Google Ads audit: method, price and deliverables

A Google Ads audit can be free, cost €250 on a freelance marketplace, or take several days of agency work. All three go by the same name. What separates them is mostly the time spent checking that the account's conversions are real, and what you actually hold in your hands at the end.

Why prices range from zero to several days of work

We found no reliable public price grid for Google Ads audits in France. Visible offers range from the free audit agencies use as a pre-sales tool to turnkey offers sold on freelance marketplaces: on Comeup, an audit with a PDF report and an action plan is listed at €250, with an €85 launch price reserved for the first five orders (prices checked on 13 October 2026). Beyond that, agencies bill a flat fee or by time spent, and rarely publish their prices.

The gap is explained first by the time devoted to measurement. Reading an account's structure, campaigns and bid strategies is quick. Establishing that the conversions on screen match real sales or real leads means comparing the account with the back office or the CRM, checking tags, GA4 imports and counting settings. In our view, this is what turns half a day of work on an account with a few campaigns and clean tracking into several days on a multi-country account with Performance Max, a product feed and imported conversions.

The free audit is paid for too: the agency doing it hopes to take over the account. That is not a flaw in itself, but its conclusions seldom recommend changing nothing. Read it as a well-argued sales proposal.

Finally, Google offers something it does not call an audit: the optimization score. Its documentation presents it as an estimate of how well the account is set to perform, from 0 to 100%, calculated in part from the impact of available recommendations. It also states that applying or dismissing a recommendation changes the score. An account can therefore gain points without creating a single extra euro of margin, and we never treat it as an audit finding.

To decide how much to spend on an audit, compare its cost with the amount at stake. Take a hypothetical account spending €20,000 a month on Search. If 8% of that spend goes to search terms that have not converted once in 90 days, the stake is €1,600 a month, more than €19,000 a year, and a few days of analysis are easy to justify. On an account spending €2,000 a month, the same work would cost more than it could save, and a tool-based diagnosis is enough.

The method starts with conversions

A serious auditor opens the conversion settings first, before structure, keywords or ads. Smart Bidding sets bids based on the conversions recorded in the account. If those are wrong, bids are optimized on false results, and every recommendation drawn from the rest of the audit rests on wrong numbers.

The first check compares Google Ads conversions with the orders or leads recorded over the same period. There is always a gap, since Google credits itself with conversions according to its conversion window and attribution model. The warning sign is a ratio close to two to one across every campaign at once: the same purchase is often counted twice, by the Google Ads tag and by a GA4 import, both set as primary actions.

Next comes the counting setting. Google offers "Every" and "One", and recommends the latter when the advertiser cares about leads rather than the number of sales. A contact form left on "Every" counts three times the person who submits it three times, and the bid strategy rewards that behaviour.

The scope of primary actions is checked at the same time. According to Google's documentation, only primary conversion actions feed the "Conversions" column and bidding, while secondary actions are for observation only. A click on the phone number or a contact page view set as primary blends events of incomparable value into the optimization. A fixed conversion value, identical for every sale, also distorts optimization: the interface still shows a ROAS, but a target ROAS strategy no longer tells a €30 sale from a €300 sale and, in practice, behaves like a target CPA.

If this step fails, we recommend stopping the audit there, fixing measurement and resuming afterwards. Analysing budget allocation on duplicated conversions produces recommendations that are wrong and yet convincing.

Where the recoverable money is

Once measurement is validated, the audit looks for spend that does not come back. The search terms report is the most direct place: it shows what people actually typed, whereas keywords only say what the advertiser bought. Pulling 90 days of search terms, sorting them by cost and isolating those that spent without converting produces a list of keywords to exclude and a verifiable amount, which remains a floor: Google leaves out of the report terms that do not generate enough searches, to comply with its data privacy standards. The reverse must be checked too, namely that no old negative keyword now blocks profitable queries.

