Everyone tells you Buc-ee's is a highway story. Big cheap parcel between two cities, interstate frontage, done. I've read that framing a dozen times. I wanted to see whether the 16-minute trade zones reveal something the highway narrative misses: a demographic screen, a household profile they repeat. Or whether they just confirm what everyone already suspects.
So I geocoded all 71 Buc-ee's locations (55 open stores, roughly a dozen of them small-format originals in the Texas home county, and 16 announced ones) and pulled the demographics inside a 16-minute drive of each.
The population spread confirmed the highway thesis immediately. The smallest trade zone holds 5,879 people (Eagle Lake, Texas); the largest holds 460,443 (Pearland, a Houston suburb). Same beaver, same format, 78× market spread. Total retail spend runs $54 million to nearly $5 billion. Ten-year growth swings from negative to +5.5%. By every variable a site model usually leads with, these stores share almost nothing.
Three numbers do hold constant across all 55 open stores: household size (2.7 people), median age (38), and per-capita retail spend ($10,962). But 2.7 is barely above the US national average of 2.5. Median age of 38 is dead-center. These aren't unusual numbers; they're just what exurban America looks like. The question is whether Buc-ee's is screening for them, or whether they're a byproduct of a parcel filter that keeps landing in the same kind of place.
The answer: the highway thesis is right, but it's right for a reason nobody talks about. Buc-ee's moved the market from the residential ring to the interstate corridor, and the trade-zone demographics are just the fingerprint the parcel filter leaves behind.
Every Size Number Runs Wild
Start with what won't sit still. Across the 55 open-store trade zones, the spread is enormous on every size variable.
Population swings store to store by a coefficient of variation of 90%. (That's the standard deviation as a percent of the average. Ninety percent is enormous.) Total retail spend in the zone: 98%. Ten-year growth: 100%. Total employment: 94%. A spread that wide means the variable is uncontrolled across the fleet; the stores don't cluster on any of these metrics.
Eagle Lake is a rice-and-cattle town an hour west of Houston. Pearland is a half-million-person Houston suburb. Buc-ee's runs a store in each, and the gap between them is the point: market size doesn't determine where they build.
The pattern holds at the extreme. The biggest Buc-ee's on earth, Luling, Texas, at 75,593 square feet, sits on a 16-minute trade zone of 13,339 people. When founder Arch Aplin explained the choice, he didn't cite a market. He cited a road. "It just didn't feel right to build the biggest Buc-ee's anywhere other than Luling," he told a local cooperative. Luling sits roughly at the midpoint of the 197-mile I-10 run between Houston and San Antonio, about 48 miles from Austin. The founder, the trade press, and the map all point the same way: stores on cheap land along major interstates between two sizable cities. The town's population is not the buy.
What this means: If your screening model ranks destination retail by trade-zone population, you're sorting on noise. Move corridor traffic and parcel access to the top of the stack; population is a floor, not a rank variable.
Two Numbers Won't Move
Now the short bars. Three variables sit under a 16% CV across the whole fleet.
Average household size: 2.7 people, a CV of 8%. The range runs only 2.22 (Daytona Beach) to 3.08 (Baytown). If you're screening destination retail, 2.5-3.0 is your band. Median age: 38 years, CV of 10%. Screen for 36-40. Per-capita retail spend (what the average resident spends at retail in a year): $10,962, CV of 16%. That's roughly $9,000-$13,000 as your floor.
The consistency is real. But I want to be honest about what it tells us. A household size of 2.7 and a median age of 38 are what you'd find in most suburban-to-exurban communities nationwide; they're not exotic numbers. If you picked 55 large parcels on interstates between cities, you'd probably land in this demographic neighborhood every time regardless of whether you screened for it.
Per-capita retail spend is a different story. Two independent operators now show the same pattern. The 45 Joann sites Burlington is converting held a median per-capita retail spend of $11,651 across a 20× population spread. Buc-ee's across the same spread: $10,962. An off-price apparel chain and a Texas travel center, and the metric doesn't budge.
Why might this one stick when everything else swings? Per-capita retail spend is a function of income and spending propensity (both slower-moving than population, both anchored by the housing stock that already exists in a trade zone). A metro can double its rooftop count in a decade; the per-capita spending profile of the people living in those rooftops barely drifts. The data only covers two formats, but if per-capita spend keeps holding across completely different operator types, it's a better floor variable than anything in the size column.
