Five levers, five closing gaps, and the model that doesn’t depend on either.

Most operators don’t know they’re running an arbitrage.

They think they’re running a restaurant.

But if you strip the floor away and look at what’s actually generating the money, what you find is a stack of bets — five of them — each one capturing a gap the operator didn’t create and won’t control when it closes. And those bets sit in conditions that are highly volatile.

That’s [Transactional Arbitrage].

Five levers. Each one a real arbitrage in the formal sense: a gap, a capture move, and an exit risk that’s already counting down.

[Concept Arbitrage] A category gets hot. The cycle picks a winner — a format, a protein, a daypart, a price point, or whatever else the future trend minds can dream up — and operators rush in to ride the heat. The build doesn’t have to be great. The category is doing the selling.

Gap: the format is more interesting than the operation underneath it.

Capture: open inside the wave, let the trend pull guests through the door.

Exit: the trend cools, the operation gets exposed, and the floor has to carry weight it was never built to carry.

Tell: when the concept’s name is doing more work than the kitchen, you’re inside this lever.

[Location Arbitrage] The address generates the traffic. Power center, mall pad, airport, downtown lunch corridor — the flow exists whether the operator earned it or not. The play is to park in the flow and convert.

Gap: foot traffic the operation didn’t build is a different asset than guests the operation pulled.

Capture: take the lease, take the volume, let the door do the work.

Exit: rent re-prices on renewal, the corner shifts when an anchor leaves or a road project lands, a competitor opens nearby and splits the funnel — and the pull was never yours to begin with.

Tell: ask what happens to the volume if a comparable operator opens across the street. If the answer is “we lose half of it,” the location is the operator.

[Math Arbitrage] This is the price-to-quality lever. The plate is engineered to read as value before the guest experiences it — portion size, price point, photo on the menu, the number that lands cheaper than the comp set. The math is built on the spreadsheet, not on the line.

Gap: engineered price math is a different thing than experienced value.

Capture: the math reads as value on the way in, and most guests don’t run the comparison until repeat visit number three.

Exit: input costs rise and the math stops working, a competitor out-engineers you, or the guest learns the difference between cheap and worth it — and trades up.

Tell: when guest frequency drops faster than guest count, the math worked once and stopped working.

[Attention Arbitrage] Paid awareness is cheaper than earned reputation, especially early. The play is to outspend the silence — paid social, paid search, influencers, geo-targeted promos — until the funnel fills and the floor turns. The spend manufactures the trial that reputation usually has to earn.

Gap: bought attention costs less today than the years it takes to build word-of-mouth.

Capture: turn on the spigot, fill the room, let volume do the validating.

Exit: ad costs climb, the channel saturates, the algorithm shifts, or the spend stops and the room goes quiet — because the reputation underneath was never built.

Tell: kill the ad spend for sixty days. If the floor empties, the spend was the operation.

[Replication Arbitrage] Take a floor-level product, clone it across many doors, and footprint starts looking like scale. The plate isn’t great, the service isn’t great, but the consistency of mediocre at scale usually beats the inconsistency of trying to be good in a hundred buildings.

Gap: replicable mediocrity reads as a system, even when the system is just a low ceiling enforced everywhere.

Capture: pick a floor most guests will tolerate, lock it, and put it in as many buildings as the capital allows.

Exit: a real operator wakes the market up to what the category can actually feel like, the floor falls below the threshold guests will tolerate, and the units start de-leveraging one market at a time.

Tell: the best unit and the worst unit in the system are within ten percent of each other on every metric. That’s not consistency — that’s a ceiling.

Every one of these works. That’s the trap.

They work until the gap closes. And the operator who built the business on the gap doesn’t have anything underneath when it does.

The alternative isn’t a sixth lever. It’s a different model entirely — one that doesn’t depend on a gap to generate the result.

That’s the work.