The world's largest Buc-ee's sits in Luling, Texas. 75,593 square feet of brisket, beef jerky, and bathrooms people write actual poetry about. 13,000 people live within a 16-minute drive of it.
It does just fine.
Nobody shopping there is local. Buc-ee's is destination retail: you don't pass it, you drive to it. Same trip you'd make for IKEA, Cabela's, a Costco run, an In-N-Out two exits out of the way. The store is the destination, not a stop on the way to one.
That one fact breaks normal site selection. Regular retail buys rooftops and sells to the people nearby. Destination retail buys a highway and pulls people in from an hour out. So when I geocoded all 71 Buc-ee's looking for a customer profile, there wasn't one: populations span 78×, from a 5,879-person town to a 460,443-person suburb.
Worst aim in retail, if population's the target.
There's a better number hiding in the same data. It barely moves, and almost nobody screens on it.
It isn't the population. The land is. Building a Buc-ee's takes 35 acres, highway frontage, and 120 gas pumps, which only fits in the cheap exurban gaps between metros. Buy 55 of those parcels and the same demographics fall out every time, screened for or not: 2.7 people per household, median age 38.
A side effect, not a target.
What these stores don't share is a customer. What they share is a fingerprint: the handful of numbers Buc-ee's holds nearly steady, and the ones it lets run wild.
The numbers that refuse to move
Three numbers hold steady no matter where they build. Household size: 2.7 people, and 2.22 to 3.08 is the whole range. Median age: 38, mostly 36 to 40. Per-capita retail spend: $10,962. The middle half of the fleet sits between $9,712 and $12,045; the full spread runs $6,583 to $15,076.
You inherit all three from the land. The question is which one you can actually trust.
Median age is the trap. It's so consistent you'd be tempted to filter hard on it. Do that, and you'd reject San Marcos and Ruston. Both of which are upcoming Buc-ee's sites. The consistency is a coincidence, not a screen, and treating it like a screen kills good deals.
Per-capita spend is the one that's different, and here's why I trust it. Burlington came in at $11,651 across a 20× population spread. Buc-ee's at $10,962 across a 78× spread. The populations are all over the map; the per-capita spend won't budge.
It's bolted to two slow things: local incomes and the housing already standing. A metro can double its rooftops in a decade; neither of those follows.
The tightest number of all: retail spend per household, divided by income. 0.37. Across all 71 sites it varies by 7%. That's the screen that rescues the deals raw spend would throw out.
Why does spend predict the trip? Because it's a proxy for the whole corridor. Highways connect regions; if the regions on each end are broke, so is the traffic past your pumps.
So the playbook for any destination concept, in order:
- Highway volume and access. The engine; everything else is downstream.
- Can you physically build it? Deceleration lanes, interchange geometry, utility load.
- Spend propensity, not headcount. Aim for 34-38% of income going to retail. A thin local population isn't a reason to kill the deal.
The town is an accident
You can hear the strategy in their own words. When the Fort Myers deal fell apart, their attorney blamed geometry, not demographics: "The roadways necessary to maintain sufficient access to the property for our volume of traffic just wouldn't work." When I-81 got the green light, their real estate director put it in one line: "I-81 is full of travelers. We are the perfect pitstop on the drive."
Getting that access approved is a two-year gauntlet: traffic studies, DOT sign-offs, interchange geometry, utility fights. The site decides where they build; the market never gets a vote. And it shows in the traffic counts, which swing as wildly as the populations do: 33,000 vehicles a day in Luling, 110,000 in New Braunfels, 200,000 at Goodyear.
Screen on a traffic number and these scatter just as badly.
One caveat before you copy this: Buc-ee's manufactures its own demand off the highway. Most concepts can't. If you run neighborhood retail, your customers really are the local rooftops, and none of this applies.
The sites they haven't built yet
If the pattern is real, it should predict the future. So I scored all 16 announced sites against the typical band, the middle 50% of the open stores.
Five are textbook: Benton AR, Gallaway TN, Mebane NC, Murfreesboro TN, Ocala FL. The other eleven miss on somewhere between one and four metrics, and the two worst offenders are Goodyear and Boerne. They break the pattern in opposite directions, which is the fun part.
Goodyear, Arizona (opens June 22) is too big: 408,395 people, off-the-charts spend and employment, 3.2-person households. By population it's the least Buc-ee's market on the board. But per-capita retail spend? $10,278. Right in the pocket.
Boerne, Texas is too rich: $129,183 median income, $483,557 home value, $16,552 in per-capita spend, way over the line. But household size (2.76) and age (40.2) hit dead-on.
Opposite problems on paper. Then you adjust for income and they're the same store. Boerne: 0.35. Goodyear: 0.34. Both inside the band.
When I re-scored all 16 on that ratio, every single one fell inside the band.
Even the two that broke four rules a paragraph ago.
Economic Pulse
Demand's cooling while input costs run hot. For site selection that's the bad combo: slower sales growth pushing against higher lease and construction costs.
The Week's Closures
Those ~200 West Marine leases are pre-entitled sites: parking, utilities, all of it already in the ground. If you're hunting locations, that dark inventory is a shortcut straight past the two-year entitlement gauntlet.
What I'm Watching
Goodyear's June 22 open. The cleanest test we'll get of whether market size matters at all. It's a 408k-person Phoenix exurb with 3.2-person households, on paper the most un-Buc-ee's market on the board. Watch its per-capita spend and household size, not its population. If those hold like the rest of the fleet, raw trade-zone population is officially dead as a filter; the corridor is the market.
Your rejected deals. Go pull three or four destination deals your committee killed for low trade-zone population. Re-score them on highway traffic (state DOT counts are free), site access, and per-capita retail spend. If they fall in the sweet spot, you didn't have a bad deal. You had a bad metric.
The data. I exported all 71 sites: lat/lon, 16-minute population, household size, age, income, home value, retail spend, per-capita, and the retail-propensity ratio. Download the CSV. 71 rows, no email gate.
-Andrew
Head of Marketing, GrowthFactor