Why classic A/B testing doesn't translate
A digital A/B test shows two versions to two random, equivalent audience slices at the same time and compares outcomes. A single physical board can only show one design at a time to whoever happens to be driving past — there's no way to randomize who sees version A versus version B on the same face simultaneously.
What a realistic version looks like
Sequential testing on one board — running creative A for two weeks, then creative B for two weeks, each with its own unique tracked phone number or code — is the most common realistic approach, with the caveat that time itself becomes a variable (seasonality, day-of-week mix) that a true simultaneous test wouldn't have.
Where budget allows, testing across two comparable locations (similar traffic count, similar demographic, similar road type) with different creative running at the same time controls for the seasonality problem, at the cost of introducing location-quality as a new variable to account for.
What's worth testing
Headline wording and the single dominant visual are usually worth testing first — they carry the most weight in a glance-length medium. Small variations (a slightly different color accent, a minor layout tweak) are rarely worth the cost of a real-world outdoor test; save that rigor for the decisions that actually move the needle, like message and hierarchy.
Key takeaways
- True simultaneous A/B testing isn't possible on a single physical board — there's no way to randomize who sees which version.
- Sequential testing (creative A then B, each with its own tracked number) is the realistic default, with seasonality as the trade-off.
- Save real-world testing rigor for headline and dominant-visual decisions — the ones that actually move outcomes in a glance-length medium.