Ad account audit and rescue

What this actually solves

The buyer of this service is stuck, not curious. Money leaves the account every month, the dashboard is full of figures that move, and none of it resolves into an answer about whether it is working. Either an agency is running it and has stopped explaining themselves, or the owner is running it and has run out of ideas. What is wanted is a second opinion from someone who is not trying to sell the next twelve months of management.

There is a specific reason this is hard to get. Nearly everyone who will look at your account for free is looking at it in order to win it, which shapes what they find — the findings arrive large, urgent and conveniently solvable by hiring them. An audit you paid for behaves differently. It can conclude that the account is broadly fine and the real constraint is your margin, or your landing page, or that the ceiling is the offer rather than the media buying. Those are useful answers and nobody gives them away.

The order of the work matters more than the thoroughness. Conversion tracking gets checked before anything else, because if the Purchase or Lead event has not been arriving, every campaign comparison underneath it is a ranking of fiction. Auditors who skip that step produce impressive documents built on numbers that were never real.

What we do, and in what order

1. Take view-only access and set the scope. Platform access through proper user management, plus read access to Analytics and to the store or CRM where the truth about revenue lives. Written scope — which platforms, which date range, and what is explicitly not included. Half a day.
2. Check the tracking before trusting a single number. Which conversion actions exist, which are set as primary, whether they fire on something real, whether the same sale is being counted twice, and whether the platform's totals reconcile with the store's. If the answer is bad, the audit says so first and everything after it is framed accordingly. One day.
3. Pull ninety days of raw data. Campaign, ad group, search term, placement, device, location, day and hour, plus the change history — which tells you what was altered, by whom and when, and frequently explains a performance cliff that has been blamed on the market for six months. One day.
4. Find where the money leaks. The recurring list is short and it is remarkably consistent — search terms that have spent real money and returned nothing, the Display Network or search partners quietly left enabled inside a search campaign, ads pointing at pages that now redirect or 404, location settings that include people merely interested in your area rather than in it, budgets sitting on campaigns that were paused months ago, and a conversion action that counts something the business does not care about. Each one is quantified from the exports, not estimated. One to two days.
5. Assess structure, budgets and creative. Whether the campaign split makes commercial sense, whether budget sits where the margin is, whether ad groups have anything being tested against anything, and whether the ad copy matches what the landing page then says. One day.
6. Write the report. Five to eight pages, plain English, with a technical term explained in a line wherever one is unavoidable. Every finding in three parts — what we found, what it means, what changes if you fix it — ordered by the money involved. An executive summary at the front with the identified waste totalled, and a first thirty days plan of five actions with the time each one takes. One to two days.
7. Walk you through it live. Because the questions that matter are asked out loud, and because a report handed over silently tends to get filed. One hour.

What moves the price

The range is $250–$2,000, and it moves with the size of the surface being examined.

How many platforms and how much spend. A single platform on modest monthly spend is a contained job. Two or three platforms, several markets and a large historical account is a different volume of data and a different amount of reconciliation.

Whether the numbers can be trusted as found. If tracking is sound, we analyse. If it is not, part of the work becomes establishing what is actually true before anything can be ranked, and that takes longer than reading a clean account.

Audit only, or audit and rescue. The rebuild is scoped and quoted separately once the audit is written, so that the diagnosis is never inflated by the size of the treatment attached to it.

What usually goes wrong

Tracking gets checked last, or not at all. The audit ranks campaigns, names winners and losers, and recommends moving budget — all from conversion data that was double-counting, or counting a page refresh, or not counting at all. Every conclusion inherits the fault. This is why it is the first thing we look at and why the report leads with it.

The report turns into a list of forty complaints. It reads as thorough and it produces nothing, because the owner opens it, feels buried, and closes it. A useful audit names five to seven things, in order, each with the difference fixing it makes. The discipline is in what gets left out, and the items we deliberately did not raise are listed at the back with the reason.

Findings are given numbers they have not earned. It is very tempting to attach a confident figure to every problem, because large numbers make a report feel valuable. Where the exports support a figure, it is stated and sourced. Where they do not, the report says "needs measurement" and explains what would have to be tracked to know. A fabricated estimate is worse than no estimate, because decisions get made on it.