How a legacy structure works
A legacy structure is the account build that made sense when the platforms gave you more manual control and less automation. It splits spend into many small ad sets: one per interest, one per lookalike band, one per age or gender group, one per placement, often duplicated again for each creative. The logic was that if you drew the boxes tightly enough, the money would only reach the right people.
Every one of those ad sets bids separately and learns separately. That was tolerable when audience targeting was the main lever available. It became a problem as the systems moved to optimising at the person and creative level, because a narrow ad set gives them a small pool to work with and a trickle of conversions to learn from.
Why a legacy structure matters
It matters because a lot of accounts still run one, usually because nobody wanted to touch a build that once worked. The symptoms are recognisable: many ad sets each spending very little, several stuck in a limited-delivery state, results that swing from day to day, and rising costs to reach the same people because the audiences overlap and bid against each other.
The wider claim — that this pattern now generally underperforms a consolidated build — is the direction most practitioners and platform guidance have moved in, but it is a judgement drawn from experience rather than a settled fact. It is worth testing in your own account rather than accepting on authority.
Common mistakes with a legacy structure
The first is diagnosing the wrong cause. When a fragmented account underperforms, the usual response is more splitting — another lookalike layer, another interest to exclude — which makes the underlying problem worse.
The second is ignoring overlap. Several narrow ad sets built from the same customer base often describe largely the same people, so the account competes with itself in the auction and pays more for the impressions it would have won anyway.
The third is ripping everything out at once. A full rebuild resets learning across the account and leaves you with a fortnight of unreadable data and no way to tell whether the new structure is better or the old one was fine.
How to act on it
Start by measuring rather than rebuilding. List the ad sets that are spending too little to leave the learning phase, check which audiences genuinely overlap, and identify the handful of ad sets carrying most of the spend and most of the results.
Then unwind in stages. Merge the clearly duplicated ad sets into one broader ad set, keep the splits that exist for a real reason such as a separate market or a separate product line, and hold the creative constant during the change so you can attribute any difference to the structure. Move your effort from settings to creative production as you go, because that becomes the lever that matters. Where you inherit an account and cannot tell what was deliberate and what accumulated, a structured Meta ads audit is usually cheaper than guessing.
