It was almost two in the afternoon and I had been driving since a little after six. Two places sat across the road from each other at the exit, and I parked in the gravel lot of the one on the north side with the engine off, doing the thing I always do, which is stall. On my side, eleven vehicles. Four work trucks with ladder racks, a sedan with a dealer frame from somewhere a long way off, two cars with booster seats visible through the back glass. On the other side, three cars, all clean, all parked in the row closest to the door.
I picked the busy one, obviously. The chicken fried steak came out dry in the specific way that means it had been sitting under a heat lamp since the lunch rush, and the gravy was doing all the work. Two weeks later I came back through the same exit and ate at the quiet one, and it was better. Not slightly. That is the part of this I keep having to relearn.
The count matters less than the spread
A full lot at noon tells you almost nothing. Everyone is eating at noon. Full at noon is a fact about the clock, not about the kitchen. What I look at now is whether the cars are spread across the hours. If I pull in at twenty to three and there are still six vehicles sitting there, that is a place people drive to because they want to be there, not because their lunch break happened to land them in front of it.
Empty at noon is a genuine warning and I do not argue with it anymore. Half full at three is a quietly good sign. Full at both is the thing you are hoping for and it is rare enough that I have stopped building a route around finding it.
Work trucks are a decent signal and a lazy one
Everybody repeats the rule about following the work trucks. I repeated it too, for years, with more confidence than I had earned. It is not wrong. People who eat lunch out five days a week are voting with their own money and they are not going to spend it twice on a bad sandwich.
But I have started paying attention to what the rule actually measures, which is speed, portion size, and price. A crew has half an hour and needs food that arrives fast and fills them up. Those are real virtues. They are not the same thing as the food being good, and I spent a lot of mediocre lunches confusing the two.
The lots I trust most now have work trucks and also something else. A car with a booster seat. An older sedan with a placard hanging from the mirror, parked near the door. Somebody’s motorcycle. That mix means the place is serving more than one kind of hunger, and that usually means somebody back there still cares about more than turnaround.
What I got wrong about clean lots
For a long time I read a tidy lot with fresh striping as a warning. Scrubbed, corporate, food shipped in frozen on a pallet. I was being a snob and I would rather say so than pretend I arrived at any of this gracefully.
On a long empty stretch one spring I stopped at a place with a repaved lot and a bright new sign, fully braced for disappointment, and got the best bowl of chili I had all year. The owners had spent money on the lot because the lot had been a problem. People kept scraping their bumpers coming in off the shoulder. That was the entire story. I had built a whole theory of the world out of asphalt.
The reverse bias is just as bad. Weeds pushing up through the gravel and a sun bleached sign do not certify anything. I have had genuinely bad food in charming ruins. Decay is not a credential.
Time of day changes what you are looking at
Before seven in the morning, the lot is mostly telling you who works nearby. At noon it tells you who is on a break. Around three it tells you who chose this. After eight it tells you whether the town has anywhere else to go, which in a lot of places is a shorter answer than you would think.
I try to name the hour to myself before I read anything into the cars. It stops me from taking credit for noticing that a breakfast place is busy at breakfast.
The tells that have held up
A few small things have survived years of testing. Staff cars parked at the far edge of the lot, away from the door, because somebody told them to leave the good spots for customers. A handwritten sign taped inside the glass saying they are out of something, which means the menu is tied to what actually came in that week. The smell in the lot, which is either specific or it is fryer oil and nothing else. And the worn spot on the door handle, the place where the paint is gone down to metal, because that takes years of hands and cannot be faked by a sign company.
None of these are proof. They are just better than counting cars.
What I actually do now
I give myself two minutes in the lot, and then I go in and look at three things before I sit down. Whether the drink station is wiped. Whether anyone looks up. Whether there is a table of people who are clearly not in a hurry, because locals lingering is worth more than strangers eating fast.
And then I accept that I will be wrong maybe a third of the time. That number used to bother me. Now it seems about right for a decision made from a car window in a town I will never see again, and being wrong a third of the time is a cheap price for not eating at the same place in every state.