Rebuilding a search account from live data, not guesswork

Search accounts accumulate sediment. Keywords added for a campaign that ended two years ago, ad groups nobody has opened since the last agency, conversion actions that may or may not still fire. When performance slips, the temptation is to rebuild from intuition — start fresh, write some sharp new ads, and trust that experience will fill the gaps.

Intuition is exactly how the waste got there in the first place. A rebuild grounded in live data — what the account is actually doing, what competitors actually run, and what searchers actually type — produces a structure that earns its budget. The order of operations is the whole game.

First, audit the account you cannot see

Before a single new ad is written, the existing account has to be made legible. Most underperforming accounts are partly dark: spend flowing through campaigns nobody is monitoring, conversion tracking that may be silently broken, query reports full of irrelevant terms quietly draining budget. Rebuilding on top of that without mapping it first means inheriting its problems with a fresh coat of paint.

The audit answers the unglamorous questions. Where is the money going? Which conversions are real? What is the account paying for that it should never have bid on? Establishing that baseline on the live account is what turns a rebuild from a guess into a controlled change — you cannot prove an improvement against a baseline you never measured.

There is a quieter benefit too. The audit surfaces what is already working, and a rebuild that throws away the winners along with the waste is its own kind of failure. Knowing which campaigns, terms, and assets to keep is as much a part of the exercise as knowing what to cut.

Then, write the ads from the query, not from the brief

Once the account is legible, the rebuild can begin where it should: with what people actually search and how the competition actually answers. Responsive search ads written from a creative brief tend to describe the product the way the company thinks about it. Ads written from live query data describe it the way the searcher is looking for it — and the gap between those two is where cost per acquisition lives.

The discipline is to ground every headline in evidence: the queries that convert, the asset slots that are pulling weight, the ad strength signals that the platform is already reporting. A rebuilt RSA set anchored to live search behaviour is not a creative exercise — it is a structural one, and it is measurable against the baseline the audit established.

And position against who is actually in the auction

Ads do not run in isolation. The same query that the account is rebuilding for is being contested by competitors with their own offers, their own copy, and their own claims. New ads written without reading the competitive set risk saying the same thing as everyone else, louder — which is just a more expensive way to compete on price.

Drafting counter-competitor copy means reading what rivals run on the shared terms and writing deliberately against it: leaning into a differentiator they cannot match, answering an objection they ignore, occupying the angle the auction leaves open. This is positioning expressed as ad copy, derived from live competitor data rather than assumed from market lore.

The sequence is the safeguard

Audit, then write from the query, then position against the competition. Each step depends on the one before it: you cannot prove a rebuild worked without a baseline, you cannot write to intent without reading the queries, and you cannot differentiate without seeing what you are differentiating from. Done in that order, a rebuild is a series of evidenced, reversible changes rather than one large bet on instinct.

The honest limit: live data narrows the guesswork, it does not eliminate it. Auctions shift, competitors respond, and queries drift. What the data buys is a rebuild that starts from reality and can be re-checked against it — which is a far better starting position than a confident blank page.