Ashland, Kansas has 746 people in its trade zone. No Casey's General Store within a 16-minute drive. And a per-capita retail spend of $13,367, higher than the median trade zone around an existing Casey's.
That number is exactly the kind of thing that gets you excited before you've checked it properly.
Casey's told investors on June 24 that it's sitting on a mountain of whitespace: about 75% of the towns under 20,000 people inside its own distribution footprint still don't have a store, per COO Ena Williams. That footprint is the region its distribution centers can supply. The company just closed a three-year plan that added 504 stores against an original 350-store target. It's projecting at least 400 more over the next three, and it walked into that investor day off a record year, net income up 31%. Wall Street bought the framing within days. There's demand out there and the map hasn't caught up.
So I went looking for the demand. First I pulled Casey's complete store network from the chain's own locator, 2,777 locations, queried point by point until no new store turned up. That matters more than it sounds: the public location datasets most analyses lean on carry only about two-thirds of Casey's stores, because their scrapers search outward from cities of 10,000-plus people and Casey's entire specialty is towns smaller than that. Rebuild a whitespace list on incomplete store data and you'll invent whitespace that a real store already fills.
Roughly 2,000 of those stores sit in six core Midwest states: Iowa, Illinois, Missouri, Kansas, Minnesota, Nebraska. The locator's 2,777 squares with the 2,944 stores reported company-wide as of April 30, 2026 once you set aside the Gulf South stores still operating under the CEFCO banner, which the locator doesn't carry. I profiled 53 existing trade zones across that six-state footprint. Then I found 13 towns that fit the company's own whitespace description, under 20,000 people, inside that same footprint, and verified against the complete network that every one sits at least 12 miles from the nearest Casey's. Ashland was one of them.
If Ashland is representative, the story writes itself: the money's already there, Casey's just hasn't gotten around to it yet.
The baseline test that broke Wayfair's screen breaks this one too
Before I trusted Ashland, I did what this newsletter did two weeks ago with Wayfair's eight stores: I ran the same spend and spend-propensity numbers against a group that had no business looking special. Fifteen small towns across the same six states, population under 20,000. Not chosen because they looked underserved. Not chosen because they were near a Casey's. Just random.
If the whitespace towns were sitting on real, unscreened demand, I'd expect their spend numbers to separate from a neutral sample the way Casey's zones separate on income. They don't. Spend barely moves anywhere: $11,800 around Casey's stores, $10,971 in whitespace, $10,634 at random. And spend propensity actually runs backward: 0.150 around stores, 0.179 in whitespace, 0.158 at random. The towns Casey's hasn't built in read highest on the metric that's supposed to find demand. If propensity screened for anything, those towns would have been the chain's first stops, not its leftovers.
Ashland isn't a hidden gem. It's just what a small town in this part of the country spends, whether or not a Casey's has found it yet.
Here's why the ratio can't find demand. The spend number is itself estimated mostly from income. Divide it by income again and the two mostly cancel, and the little variation that survives runs against income, which is why the richest group in this sample, Casey's own zones, posts the worst ratio of the three. That's a property of the math, not a market signal. The Wayfair screen died on the same math two weeks ago, in wealthy suburbs instead of the rural Midwest.
So don't trust a spend-based affordability ratio until you've run it against a neutral, unselected sample. Tight inside your approved deals could mean the screen is working. It could also just be what income-derived spend ratios always look like, anywhere. Ashland proves the second reading.
Casey's zones are 20% richer, and it isn't a metro artifact
So the money argument is dead. Something in that 75% number is still worth explaining, because the whitespace towns and the random towns sit in almost the same place on every variable I checked. Casey's own zones don't.
Look at home value first. Whitespace towns: $167,788. Random towns: $167,603. Those medians come from samples of 13 and 15 towns, so the eerie $185 gap is partly luck; drop one town from each group and the gap can stretch to almost $7,000. What no version of the sample changes is the shape: the two groups never separate by more than about 4%, and income repeats it, $64,470 against $63,947. If Casey's zones were no different from the other two groups, I'd expect them to sit near those same numbers. They don't. Casey's is at $76,998 on income, about twenty percent higher than both, and $227,027 on home value, some $59,000, or 35%, clear of the pair.
Grad-degree share leans the same way but wobbles: whitespace actually reads lower than random, 4.6% against 5.4%. That wobble made me suspicious of the number even before the control below killed it. Age runs the whole thing in reverse. Casey's zones are the youngest of the three groups by several years, whitespace towns the oldest.
