FB Motors. One system, from ad to resale.

The problem

By early 2026, the creatives had worn out: the same ads were bringing in fewer and fewer sellers, and the cost per vehicle bought was soaring.

What we built

All acquisition runs through Meta (Facebook and Instagram). The ads speak to owners of a vehicle that the mainstream market turns down or undervalues. We keep what tests new ideas separate from what runs the proven ones, and every creative is replaced as soon as its results wear off: detection is automatic. The multi-step form collects exactly what is needed to value the vehicle before the first call.

What changed

The ads now attract only sellers FB Motors can actually buy from, and a creative that is wearing out gets caught before it becomes expensive.

The problem

Most ad accounts send Meta a single signal: “a form was filled in”. So Meta looks for people who fill in forms, not sellers you actually do business with.

What we built

At FB Motors, every stage of the journey is sent back to Meta, and the last two carry a value in euros: the estimated margin at purchase, then the actual margin at resale. Every event is sent in real time, server-side, and attributed to the original ad.

What changed

We know which creative produced which margin, not just which forms. Rejected enquiries also feed an exclusion audience, so we stop paying for the same profiles.

The problem

The vast majority of enquiries will never turn into a deal, and every pointless call eats into the team's time.

What we built

We first built a score out of 100, across five tiers. At FB Motors' real volumes it would not calibrate, so we replaced it with clear-cut elimination rules that automatically rule out off-target vehicles (mileage, engine, segment, asking price). What remains goes to the team, with a price indicator.

What changed

The team now only sees the enquiries that can lead somewhere.

The problem

Offer a few hundred euros too much and the margin is gone. Offer too little and the seller goes elsewhere.

What we built

Each enquiry is compared with listings on the Belgian market, from the most precise to the broadest: model, engine, gearbox, year, then adjusted for mileage. The system applies the depreciation learned from FB Motors' actual resales, and shows whether the seller's asking price is a good deal, fair or too high. When an enquiry is too high, a counter-offer is calculated. On every resale, the gap between the estimate and the price achieved is recorded, and the next estimate takes it into account. The final decision stays human.

What changed

Every offer starts from a calculation, and every resale makes the next one more accurate.

The problem

Sellers sell to whoever calls back first. Before, too many enquiries went unanswered.

What we built

Every shortlisted enquiry lands on Telegram, with the vehicle, the asking price and the estimate. One click accepts or rejects it. Once accepted, it becomes an opportunity, automatically assigned to the least busy setter.

They call every seller back, check the vehicle's condition, confirm the price and book the inspection. They move every file forward through the pipeline and never let go of a seller who does not answer: call-backs are scheduled at 9 am and 6 pm, and the system follows them up automatically. The final price and the negotiation stay in FB Motors' hands.

What changed

Every file has an owner, a next step and a scheduled call-back. At the Agreed stage, the purchase and the stock entry are created automatically, and the purchase signal goes to Meta.