The claim is only as clean as its first text message.

Agreement age predicts product type

PCP moved from a minority product to roughly nine in ten financed new cars in a little over a decade. That steep, one directional curve means the year of an agreement is itself a strong PCP signal. A blank product field on a credit file dated 2017 onwards is overwhelmingly likely to be PCP, and the model scores it accordingly.

The surge sits inside the scheme window

The fastest PCP growth, 2012 to 2016, falls entirely within the FCA redress window of 2007 to 2024 and straddles the 31 March 2014 break point between the two schemes. The years when dealers pushed PCP hardest are exactly the years the regulator has priced at the greater loss rate. Volume and liability peak together.

Fewer sites, denser clusters

The franchised network roughly halved while finance penetration nearly doubled. Origination did not spread, it concentrated. At today's network size the national average exceeds 2,600 eligible agreements per franchised site. Each surviving dealership is a dense, mappable claims cluster rather than a needle in a haystack.

The showroom desk made the decision

Only 38% of motor finance customers shopped around, against 94% for motor insurance. For most buyers the dealership desk chose the lender, the product and the rate in a single conversation. That point of sale capture is the behavioural mechanism behind the commission findings, and it is why dealership environment reconstruction matters.

One client, several agreements

Over 80% of current holders on a new vehicle had used motor finance before. Serial PCP renewal every three to four years means a single client frequently carries two, three or four eligible agreements across the window. Real client value routinely runs at a multiple of the £830 per agreement headline.

The regional skew is bankable

Registrations concentrate hard: the South East and North West alone account for over a third of the national mix, and modelled regional books range from £2.25bn down to £0.25bn. Origination effort that follows the spikes reaches the largest books first. The cluster spike map turns that skew into a targeting sequence.

Product type prediction

Credit files often list motor finance without saying if it was PCP or HP. The model uses location, purchase year, dealership density and historical sales trends to calculate the probability the finance was PCP, turning a blank field into a likelihood score.

Dealership environment reconstruction

The model rebuilds the point of sale: how many local dealers were actively selling finance, which areas were heavy on PCP promotions, and the competitive pressure at the time of purchase. This supports inference on what the client was likely told or not told.

Timeline gap filling

Where the file holds no agreement date, renewal date or settlement behaviour, the twenty year sales reconstruction maps the client's purchase window, the PCP surge pattern for that year, and the product dominant at the time, producing an estimated timeline.

Contextual risk overlay

Credit files carry no mis selling risk signal. The model adds area level exposure risk, mis selling pressure zones, PCP saturation in the postcode, and historical lender behaviours in the region, converting a bare credit entry into a contextualised risk profile.

Lender ambiguity resolution

Where a lender name gives no product signal, the model matches the lender's historic PCP usage, regional sales footprint and product mix in the client's purchase year, inferring the missing product type from lender behaviour.

Client scoring output

For each client the model generates:

• PCP probability
• exposure rating
• cluster intensity score
• local sales saturation score
• likely dealership influence