A valuation is only an opinion until you can show what it’s built on. For used equipment, that evidence is comparables — real machines that recently sold, at known prices, in conditions you can reason about. Get the comparables right and the number defends itself. Get them wrong and you’ve dressed up a guess.
Realised prices, not asking prices
The first rule: an asking price is a wish, a realised price is a fact. Dealer listings tell you what someone hoped to get; auction results tell you what a machine actually fetched on the day. ExactVal anchors to realised sales — the hammer price someone genuinely paid — and treats listings only as a weak, secondary signal.
Where the data lives
Used-plant sales are spread across physical and online auction houses, manufacturer-approved remarketing, ex-rental disposals and cross-border trade between the UK and Europe. Each source has its own format, its own quirks, and its own view of condition. None of them, alone, sees enough of the market to be representative — which is exactly why a single dealer’s gut feel is so unreliable.
The honest version of a comparable carries its condition, its date, and its source. Strip those away and you’re comparing a hammer price to a hope.
Why it’s expensive
Good comparable data costs money for boring, structural reasons: it has to be collected from many sources, cleaned of duplicates and withdrawn lots, normalised to a common spec, and corrected for condition and currency. The sources that do this well guard it, because it’s the asset. That collection-and-cleaning cost is most of why a desktop valuation has historically run to £250.
Using comparables well
Raw comparables aren’t a valuation — they’re the raw material. The work is in the adjustments: aligning a comparable’s hours and condition to the subject machine, weighting recent sales over stale ones, discounting a thin sample, and being honest about uncertainty when the market is quiet. That last part is the confidence score: when the comparable set is deep and consistent, trust the number; when it’s sparse, the model says so instead of pretending.
Every ExactVal figure ships with the comparables behind it — the sales, their dates, and the adjustments applied — so the number arrives in the credit file as an argument, not an assertion.
Want to see the comparables behind a valuation on your own machines? Book a demo.