Horizontal scaling is growing a profitable Meta Ads campaign by adding new ad sets, audiences, creative angles, placements or regions, rather than raising the budget on what is already running. Raising the budget on the existing ad set is the opposite approach, known as vertical scaling.
How horizontal scaling works
You take something that works and extend it sideways. Typical moves include:
- launching the winning ads to a new audience, such as a lookalike of your best customers or a broad audience with no interest targeting;
- opening a new region, for instance a London trades business extending into Surrey and Kent once it has the capacity;
- building new creative concepts that speak to a different customer motivation, such as price, speed or reassurance;
- running a separate campaign with a different objective, such as retargeting people who visited the site.
Each new ad set starts its own learning phase and needs enough budget to gather results. For a long time, the most common version was duplication: copying a winning ad set several times with the same audience. That is far less useful now. When identical ad sets target the same people, they enter the same auctions and Meta has to choose between them, a problem called auction overlap.
At the time of writing (October 2026), Meta’s own guidance favours fewer, larger ad sets with broad targeting and a wide range of creative. In practice, horizontal scaling today is mostly about adding genuinely different creative and new markets, not about multiplying audiences.
Why it matters
Vertical scaling runs out of road. Push the budget far enough on one ad set and the cost per result rises as Meta reaches less likely buyers, and frequency climbs as the same people see the ads again. Horizontal scaling looks for new pockets of demand instead of squeezing the old one harder, which is often how a business grows Meta Ads spend without the cost per result running away.
The danger is spreading a modest budget too thinly. Five ad sets on £10 a day each will often perform worse than one ad set on £50, because none of them gathers enough data to settle. Many UK small business accounts would do better with less structure, not more.
Common mistakes
- Duplicating the same ad set with the same audience and calling it scaling.
- Creating so many ad sets that each one sits in learning limited.
- Expanding into regions or audiences the business cannot serve well, such as postcodes outside your delivery or call-out area.
- Forgetting to exclude existing customers from prospecting audiences.
- Judging a new ad set in its first two or three days, while it is still learning.
How to act on it
Scale only from a stable base: an ad set that has delivered results at an acceptable cost for a couple of weeks. Then decide what is new about the expansion. If the answer is “nothing, it is a copy”, raise the budget on the original instead.
Give each new ad set enough budget to reach roughly 50 optimisation events in a week, which is the level Meta’s guidance associates with leaving the learning phase. If that is not affordable, add the new creative into the existing ad set rather than splitting. Check overlap between audiences before you launch and keep exclusions tidy.
Deciding between vertical and horizontal moves, and when to consolidate instead, is a regular judgement call in the Facebook ads management I provide.