So the trade zone has a consistent shape: mid-income, family-structured, prime-of-life. Whether it holds 6,000 people or 460,000. The household doesn't explain Buc-ee's; it describes where Buc-ee's ends up.
What this means: When evaluating any format that draws customers beyond its immediate 16-minute trade zone, screen with per-capita retail spend instead of total trade-zone population or household size. The housing stock anchors it; the parcel doesn't change it. Watch whether it holds as I add more retailers to the comparison set.
What They're Actually Buying
So if the demographics are a side effect, what are they actually buying? The corridor. A Buc-ee's doesn't sell to the residential ring. It sells to the traffic flowing past the parcel; drivers on I-10 between Houston and San Antonio, drivers on I-81 in Virginia, the people seeking a 30-minute break during a four-hour highway push.
To capture that corridor traffic, the parcel has to clear a specific gauntlet. Buc-ee's needs 20 to 35 acres, direct interstate frontage, and clean access for a store that generates its own traffic jam. Room for 100-to-120 fuel pumps and a 50,000-to-75,000-square-foot box. That parcel only pencils in one kind of place: cheap, low-density, exurban land on a highway between two metros.
And "access" isn't a number you sort on; it's a two-year gauntlet of traffic studies, DOT curb-cut approvals, and interchange agreements. That gauntlet is exactly why the parcel, not the market, is the binding constraint. Buc-ee's attorney Jeff Nadalo explained why a Fort Myers, Florida deal died: "The roadways necessary to maintain sufficient access to the property for our volume of traffic just wouldn't work." They killed a site over the road, not the market. When real estate director Stan Beard described a green-lit one (the I-81 store in Rockingham County), he framed it as a corridor play: "I-81 is full of travelers. We are the perfect pitstop on the drive."
The corridor traffic counts reinforce the point. I pulled state DOT AADT (annual average daily traffic) figures for a cross-section of the fleet: I-10 at Luling carries 36,400 vehicles a day; I-81 at Mount Crawford, roughly 50,000; I-95 at Daytona Beach, 76,800; I-10 at Goodyear, 201,400. A 5.5× spread; CV north of 60%. Per-capita retail spend holds at a 16% CV across the same stores. If Buc-ee's were screening on corridor volume, the traffic numbers would cluster the way household size does. They don't. The interstate is a format requirement; the traffic count on it is not the variable they optimize.
So "Buc-ee's ignores market size" isn't quite right. The market is enormous; it just doesn't show up in a 16-minute drive of the parking lot. A Buc-ee's in Luling isn't selling to 13,339 residents; it's selling to the traffic flowing down I-10 between two of the ten largest cities in America.
Which is why the household numbers behave the way they do. Screen hard on "huge cheap lot, interstate frontage, between two cities," and you keep landing in the same kind of place: exurban, family-shaped, mid-income. No matter how big or small the town attached to it. The 2.7-person household and the $10,962 per-capita spend aren't what Buc-ee's is buying. They're the fingerprint the parcel filter leaves behind.
One caveat: roughly a dozen of the 55 open stores are small-format originals clustered in Buc-ee's home county (Brazoria County, Texas). That cluster shares a tight demographic neighborhood, which mechanically pulls household size and median age tighter than a travel-center-only sample would. Excluding them, the CVs on household size and median age rise a few points but stay well under 15%; the fingerprint holds, just less dramatically. I left them in because they're real open Buc-ee's, and the question was what the whole fleet holds constant.
One guardrail before you borrow any of this. Buc-ee's manufactures its own demand off the corridor; most formats don't. If you run co-tenancy or inline retail, your customers are the rooftops and the daytime workers. The 16-minute trade zone is exactly the number you should lead with, and ignoring it would be malpractice. The transferable lesson: measure demand at the scale your format actually draws from. The corridor for a destination store, the trade zone for a neighborhood one. Treat your match variables as a floor to clear, not a target to hit.
What this means: Next time you're underwriting a destination format, build the model in this order: corridor traffic first, parcel feasibility second, household character as a pass/fail gate. Per-capita retail spend is your floor variable. Population is a footnote.
I Scored the 16 Announced Sites Against the Fingerprint
If the fleet has a fingerprint, the announced stores should fit it. So I scored all 16 against the typical band (the middle 50% of open stores on each variable) and counted how many each one breaks.
