Last month’s best ad is getting more expensive. The obvious move is to refresh it. A few new openings, a different edit, another cut for Reels. With AI, you could have 50 versions ready by Friday.
Before you brief any of them, ask a more basic question. Which customers are we trying to win, and what would make them choose us?
If every new ad makes the same promise to the same kind of customer, you’ve added production volume. You may still be leaving the same growth opportunity untouched.
Start with the customers worth winning
Imagine you run an airport transfer service. A family planning a holiday wants to know what the whole journey will cost and how they’ll get from arrivals to the hotel with children and luggage. A traveller who’s already landed might want the next departure, the pickup point and a ticket they can buy on their phone.
Both need a transfer. They’re making different decisions.
For the family, an ad could show the full price for the group and what happens after landing. For the traveller in arrivals, it could show the pickup point and how to find the next available departure. Each idea makes a different promise and needs different evidence to back it up.
Which opportunity deserves more investment depends on what you can serve profitably. You might prioritise advance bookings on a route with spare seats or focus on same-day demand where departures are frequent and mobile checkout is quick. That choice gives the creative brief a commercial job to do.
Scale the ideas, then the executions
Once you know which customers you want to win, build a handful of ideas around distinct reasons to book. For the family, price certainty and an easy arrival are two separate hypotheses. For the traveller who’s already landed, speed and finding the pickup point may be more compelling. Each gives you a different story to tell.
From there, AI can help turn a seed set of ideas into dozens of executions. A price-certainty concept might become a creator walkthrough, a static fare breakdown and a short video showing the booking flow. They give the delivery system different ways to present the same proposition. A pickup-point walkthrough adds a proposition for another buying situation.
Meta’s Andromeda retrieval system is designed to handle a growing pool of creative and find relevant ad candidates for each person. Broad delivery can give the system room to find those matches without requiring a separate ad set for every segment. You still have to choose which demand to pursue, what to promise and what you can deliver. And the number of variants worth putting live depends on how much delivery and booking data you can learn from.
Work out what actually changed
A familiar ad getting weaker can be a sign of fatigue. It can also happen because the mix of customers has shifted, the offer is less competitive or the booking journey has become harder. If an airport transfer campaign moves from peak holiday planning into a period with more same-day demand, a new edit of the advance-booking ad may leave the real problem in place.
Start with where the decline showed up. Did people stop clicking, or did they still click and book less often? Did the decline affect several creative concepts or mainly one? Look at the relevant route, booking lead time, fare and availability alongside the ad data. Those checks help you decide whether to refresh an execution, introduce a new proposition or fix something after the click.
Read results at the concept level as well as the ad level. Did the group-price idea drive bookings at an acceptable cost? Did the pickup-point idea get enough delivery to assess? The platform may give one concept much more exposure than another. Treat that as evidence about how the campaign delivered. If a decision depends on whether one proposition actually caused more bookings, use a controlled test designed to answer that question.
Before you commission the next batch, write one sentence for each concept. “We want to win this customer, in this buying situation, by showing them this reason to choose us.” If all 50 ads trace back to the same sentence, you know exactly what you’ve scaled. The next brief can decide whether that idea needs more executions or whether another group of customers deserves an idea of its own.