Fifth Third Bank.

Everyone was bidding on the same 400 keywords

Fifth Third’s acquisition cost for a new checking household had been climbing for six straight quarters. The reason was not creative and it was not offer — it was inventory. Every regional bank in the footprint was bidding against the same finite pool of high-intent search terms, and the auction only goes one direction.

The brief was blunt: find reach that competitors cannot bid against, in the specific ZIP codes where branch density made a new account profitable, and prove it moved something other than a brand-tracker score.

We planned it like a digital buy, then drove it

We started from the branch list, not the map. Each location contributed a catchment of target ZIPs, weighted by existing household penetration — so the fleet spent its hours where an incremental account was most likely and most valuable, rather than simply where traffic was heaviest.

Routes were then built around daypart behaviour: commuter corridors at morning and evening peak, retail and grocery clusters midday on weekends, and dwell time deliberately concentrated at intersections with long signal cycles. A truck stopped at a red light for ninety seconds is a very different media unit than one at speed.

Creative was art-directed for glance value at distance — one message per side, a single proof point, and the branch offer legible from three car lengths back.

Every mile logged, matched and attributed

Each vehicle streamed GPS at fifteen-second intervals, which gave us a verified impression count rather than a modeled one — audited by a third party and reconciled against the plan weekly.

To connect exposure to outcome, exposed ZIPs were matched against demographically similar control ZIPs with no fleet presence. Branded search volume, site sessions and — critically — actual account openings were compared across the two groups for the length of the campaign.

That control design is what let us report a real incremental number instead of a correlation. The lift held after controlling for concurrent digital spend and seasonal patterns.

Break-even inside a year, and still compounding

Branded checking search clicks nearly doubled in exposed markets against control. Household production lifted eight percent, delivering 6,802 incremental checking households over the campaign.

Because a checking household carries a multi-year value, the campaign passed better-than-break-even ROMI in under twelve months and continued returning after the fleet stood down. The programme has since expanded into three additional metros.