The lane lies. Your sold history doesn't.

The problem

Every store has a car it should never have bought. It looked right on the lane. The book said it was fine. The buyer had bought one like it two years ago and it sold. Ninety days later it is on the wholesale truck at a four-figure loss, and nobody ever wrote that loss down because it was spread across a buy fee, a transport bill, two months of floor plan, a reconditioning ticket and a price cut.

The reason is not that buyers are careless. It is that the information a buyer needs at the moment of the bid lives in four places that do not talk: the DMS knows what sold and what it grossed, the inventory tool knows what is on the lot, the auction portal knows what is running, and the retail market knows what the car will actually bring. The buyer carries all four in his head and bids in seconds.

app.dealercortexai.com / sold-replacements

What it costs

• The wrong car. A model the store has not turned in 45 days in two years, bought because the lane was hot. It ages, gets cut, gets wholesaled.
• The right car at the wrong price. The store wins the bid by $600 because the book said so. The retail market for that trim at that mileage in that radius never supported the book.
• The car with the hidden announcement. A green light that also says yellow light. A condition report with "confirmed codes" buried under mechanical. A branded title in the structured field, not the announcement text.
• Two buyers, two doctrines. The same store, the same sale day, two people buying by two different sets of rules, and an owner who cannot tell which one is right.

How DealerCortex handles it

Sold Cars and Replacements. Every car the store sold in the last 30 days becomes a cohort: year, make, model, trim, mileage band. For each cohort the board shows days to turn and front and back gross from the store's own DMS history. Then it goes and finds the replacements running at Manheim this week, lane by lane, with the light, the condition grade and the MMR. The buyer opens it on sale morning and reads: you sold twelve F-150s at $2,900 front in 31 days; here are the ones on the block today that match.

The Pre-Sale Buy List. For each sale day, every car on the run is scored against the store's model set and rules. Red lights, as-is, branded titles, engine, transmission and check-engine announcements are stripped before a buyer sees the list. The system reads Manheim's own announcement text, so a car that says both green and yellow is treated as yellow. The board is ready before 7 AM Eastern on every sale day.

Profit Gap and Max Bid. For any car on the run the system builds the full cost stack: hammer, buy fee, transport by actual mileage from the sale location to the store, reconditioning estimated from the condition report, carrying cost at the store's turn rate. It sets that against the retail market for that exact car at that mileage from live retail listings and shows what is left at 80, 85, 90, 95 and 100 percent of market. The result is a maximum bid in dollars.

Store buying rules. The store encodes its doctrine once: ceiling as a percent of MMR, condition-grade floor, mileage limit, radius, what to do with frame cars, which years and bodies it wants. Every board and every bid applies the same rules. Two buyers at the same store cannot buy two different ways.

The OVE Buy Agent and Auction Search. A daily scan of every live OVE car tiers them: clean cars under MMR in tier one, cars with soft flags in tier two, heavy sheet-metal cars with a clean history in their own tier, mechanical and storied-title cars excluded. Auction Search scans the whole block with saved searches per buyer.

What changes in the first 30 days

The store's sales history loads on day one. By the first sale day the replacement board is live. The buyer bids with a max number and the owner sees, cohort by cohort, whether the store is buying the cars it turns. The wholesale truck starts to leave lighter.