From Pump to Counter
A fuel and convenience retailer treated the few minutes at the pump as a moment of intent and served ten-minute offers to drivers while they were fueling. Verified store entries rose 48%, holdout-adjusted, and second visits within 30 days roughly tripled.
Revised edition of a 2024 engagement. Client name withheld under NDA.
The numbers
- +208%
- second store visits within 30 days (6.2% to 19.1%)Exposed drivers who made a first visit, against control
- 11.7 → 6.9
- median days between first and second visitA 41% shorter gap among exposed drivers
- +21%
- food-program purchases attached to fuel tripsAgainst the pre-period; modeled from time-aligned register data where no loyalty link existed
- 3.2×
- media return on net food-program marginMedian across the four markets
- $0.78
- cost per incremental store visitMedian across markets; $0.61 in the best-performing decile
Store entries per 1,000 fueling devices
Door crossings with at least 60 seconds inside; eight-week flight across four markets.
Drivers who came back within 30 days
- Before
- After
Before is the control group; after is exposed drivers who made a first visit.
Show the numbers as a table
| Baseline | 142 |
|---|---|
| With pump-side offers, holdout-adjusted | 210 |
| Second visit within 30 days | 6.2% before, 19.1% after |
The challenge
The forecourts were busy and the stores beside them were not. Pump transactions carried thin margins, while the hot-beverage and quick-serve food program sat under-used through the morning and lunch commutes, exactly when it should have been busiest. Conventional geofencing treated everyone near a station as one broad audience, and the retailer wanted proof of incremental store visits, not impressions.
The insight
A driver at the pump has already stopped, has a few idle minutes, and is standing a short walk from the counter. That is genuine intent, but it lasts only as long as the fill. The offer has to arrive while the driver is fueling, be worth the walk, and expire before it becomes noise. And a single visit is only a trial. The value lies in the second and third, when a stop becomes a routine.
What we did
- 01
Find the driver who is fueling now
Pump islands and store entrances were mapped with fine-grained geofences, and dwell patterns separated drivers who were fueling from traffic passing through.
- 02
Fit the offer to the hour
A $1 morning coffee and a $5 lunch offer were served by daypart, weather and walking distance, as one-tap wallet passes that expired in ten minutes.
- 03
Design for the second visit
Drivers who fueled again without coming inside received a reminder on day two and day five, with value that stepped up for repeat visitors.
- 04
Hold back to prove it
Ten to fifteen percent of pump clusters in each market received nothing, and synthetic controls adjusted for fuel prices and weather.
How it was measured
Store entries were counted as door crossings with at least 60 seconds inside, per 1,000 devices detected fueling, over an eight-week flight in four markets. Ten to fifteen percent of pump clusters in each market were held out, and synthetic controls adjusted for fuel-price swings and weather. The holdouts showed no material spillover within each market. Food-program attachment was modeled from time-aligned register data where no loyalty link existed.
Results
Store entries rose 48%, holdout-adjusted, from 142 to 210 per 1,000 fueling devices. Among exposed drivers who made a first visit, 19.1% returned within 30 days against 6.2% in control, and the median gap between visits fell from 11.7 to 6.9 days. Food-program attachment on fuel trips rose 21%, media returned 3.2× on net food-program margin in the median market, and an incremental visit cost $0.78, or $0.61 in the best-performing decile.
Results describe a specific engagement and are not a guarantee of future results. Platform-reported figures are labeled as such.