How to price a used car so it actually sells

Why the median sold price is the wrong target for an independent dealer, what the fortieth percentile is and why it matters, and how to adjust for mileage without guessing.

Updated 1 August 2026 · 6 min read · Written for UK independent dealers

There are two questions hiding inside "what is this car worth", and confusing them is why a lot of stock sits.

The first is what is it worth — the number a valuation book gives you, the average of what similar cars change hands for. The second is what will move it — the number that gets it off your forecourt inside the time you can afford to wait. They are not the same number, and for an independent dealer the gap between them is where the business lives.

Start with what actually sold, not what is advertised

The single biggest improvement most dealers can make is to stop pricing against advertised prices and start pricing against sold prices.

Advertised prices tell you what other dealers hope to get. Some of those cars have been sitting for four months at a price nobody will pay. Averaging them gives you an optimistic number that feels reassuring and prices you out of the market.

Sold prices tell you what buyers actually paid. That is the only number with any authority.

Where do you get them? A market data service will give you sold records. If you do not have one, the crude version is to watch specific cars disappear from the portals and note where they were priced when they went — laborious, but better than guessing.

Match the specification properly

A "2019 Golf" is not a specification. There are enormous differences between a 1.0 TSI Life and a 2.0 TSI R-Line, and averaging them produces a number that fits neither.

When you look for comparable sales, match in this order of importance:

1. Make, model, engine and trim. The badge on the boot is not enough.
2. Year — and be careful with facelifts, which can create a real price step mid-generation.
3. Gearbox. Manual and automatic versions of the same car can differ by close to a thousand pounds, and in some segments the manual is nearly unsellable.
4. Mileage band — within about 20,000 miles.
5. Fuel type, obviously, but also be alert to diesels in urban markets.

If you cannot find at least four genuinely comparable sales, widen the year band before you widen the trim. A same-trim car a year older tells you far more than a same-year car in a different trim.

The fortieth percentile, and why the median is a trap

Here is the part that most guidance misses.

Suppose you find twenty comparable sales. The median — the middle one — is the obvious target. It is also, for an independent dealer, usually slightly too high.

The reason is who else is in that dataset. Those twenty sales include franchised dealers with manufacturer-approved warranties, finance offers the customer trusts, and a brand name on the forecourt. They include car supermarkets with enormous choice and a national marketing budget. Those sellers command a premium, and their sales pull the median up.

You are none of those things. What you have is the car and the price. So the number that actually moves your car is not the middle of that distribution — it is a bit below it. In practice, somewhere around the fortieth percentile of recent sold prices is the level at which a no-brand independent shifts stock reliably.

That is not "underselling". It is pricing to the position you actually occupy in the market. Mid-pack pricing is a luxury that comes with a brand, and paying for that luxury with an extra six weeks of holding cost is a bad trade.

The last row matters too. Being the cheapest is not a strategy; it is a way of donating margin and of signalling to careful buyers that something might be wrong.

Adjusting for mileage

Two cars of identical specification with a 40,000-mile difference are not worth the same, and the market's median assumes an average odometer.

The workable approach is to compare against expected mileage for the car's age. Around 8,500 miles a year is a reasonable UK average. A four-year-old car would be expected to have covered roughly 34,000 miles.

• Materially below expected: a modest upward adjustment is justified, but be careful — very low mileage on an older car raises questions about long standing periods, and buyers know it.
• Materially above expected: adjust down, and adjust more sharply as you cross psychological thresholds. The step from 99,000 to 101,000 miles costs more than the step from 89,000 to 91,000, because a great many buyers filter at 100,000.

Do not stack adjustment on adjustment. If you have already matched to a comparable mileage band, adjusting again for mileage double-counts.

Sanity-check against a ceiling

Whatever your market read says, cap it against a trade guide's retail figure for the car. If your calculation lands materially above that, something has gone wrong — usually a thin sample or a mismatched trim — and you should trust the ceiling rather than the calculation.

This matters most on unusual cars, where you might find only two or three comparable sales and one of them was an outlier. A small sample producing a confident number is how a dealer talks themselves into paying too much at auction.

If you have fewer than about four genuinely comparable sales, treat the resulting price as indicative and lean conservative. A thin sample is not the same as no data, but it is much closer to it than the tidy number on the screen suggests.

Price for the window you can afford

Pricing is a trade between money and time, and the right answer depends on your own position.

A car costs you money every day it sits: capital tied up that could be buying something else, floorplan interest if you use it, insurance, space, and the depreciation that carries on regardless. As a rough discipline, work out what a day of holding actually costs you and use it to judge whether an extra fortnight is worth an extra £300.

If your cash is tight or your forecourt is full, price keener and turn faster. If you have room and the car is genuinely rare, patience is a legitimate strategy. What is not legitimate is holding at an optimistic price by default and calling it patience.

Repricing: when and by how much

Set a review cadence and stick to it — a look at every car at 14, 30 and 45 days is a reasonable rhythm.

At each review, ask the diagnostic questions rather than reaching for the price. If the answer is genuinely price, move decisively. Small repeated cuts of £100 are the worst pattern available: they signal to watching buyers that more cuts are coming, so everybody waits.

One meaningful cut that lands the car just above the sold median will do more than four token ones, and it costs less in total because the car goes sooner.

Remember the margin scheme relief while you are deciding: on a margin car, a £600 cut costs you £500 of retained profit, because a sixth of the margin was always going to HMRC. Cuts are cheaper than they feel, and holding is dearer.

A working sequence

1. Pull sold prices for the exact specification, within a sensible mileage band, over the last 90 days.
2. Discard anything that is not genuinely comparable, especially on trim and gearbox.
3. Find the fortieth percentile rather than the median.
4. Adjust for mileage against age-expected, once.
5. Cap against the trade guide retail figure.
6. Sense-check the days' supply: in a tight market you can sit slightly higher; in a glut you cannot.
7. Set the price, then diarise the 14, 30 and 45 day reviews before you forget.

Do that consistently and the effect is not dramatic on any single car. Across a year it is the difference between turning your stock eight times and turning it five.