The obvious objection: my store sample mixes true small-town Casey's locations with metro-fringe zones, Wichita, Mission, and Oswego among them, and metro money could be manufacturing the whole gap. So I cut the sample down to just the 19 store zones under 20,000 people, the same universe the whitespace list lives in, and reran everything.
The income gap survives. Small-town Casey's zones: $74,698 median household income. Whitespace towns: $64,470. Still 16% clear. Home value narrows but holds ($195,328 against $167,788). Grad-degree share does not survive: it collapses to 4.5% against 4.6%, a tie, which means that bar in the chart above was a metro artifact all along. Strike it.
And one variable gets stronger, not weaker, under the control. Even inside the under-20,000 universe, the median Casey's trade zone holds 9,701 people. The median whitespace town holds 3,512. The towns Casey's actually built in are nearly three times the size of the towns it says it's missing. Age narrows under the same control but keeps its direction, 41.6 against 44.0, so call it a lean, not a floor.
Put a threshold on it instead of a median and the floor gets easier to see. Back on the full 53-store sample, metro-fringe included: 45 of 53 Casey's zones clear $65,000 in median household income, against 6 of 13 whitespace towns and 6 of 15 random towns. It's not the metro stores carrying that either; inside the 19 small-town-only zones, 15 of 19 clear the same bar. Same shape at $180,000 in home value: 44 of 53 store zones clear it, against 4 of 13 whitespace towns and 5 of 15 random towns.
Poverty rate never resolves into a pattern either direction. Leaving it as noise, not forcing it into the thesis.
If a chain's growth story leans on a whitespace count, ask what "whitespace" is actually measuring before you build a forecast around it. Casey's isn't wrong that 75% of those towns lack a store. It's the leap from "lacks a store" to "qualifies for one" that the raw count skips, and the skip is worth exactly the size of the population and income gaps above.
I rebuilt the 75% claim from scratch. It checks out, and it's 890 towns
Everything above tests whether the whitespace is worth building. It's fair to ask the prior question: is the 75% number even true? Nobody outside Casey's real estate team had checked, so I did.
The full store network was already in hand. For the town side I took every incorporated place in the Census with between 100 and 20,000 people, across all nineteen states Casey's operates in, within 500 miles of one of its three self-distribution centers in Ankeny, Iowa; Joplin, Missouri; and Terre Haute, Indiana, the same "distribution footprint" Williams scoped her number to. That's 9,502 towns. Then I measured every one of them against the complete 2,777-store network.
7,081 of those towns, 74.5%, have no Casey's within five miles. Williams said about 75%. Rebuilt from the outside, on public data, her number comes in within half a point. The other Investor Day soundbite survives the same treatment: Casey's says roughly 2,000 of its stores sit in just six states, and the locator network puts 1,992 of 2,777 in those six. Whatever else the whitespace framing is, it isn't invented. The serve radius is the one judgment call, so I moved it: at 3 miles the unserved share is 80%, at 5 miles 75%, at 10 miles 51%. Their number corresponds to a town having a store within about five miles, which is a reasonable definition of "has a Casey's."
Here's the part the soundbite skips. Apply the floors this piece just found to all 7,081 towns, using town-level Census numbers rather than the trade-zone profiles above, so the bar is the median of the towns Casey's already serves: population at or above 1,919 and household income at or above $63,235. 890 towns clear both. A median bar is strict by construction (half of Casey's own towns sit below it on any single variable), so I also ran the 25th percentile of Casey's own towns, the level most of its existing footprint clears: 2,161 towns make that cut.
So the honest version of the whitespace claim runs: 75% of small towns in the footprint lack a store, and somewhere between 900 and 2,200 of them look like towns Casey's actually builds in. That's still a deep pipeline. The new three-year plan needs roughly 200 new-build sites, and even the strict floor covers that four times over. But it's an order of magnitude smaller than 7,081, and the gap between those two numbers is the difference between a real estate strategy and an investor-deck flourish.
So what is Casey's actually missing?
Put the three findings together and the 75% number reads differently. The towns aren't sitting on hidden spend the map missed; money per person is flat across the whole region, regardless of store presence. What correlates with the built footprint is people and income: nearly three times the trade-zone population, 19% higher household income before the metro-fringe control, 16% after it.
That's a demographic pattern, not a proven cause, so I went and checked the most obvious alternative: maybe it isn't income at all. Maybe someone else already built the gas station.
I pulled the nearest businesses around all 13 whitespace towns and a sample of ten built Casey's zones. None of the ten built towns have a competing fuel brand parked nearby. Five of the thirteen whitespace towns do: a Huck's in McLeansboro, a Kwik Shop in Ellsworth, MFA Oil and a Flash Market in Steele, a Sinclair in Golconda, a BP in Ord. For those five, the floor isn't the story. A rival already sits on the obvious corner.
