At a dealer group with sixteen branches, two buyers arrived at figures for the same car that were five thousand euros apart. Both experienced. Both good.
The managing director told it as an anecdote. It is not an anecdote, it is the heart of the problem.
Where that difference comes from
Appraising a car is ten minutes of work before the appraising even starts. Pull the RDW data. Look up the market value. Read through the damage report. Find a comparable car and see what it went for.
Each of those four steps involves choices. Which source do you take for the market value. How heavily do you weigh that damage. Which comparable car is genuinely comparable. Two people who are both good make those choices slightly differently — and then you are five thousand euros apart.
On one car that is annoying. On five thousand cars a year it is no longer noise, but a structural hole in your margin that nobody knows the size of. Because to know that you would have to have every car appraised twice, and nobody does that.
What it is not
It is not that buyers should pay closer attention. That is the reflex — more training, better agreements, a session about it at some point. That works for as long as somebody remembers it.
Nor is it that you should take an average. Averaging two estimates gives you a more precise wrong answer.
It is that the appraisal is written down nowhere. The method sits in the buyer's head, and every head has its own version.
What does work
Put the method on paper instead of in a head. Which sources, in which order, with what weight. Which margin. How damage is settled, against which of your own price lists.
Once that is explicit, a tool can carry it out. Not because a machine is smarter than your buyer, but because a machine does it the same way every time.
At this group the tool now appraises a complete list in one run: RDW check, market value, likelihood of sale, recommended retail price and a purchase price with a margin the group sets itself. Damage reports are read out and settled automatically.
Half a day of appraisal work became about eight minutes for twenty cars.
What this is really about
Those eight minutes are nice. The consistency is worth more.
Because the buyer has not been replaced. He now starts where he used to get to only after ten minutes of preparation: at the judgement. Whether this car fits this region and this customer demand. That is what he is good at, and there is still no model for it.
If two good people arrive at different answers, it is not the people. It is the fact that nobody ever wrote down how the answer comes about.