.
How does GrowthFactor compare to Placer.ai for void analysis?
Placer.ai publishes a void analysis guide and ranks potential tenants for a property using what it calls a Relative Fit Score, drawn from foot traffic and co-tenancy patterns. GrowthFactor covers what comes after the market read: trade areas on real drive times, candidate sites scored against each other and against your existing stores, and the deal tracked through to signing. Books-A-Million, the #2 book retailer in the US, screens more than 3,000 sites a year through it. Plenty of teams run both.