The other eight don't. Ashland is one of them: no incumbent fuel brand nearby, just a Health Mart and a Do It Best, the same kind of thin commercial mix a built Casey's town shows. No spend advantage, and 746 people in the whole trade zone against the 9,701 a median small-town Casey's draws from. For that larger half of the list, the population and income floors are still the best explanation I've got.
I still haven't ruled out distribution-center radius, parcel availability, traffic counts, and interchange access, the things a fuel-pad broker screens before any census number does. Nor the legacy highway routes Casey's happened to grow along before it ever screened anything.
And the nearest-businesses lookup behind those competitor checks only surfaces the closest 20 or so businesses. "No competitor found" means none turned up close by, not that one is confirmed absent.
Some of that 75% will get built. Casey's fiscal 2027-29 (FY27-29) plan calls for at least 400 more stores, split roughly evenly between new construction and small acquisitions, aimed partly at exactly this kind of infill. But the 7,081-town count splits three ways: towns with a rival already on the corner, towns that wouldn't clear Casey's own revealed floor today even if the map showed empty space tomorrow, and the 890 to 2,161 that actually look like Casey's towns.
And the acquisition half of the plan is crowded. Sunoco has bought 140 stores in three separate deals since January, and Cumberland Farms closed on Coen Markets' 54 sites on June 1, all part of a broad 2026 rebound in convenience-store dealmaking after a quiet 2025.
I can't see whether any target list overlaps Casey's. But the buyer pool for small regional operators with sites that already clear somebody's floor is visibly deeper than it was a year ago, and that's the pool Casey's is counting on for half its 400.
Discount any whitespace count in an investor deck to the subset of towns that clear the chain's own revealed floors, and read the acquisition half of a store-growth target as competing in an M&A market, not as new pads. The towns that get Casey's roughly 200 ground-up builds will look like the Casey's bars in the chart above, not the whitespace bars.
If you're a QSR or small-format tenant pricing Casey's-adjacent pads (Casey's is the fifth-largest pizza chain in the country, so pizza operators most of all), size the real new-construction pipeline at half the headline and screen the town first. The $74,698 and 9,701 floors are Midwest convenience-store numbers, from the small-town-only control. But the method travels to any format and region: build the floor from a chain's built stores, then test its whitespace claim against a neutral sample of towns nobody picked.
All 53 stores and all 28 comparison towns sit in the six legacy Midwest states. Casey's newest growth engine, the $1.145 billion acquisition of Fikes Wholesale, parent of the CEFCO convenience-store banner, closed November 2024 and put it into Texas, Alabama, Florida, and Mississippi for the first time. Nothing in this study covers those markets.
Three things that would tell us whether this floor holds
- Which towns Casey's actually builds in next. The FY27-29 plan names a number, not a list. When the next batch of site announcements arrives, check them against the income and home-value floors above. If Casey's keeps clearing them, the floor keeps its predictive value, whatever mix of income, traffic, and land economics actually produces it. If it starts building meaningfully below them, the floor was softer than this data suggests, or the acquisition pipeline is pulling in weaker sites than the organic one did.
- Whether the Gulf South screens the same way. CEFCO and Fikes came with their own site history before Casey's owned them. Once there's enough of a sample, a same-style profile of the Texas and Florida stores will show whether "Casey's floor" is a company-wide standard or a Midwest habit the newer markets haven't inherited yet.
- Whether any of these 13 towns actually gets a Casey's within the next year. Ashland and a dozen towns like it (the full list is in the CSV below) are a standing test now. A future announcement hitting one of them is the cleanest real-world test available: does the company's own stated whitespace convert, or does it keep sitting open because the real floor still isn't cleared.
Economic Pulse
Redbook accelerated to +11.5% and fuel is finally falling, which on the surface is a great fortnight for a rural gas-and-pizza chain. Then read the Fed minutes again: the spending is increasingly financed by credit at the bottom of the income distribution, and the customers "facing disproportionate pressures from elevated gasoline and grocery prices" are, almost verbatim, the median resident of the 81 trade zones above. Underwrite the income tier of a trade zone, not the national sales headline. The floor this piece found is consistent with Casey's doing exactly that.
The fortnight's closures
I exported all 81 trade zones behind this piece, the 53 Casey's stores, the 13 whitespace towns, and the 15 random towns, with every variable above: population, income, home value, age, spend, propensity. If you want to check my math, run your own floors, or watch which of the 13 whitespace towns gets a store first, it's all here.
-Andrew
Founding Team Member, GrowthFactor