Budget allocation is read by comparing, campaign by campaign, the share of spend with the share of conversion value. Two checks complete the picture on most accounts: cost per acquisition recalculated separately for brand and non-brand queries, and the share of brand queries captured by Performance Max. A satisfactory account average can hide loss-making non-brand acquisition.

Bid targets set far from the account's history also deserve a look. A target CPA well below the observed CPA limits delivery, and the campaign then spends a fraction of its budget without the interface clearly flagging it.

Quality Score, lastly, is a diagnostic. Google states that it is not used in the ad auction and breaks it down into expected clickthrough rate, ad relevance and landing page experience. Its value lies in pointing to the component at fault: ad relevance is fixed in the ads, landing page experience is fixed on the website.

The deliverables to insist on

The final document says a lot about the quality of the work. Before signing, ask for a sample deliverable and check that it includes:

• the state of measurement, with the quantified gap between Google Ads and your real sales, and its explanation;
• dated findings, each tied to the data and the period behind it;
• the amounts at stake, with how they were calculated;
• an action plan ranked by estimated impact and effort;
• items you can use directly (terms to exclude, settings to change), not just their description;
• a clear line between what was observed and what remains a hypothesis.

Three signs should make you doubt: an overall score without the checks behind it, screenshots of Google's recommendations presented as findings, and a guaranteed CPA reduction before the auditor has even opened the account.

An account audit also leaves two questions unanswered. It records the conversions Google Ads credits itself with, without measuring what would have happened without the spend. A brand campaign with a high ROAS may mostly reach customers who would have come anyway, and only a test with a control group (a geo test, an incrementality study) settles that. Nor does it know your margin: provide it from the start, otherwise any verdict on a CPA being "too high" has nothing to stand on.

What CheckUp delivers

CheckUp is the audit tool BRAIKE develops and operates. It connects to Google Ads, Meta and TikTok accounts through OAuth, read-only, and runs 195 checks per account, grouped into six pillars: measurement, audiences and signals, structure and hygiene, steering and performance, creative and brand safety, plus checks specific to the advertiser's industry.

The deliverable includes a health score from 0 to 100 per platform and per pillar, the verdict for each check, a euro estimate of the potential gain from each recommendation to help prioritise, the likely cause of each issue with the fixes to plan, and a shareable report.

On price, CheckUp is not yet openly available. During the private beta, BRAIKE runs these health checks free of charge and delivers them in about 24 hours, depending on the team's availability. A user licence has been announced for the end of 2026.

A tool does not replace a consultant's judgement, though. CheckUp knows neither your margin nor what happens outside the account (a stock-out, a price increase, a site redesign). We use it to rerun the checks regularly and keep the human audit for structural decisions: taking over an account, changing bid strategy, rebuilding measurement. We explain why we prefer frequent health checks to the annual audit in the end of one-off audits.

Before you order an audit

Prepare three things, whoever the provider: an export of your orders or qualified leads over the last 90 days, your margins by product family, and read-only access to Google Ads that you will remove at the end. Then ask each candidate what they check first. We would pick the one who starts with conversion measurement, even if they are not the cheapest. For a first, no-commitment assessment, you can request a CheckUp health check during the beta.

Frequently asked questions

Do you need to grant admin access for an audit?

No. Google Ads offers a read-only access level, enough to audit an account without being able to change it. CheckUp also connects read-only, and access can be revoked at any time.

Does a Google Ads audit cover GA4?

Only partly. The audit checks the link between Google Ads and GA4, the imported conversions and consent mode on Google tags, which in Europe determines the share of conversions that are observed and the share modelled by Google. The full GA4 setup (events, parameters, reports) calls for a dedicated GA4 audit. Have that scope spelled out before signing.

How often should an audit be repeated?

A structural audit makes sense when a new provider comes in, then after any major change (site redesign, new bid strategy, a new country). In between, regular checks catch breakages, such as a deleted tag or a campaign that has become limited by budget, far sooner than an annual audit.

Sources checked on 13 October 2026.