Why a human-calculated car valuation beats an algorithm — every time

Almost every online car-buying service uses an algorithm. You enter the reg, a computer pulls historical sale data, applies a few rules, and spits out a number. It's fast. It's cheap to run. And it's almost always lower than what your car is actually worth.

At Carmora we don't do that. Every valuation we give is calculated by a real person, looking at the actual market the day you ask. Here's why that matters.

Algorithms are trained on averages — your car isn't average

An algorithm sees a 2019 Volkswagen Golf 1.5 TSI Match with 60,000 miles and matches it against thousands of historical sales of similar cars. It then takes the average, knocks a margin off, and offers you that.

The problem is your specific Golf isn't the average. It might be a higher-spec variant that wasn't common in the training data. It might be one of the last of the line before a refresh, which makes it more desirable. It might have a panoramic roof that adds £400. It might be in the colour everyone wants. The algorithm doesn't know any of that.

The used car market moves week to week

Used car prices move faster than most people realise. A surge of fleet returns can drop the price of one model by £500 in a fortnight. A high-profile recall on the petrol variant can spike demand for the diesel. A celebrity tweet about a particular brand can move metal.

Algorithms typically retrain monthly or quarterly. A real person reading the trade press and watching the auction halls knows about these shifts the day they happen.

Condition is everything — and condition is subjective

Two identical cars with identical mileage can be £2,000 apart on price because of condition. A car that's been garaged its whole life with one owner is night-and-day different from a fleet hire-car. The interior, the smell, the wear on the steering wheel, the kerb damage on the alloys — none of this is in the data the algorithm sees.

When you send Carmora photos and a brief description, a person looks at them and prices what they actually see, not what an algorithm guesses.

Algorithms are designed to underbid

This is the cynical bit. The big online-buyer algorithms aren't tuned to give you the fair price — they're tuned to give you a number low enough that the buyer can profit even after some on-the-day downgrade.

That's why so many ‘we buy any car’ style services give a high initial quote on the website, then reduce it when you arrive at the forecourt. The algorithm doesn't expect to pay what it quoted.

Carmora's quote is the quote. We don't downgrade on the day for age-related wear, light kerb damage on alloys, or other ‘used car’ things. If the car is materially different from how you described it — for example, undisclosed accident damage — that's a different story. But age-related marks are baked into the original quote.

What this means in practice

You can prove this to yourself for free. Get a quote from one of the algorithmic services, then enter the same reg on our home page. Compare.

You'll almost always find Carmora a few hundred pounds higher — sometimes more on cars that don't fit the algorithm's neat bell curve. Try it here.