How to find the best cars at car auctions

Every buyer has bought a car that looked fine in the photos. The listing had thirty images, the announcements were clean, the price was under book, and it still cost two thousand dollars more to retail than it should have. The problem is rarely that the information was missing. It is that a buyer has forty minutes and two hundred lots.

This is the method that survives that constraint: what to read first, what the photos actually tell you, how to price the work before you bid, and how to set a number you will not talk yourself out of when the bidding gets quick.

Read the announcements before the photos

Photos are seductive and announcements are boring, which is exactly backwards from how useful they are. The announcements decide whether a car is worth your attention at all.

Three things to check before anything else:

• Title brand. Clean, salvage, rebuilt, flood or lemon changes both what you can retail the car for and who will finance it. A rebuilt title is not automatically a bad buy, but it is a different buy, at a different number.
• Odometer disclosure. "Actual" is what you want. "Not actual" or "exempt" means the reading is not warranted, and you are pricing a car whose real mileage you do not know.
• Structural and mechanical announcements. Frame damage, airbag deployment, and "as is" declarations tell you what the auction is refusing to stand behind.

If any of those disqualify the car for your lot, stop. You have spent thirty seconds. The buyers who get hurt are the ones who fall for the photo set first and then rationalize the announcement.

Work the photo set as a set, not as pictures

A single photo tells you almost nothing. The same photo compared against the other twenty-nine tells you a great deal. What you are looking for is inconsistency.

Paint that does not match. Look along the side of the car in the daylight shots. Panels repainted separately reflect light differently: a door slightly flatter than the fender beside it, a quarter panel a half-shade off. That is previous repair work the announcements did not mention.

Panel gaps that change width. Follow the gap between hood and fender from front to back. Factory gaps are even. A gap that tapers means a panel has been off the car, and something moved it.

Tires that disagree with each other. Four different brands is a car that was maintained by whatever was cheapest that week. Uneven wear across one tire's width points at alignment, and alignment problems often point at something structural.

Photos that are missing. A set with no undercarriage shot, or one wheel photographed from a careful angle, is a set with something to hide. Absence is information.

The interior tells you the truth about mileage. A worn driver's seat bolster, a shiny steering wheel and worn pedal rubbers on a car showing 40,000 miles do not go together.

Price the work before you bid, not after

The single most expensive habit in wholesale buying is deciding the number first and discovering the reconditioning later. Price the work from the photos, panel by panel.

Rough bands most buyers can carry in their head:

• Cosmetic scuffs and light scratches: a few hundred dollars a panel, less if it polishes out.
• A bumper cover that needs refinishing: more than a scuff, but bumper covers are replaceable parts and the cost is predictable.
• A creased door or fender: several times a scuff, because it is bodywork and paint rather than paint alone.
• A quarter panel: the expensive one. It is welded to the car, so it is repaired rather than replaced, and blending the paint pulls in the panels either side.
• Wheels, glass and lamps: individually small, collectively significant. Four curbed wheels and a cracked headlamp add up faster than buyers expect.

Then add a contingency for what photographs cannot show. Buyers who have done this a while carry a fixed allowance for the surprise, and the discipline is that the allowance does not get spent in advance to justify a bid.

Set the walk-away number backwards

Work from realistic retail, not optimistic retail. Take what the car actually sells for on your lot, in your market, in the time you are willing to hold it. Then subtract, in order:

• reconditioning, including the contingency;
• transport;
• auction fees and buyer premium;
• floorplan for the days you expect to hold it;
• the gross you need for the deal to be worth doing.

What is left is your maximum bid. Not your target, your maximum. Write it down before the lane runs. The purpose of writing it down is that bidding is fast and competitive and your judgment is worse inside it than outside it.

Anchor on comparables, not on book

Book value is a starting point that averages away everything specific about the car in front of you. Three live listings of the same year, make, model and trim, with similar mileage, in your region, are worth more than any single number. Sold prices are worth more than asking prices, and days on market tells you how long your money will be tied up.

The real constraint is volume

Everything above is what a good buyer already knows. The reason it does not always happen is arithmetic: two hundred lots, forty minutes, and thirty photos each is six thousand photographs. Nobody reads six thousand photographs.

So buyers triage. They skim the hero shot, glance at the announcements, and lean on the price. That is the moment the expensive mistakes are made, and it is the part of the job worth automating.

This is what we built Bucepha to do. It reads the listing's photographs, up to sixty, marks the damage it finds, consolidates the same defect seen from four angles into one repair rather than four, prices the work, and scores the opportunity against market value. It is honest about uncertainty: photos it could not read lower the confidence, and a wide repair range lowers the score even when the midpoint looks attractive.

It is not an inspection, and we do not pretend otherwise. Photographs cannot show frame damage, flood history or a tired engine. What it does is put the twenty lots worth your attention in front of you, so the forty minutes go to the cars that deserve them.

The Starter plan is 300 credits a month with the first month free, and a photograph costs 5 credits. If you want to see it against a real lane, start there.

The short version

• Announcements first: title, odometer disclosure, structural declarations.
• Read the photo set as a set. Look for inconsistency, not for damage.
• Missing angles are information.
• Price reconditioning panel by panel, then add a contingency you do not spend in advance.
• Build the maximum bid backwards from realistic retail, and write it down.
• Anchor on comparable sold prices, not on book.
• Spend your limited attention on the cars that survive all of the above.