How an Agency Rebuilds a Client's Audience Library After Inheriting the Account

The handoff document was two paragraphs and a spreadsheet password. Everything else about the inherited account — what audiences existed, which were current, who built them and why — had churned out the door with the previous agency. This is a story about an agency inherited ad account audience cleanup: how a performance shop took over a brand mid-flight, found dozens of duplicated and drifting custom audiences, and rebuilt the whole mess into one reusable library before it dared to relaunch a single campaign.

This is a composite story, but every beat is real for any agency that has taken over an account from a shop that lost the client. The names and exact numbers are illustrative; the inheritance problem is not.

The inheritance: a black box with spend in it

The brand had fired its previous agency after a quarter of flat results and hired this team to turn it around. The account itself transferred cleanly enough — partner access granted, billing moved — but the audience layer was a black box. Across two Meta accounts and a Google property, the new team counted more than forty custom audiences and lookalikes, most with names like "LAL 1%", "LAL 1% (2)", "Purchasers_final", and "Purchasers_final_v2". Several were live in active ad sets. Nobody on the new team knew which ones were real.

That ambiguity was not cosmetic. The agency was being paid to fix performance, and performance was running on targeting it could neither read nor trust. Relaunching on top of that — layering new campaigns over forty undocumented audiences — would mean inheriting every hidden overlap and stale seed the previous shop had left behind, and owning the results as if they were its own choices.

The cost of inherited disorder is well documented. Gartner estimated in 2021 that poor data quality costs organizations an average of $12.9 million a year, and Nielsen reported in 2023 that audience-targeting accuracy across digital ad platforms frequently falls well below half — meaning a stale, unverified seed is not a neutral starting point but an active drag on every dollar relaunched against it.

Step one: reconstruct the past with the action history

Before touching anything, the agency wanted to understand how the account got this way. The previous shop's people were gone, but the account's history was not. The team connected the account to its operating layer and opened the action history, which reads through the official platform APIs and surfaces a timestamped record of changes — including changes made before the agency took over, to the extent the platform retained them.

What they reconstructed was illuminating. The duplicate purchaser lookalikes were not redundant accidents; they had been built three months apart, off two different customer exports, and both left running. The "LAL 1% (2)" had been created the week a junior at the old shop covered for a colleague, off a seed that was already stale. Reading the history turned forty mystery audiences into a story the team could act on, the same investigation method we describe in how to investigate ad account changes: narrow the window, read the attributed entry, and decide.

Step two: inventory everything in one view

With the history understood, the team needed a single map of what currently existed. The trap most agencies fall into here is inventorying account by account — opening each native manager, listing audiences, and trying to mentally diff them. That is exactly how the mess was created, and it is how a cleanup quietly recreates it.

Instead the team listed every audience across both Meta accounts and the Google property in one screen, scoped to the accounts they had access to, using Wevion's Audience Hub. Seeing all forty-plus in one place made the duplication obvious in a way no per-account view ever had: the three purchaser lookalikes lined up side by side, the customer lists with near-identical names sat together, and the genuinely unique segments — a handful of high-value retargeting audiences worth keeping — stood out from the noise. This is the same scatter-to-library shift covered in how to fix scattered ad audiences across accounts, applied to an account someone else made scattered.

Step three: pick one authoritative seed per segment

Now the actual consolidation. For each real segment — the purchaser lookalike, the high-value customer list, the product-page retargeting pool — the team picked one authoritative version and rebuilt from a clean source rather than trusting any inherited seed.

The customer lists were the riskiest. A list named "Purchasers_final" might be a year old, hashed from a partial export, and matching poorly — but its name gave no hint. So the agency went back to the client, pulled a fresh authoritative purchaser export, and re-uploaded it once through the hub, reading the valid-versus-invalid match count so it knew exactly how much of the list actually resolved. Anything it could not verify against a known-good source was treated as something to rebuild, not adopt. The discipline mirrors our guide to building reusable audiences across accounts: the seed is the asset, and a seed you cannot verify is a liability you inherited.

Checking the overlap the previous shop never could

Once the clean seeds were in place, the team ran the overlap check that the inherited chaos had made impossible. On the larger Meta account, the rebuilt retargeting pool and the purchaser lookalike overlapped heavily — the previous shop's campaigns had been bidding against themselves, inflating CPMs that the brand had been told were "just a competitive auction." The agency consolidated the overlapping audiences and added exclusions before any relaunch.

That single check reframed the whole engagement. Instead of inheriting an unexplained CPM problem and owning it, the agency could show the client, at the very first review, a concrete inefficiency the previous shop had left running — and the fix already in place. It turned the inheritance from a liability into a quick, visible win, the kind of overlap discipline the agency one-library playbook treats as standard hygiene.

Step four: relaunch on the clean library — with a human approving every step

The relaunch was deliberately boring. With one verified library per segment, the buyers chose which rebuilt audiences to run, how to layer and exclude them, and what budgets to test — and approved each campaign before it went live. Nothing relaunched automatically. The hub had removed the cleanup tax and surfaced the overlap; it did not make a single targeting or launch decision, and the team would not have wanted it to.

A few practical realities shaped the rollout. The sync ran about every 15 minutes through official APIs, not live, so a freshly rebuilt audience needed a short delay before it populated and was safe to launch against. Platform depth differed — Meta carried the fullest set of build actions, with Google user lists covered for the core flows. And the old duplicates stayed in place, paused, until the rebuilt versions were verified live, so nothing was deleted on faith. For the platform layer that hosts the inventory, the rebuild, the relaunch and the record together, see our best ads management software for agencies roundup.

What the cleanup changed — and what it did not

A few weeks in, the differences were concrete. The forty-plus mystery audiences had become roughly a dozen verified, named segments the whole team could read at a glance. New campaigns launched off seeds the agency had built and could vouch for, not inherited fog. And when the client asked why early results looked different from the previous shop's reports, the agency had a real answer — the overlap it had removed, the stale seeds it had replaced — drawn from the action history and the consolidation work, not a theory.

What did not change is worth stating plainly. The buyers still made every targeting call. The hub listed, synced, built and compared audiences and stopped there; it did not optimize, launch, or act on its own. The sync was about every 15 minutes through official APIs, not live. And the cleanup did not magically improve the client's product or offer — it removed the operational debt the agency had inherited so the team's actual strategy had a clean surface to run on.

The same playbook generalizes beyond inheritance. Any agency reassigning a buyer onto an existing client, or a multi-account operator absorbing a brand mid-flight, hits the same wall: audiences someone else scattered, with no record of intent. The cleanup motion — reconstruct, inventory, verify, consolidate, relaunch with approval — is the same whether the mess came from a churned competitor or a departed colleague.

The takeaway for any agency inheriting an account

If you take over an inherited ad account, the dangerous move is to relaunch fast on top of audiences you did not build and cannot read. The forty mystery seeds are not someone else's problem the moment you launch over them — they are yours, overlap and stale lists and all. Reconstruct the past with the action history, inventory everything in one view, verify one authoritative seed per segment, check the overlap the previous shop never could, and relaunch only on the clean library with a human approving each step.

To see how a central audience library and an action history handle your next inherited account across Meta and Google — with per-client scoping, sync about every 15 minutes through official APIs, and every targeting and relaunch decision left to your buyers — start a 14-day Wevion trial alongside the permanent free plan and clean up your first inherited account this week. For the rules-and-automation comparison many agencies weigh at the same time, see how Wevion stacks up against Revealbot.

This case study is part of our agency tools hub — explore the full cluster for related playbooks.