Five land dead-center, inside every band: Benton AR, Gallaway TN, Mebane NC, Murfreesboro TN, and Ocala FL. Textbook Buc-ee's. The other eleven miss on one to four variables, and the two biggest misses are the most interesting sites in the class.
Goodyear, Arizona opens June 22, and it breaks every size rule at once. 408,395 people in the trade zone, dense Phoenix exurb, total spend and employment off the top of the chart. On a population map it's the least Buc-ee's-looking market they've ever signed. Its household size (3.2) runs above the fleet's 2.5-2.8 band; that breaks too. But its per-capita retail spend is $10,278. Right in the fleet's band. The market is enormous, the household is larger than typical, and the one number that holds is the per-capita figure.
Boerne, Texas breaks the opposite way: it's too rich. Median income $129,183. Median home value $483,557. Per-capita spend $16,552, well above the band. But its household size (2.76) and median age (40.2) are dead-on. It misses on affluence, not on household shape.
Both outliers tell the same story on one variable: per-capita retail spend holds, even when everything else breaks. Boerne's household and age are dead-on despite affluence running hot. Goodyear's household is larger than the fleet's typical band, but the per-capita figure lands squarely in it. The one number that travels with the parcel filter, not the market, is the spending figure.
What this means: Shift Goodyear from a market-size test to a format-viability validation. When opening-week traffic data lands, pull it against per-capita retail spend and household size, not population. If Goodyear sustains the fleet average on those two (despite 78× the market size), rerank your destination-format underwriting; move corridor traffic and parcel feasibility to the top, drop population from a ranking variable to a floor check. You're not watching a test. You're validating a reframe that changes how you score.
Demand Cools at the Top While Costs Reheat Underneath
The read: demand is cooling at the top (confidence down, consumers cutting back) while costs reheat underneath (ISM services prices at a near-three-year high). Labor still firm. For site selection, that's the squeeze: softer sales against stickier occupancy and build costs. The May jobs report drops Friday; ADP's beat hints at an upside surprise.
The Week's Closures
A quieter week for fresh filings, but one iconic box went dark and the 2025-26 wind-downs kept grinding. Closures are the supply side of the same parcel math; every box that goes dark is a pre-entitled site for whoever hunts it next.
What this means: West Marine's ~200 leases are specialty boxes (3,000-25,000 sq ft) spread across 34 states, most in strip centers and freestanding locations where the parking, ingress, and curb cuts already exist. For a tenant chasing hard-to-build locations, that dark inventory is the shortcut around the two-year entitlement gauntlet that makes a Buc-ee's-grade parcel so scarce.
What I'll Be Watching by Edition #40
- Goodyear's June 22 open. It's the cleanest test the announced class offers. A 408,000-person Phoenix exurb where the average household runs 3.2 people is the most un-Buc-ee's market on the board. If it draws like the rest of the fleet, the case is closed: the 16-minute trade zone is the wrong denominator for destination retail, and corridor traffic is the real market. If it underperforms, there's a market size even Buc-ee's can't ignore, and it's worth knowing where that line sits. I'll pull the Google Maps traffic overlay on opening day and check local news for line reports; if Buc-ee's releases any traffic figures (they occasionally do for flagships), I'll run them against the fleet average in edition #40.
- Your last cycle's rejects. Pull three or four destination-format deals your committee killed for low trade-zone population. Re-score them on corridor traffic (state DOT AADT data is free; StreetLight or Replica if you need more granularity), parcel access, and per-capita retail spend. If any of them land in the fingerprint band, they were killed by the wrong variable. That's the cheapest pipeline you'll find this quarter.
- The box keeps growing, the trade-zone logic doesn't. Luling holds the record at 75,593 square feet; a Fort Pierce, Florida store is already drawn up at 76,245. Bigger box, bigger parcel, even more remote exurban land. That should tighten the household fingerprint even as the surrounding markets get more random. Worth a re-run once Fort Pierce and the 2027-28 class break ground.
If you want to cut your own angle on the data, I exported all 71 sites: lat/lon, 16-minute trade-zone population, household size, median age, income, home value, retail spend, and the per-capita figure that keeps holding. Download the CSV (71 rows, no email gate).
—Andrew
Head of Marketing, GrowthFactor