Why we stopped pricing cars by hand

For most of my working life, pricing was a Monday job. A manager sat down with a coffee, a list of stock, and Auto Trader open in another tab, and worked through as many cars as they could before the phones started. On a good morning that was forty cars. We had 1,900.

So most of our stock was priced on the day it landed and then left alone until it had been sat long enough to annoy someone. In between, the market moved and we didn't. We weren't unusual. I think most dealers are still doing this, and I think most of them know it.

The first agent we built did one thing: it looked at every car in stock, every day, against every comparable advert it could find, and told us whether to hold, raise or cut. It wrote its reason down each time. That last part turned out to matter more than the recommendation, because it meant we could argue with it.

And we did argue with it. Early on it was too keen to cut anything that had been sat a while, and it didn't understand that some cars are slow because they're rare. We lost a bit of margin before we taught it the difference. That's in our results page because I'd rather you heard it from us.

What changed once it settled down was not that prices got lower. Some went up. What changed is that every car had a price that was current. Stock turn is up 48% since. The manager still prices cars — but now they price the ten that need a human, not the forty they could get to before the phones rang.

Nobody misses the spreadsheet.

George Manning is the founder of OmegAi and a dealer principal who runs a UK car supermarket group. About →