Google stops reading your advert and starts reading your paperwork

What the new Google Ads policy actually says#

A fox does not argue with a fence. It walks the whole wire, watching where the hedgerow thins, and picks its path in daylight rather than meeting the barbs at midnight. That patient prowl is precisely the posture advertisers need this autumn, because the Google Ads policy covering government documents and administrative services is about to stop reading your advert and start reading your paperwork.

Here is the change, stated plainly. From 5 October 2026, advertisers promoting government-issued documents and administrative services must prove explicit government authorisation for the specific service being advertised, according to Search Engine Land, which reported the tightening on 5 August. Acceptable proof includes a domain linked from an official government website and clearly identified as an authorised provider, a listing in a government-managed directory of approved partners, or a named authorisation on an official regulatory portal.

The refusals are as instructive as the requirements. Business licences, commercial contracts, company registrations and government-hosted blog posts are all explicitly ineligible as evidence, per the same Search Engine Land report. Targeting must also be trimmed to the relevant geographic regions, with a sensible exception carved out for inherently cross-border services such as electronic travel authorisations and border entry documents.

What Google's own page says, and what it does not#

A trade report is news of record, not a primary source, so we went and read the policy properly. Google's live Advertising Policies Help page for government documents and services states the standing rules in flat, unglamorous language: only certified governments and authorised providers may run ads that promote direct acquisition of specific government documents and services. It defines an authorised provider as an advertiser given explicit permission by a government to provide a specific document or service, and it demands that the provider's domain be linked from an official government website and named there as permitted for that exact service.

It does not, as of this morning, publish the 5 October date. That gap matters and we will not paper over it. The deadline in this piece comes from Search Engine Land, dated 5 August 2026. The rules quoted come from Google. Where the two ever differ, Google's page wins, and a dated screenshot of both belongs in the same folder as your evidence.

Two structural details are easy to skim past. The policy sits inside Other restricted businesses, a bucket of twelve categories with government documents listed first. And certification here is not one step but two, a point Google's restricted products application page makes plainly: in addition to completing advertiser verification, certain industries require you to apply to advertise that category at all. Two clocks, two queues, one deadline.

Sixty days sounds generous until you notice that most of that clock belongs to somebody else. The quarry here is a signature sitting on a government website, and no bid strategy has ever hurried a civil servant. That single fact is what makes this Google Ads policy different from every other policy note in your inbox this month.

Evidence beats enthusiasm: the proof hierarchy nobody reads#

The most expensive mistake under this Google Ads policy will not be a badly written advert. The snare is subtler and slower: a capable growth team confidently submitting the wrong document, waiting three weeks for a rejection, then discovering that the right document takes six.

There is a hierarchy here, and it is worth spelling out slowly, because the four rejected evidence types are precisely the four things a marketing team already has to hand. A company registration proves you exist. A business licence proves you may trade. A commercial contract proves somebody in the public sector once bought something from you. A government blog post proves a communications officer typed your name. None of them prove that a government authorised you to sell this specific service, which is the only question the Google Ads policy asks.

Read the lead-time column as a warning rather than a schedule. Those figures are modelled from folkfox certification work across gambling, financial services and healthcare, not measured from a public dataset, and the honest headline is that the accepted routes are the slow ones. The refused routes are quick precisely because they are things you can generate yourself.

There is a quieter cost sitting under all of this. Google says it will automatically attach a 'Not a government website' disclosure to Search ads in these categories unless the advertiser is certified as a government provider, per its own policy documentation. So the uncertified advertiser is not simply running as before. They are running with a label that tells every searcher exactly what they are not.

That disclosure is a click-through rate change dressed as a compliance detail, and it will land in your reporting long before anybody in the business connects it to a policy page nobody read. Budget bleeds quietly in that gap, which is why the Google Ads policy deserves a diary entry rather than a forwarded email.

Why the Google Ads policy update 2026 is a documentation project#

Marketing teams treat policy changes as a checkbox because most policy changes are one. Rewrite the headline, swap the landing page, resubmit, done by Thursday. A certification regime does not behave that way, and every team that treats this Google Ads policy like a copy edit will discover the difference on the day enforcement starts rather than the day the work should have begun.

The distinguishing feature of a certification regime is that the critical path runs outside your building. You can rewrite an advert in an afternoon. You cannot make a government department reply in an afternoon, and the whole Google Ads policy hinges on that reply existing in public, on their domain, naming you. Everything else is undergrowth.

Reading the Google Ads policy like a compliance file#

Two of those eight weeks belong to Google rather than to you. Advertiser verification is its own process with its own clock, and Google's advertiser verification documentation says it will notify you in-account or by email when verification is required and what your deadline is, that some of your ads may be restricted while it runs, and that in certain cases Google Ads will pause the account until verification is complete. A pause is not a warning. It is the outcome.

The consequences layer is harsher than most media plans assume. Google's account suspension documentation states that egregious violations bring immediate suspension with no prior warning, that repeat violations carry a notice sent at least seven days before suspension, and that new accounts created by a suspended advertiser may be suspended too. The escape hatch some teams imagine, spinning up a fresh account and carrying on, is explicitly closed.

Scope is the second trap, and it is subtler than the first. An authorisation to help citizens with one document is not an authorisation to advertise the adjacent three, however similar they look in your campaign structure. If a single ad group mixes a permitted service with a neighbouring one, the whole group is exposed. Split them now, while splitting them is a spreadsheet task rather than a suspension appeal.

Geography quietly rewrites the media plan#

The targeting clause reads like housekeeping and behaves like a budget cut. Authorised providers must advertise only the services their authorisation covers and restrict targeting to the relevant geographic regions, as reported by Search Engine Land. Inherently cross-border services, electronic travel authorisations and border entry documents among them, are excepted.

Consider what that does to a media plan built on national reach. If your authorisation came from one region, your addressable audience shrinks to that region, but your fixed costs do not shrink with it. Creative production, landing page maintenance, compliance review and account management all stay where they were while the denominator collapses. Cost per acquisition does not merely rise, it changes shape, and the smart bidding history that made your account efficient was trained on a market you are no longer allowed to buy.

Notice that nothing in that ranking is a creative problem. The work that decides whether you are still advertising in October is administrative, unglamorous and entirely unsuited to the last fortnight before a deadline. Teams that lose this lose it in an inbox, not in an ad account, and no Google Ads policy has ever been outfoxed by a better headline.

There is a second-order effect worth naming. Once national campaigns fragment into regional ones, the account structure that survives is the one that was already built around geography rather than around product. Rebuilding that structure under deadline pressure is how conversion tracking breaks, so if a rebuild is coming, do it before the certification file lands rather than after. Our PPC practice spends a surprising share of its time on exactly this: account architecture that anticipates a regulator rather than reacting to one.

Google is also consolidating its grip on lead-generation categories more broadly. On 5 August it published three new Help Centre pages covering Performance Max pay-per-lead goals ahead of Local Services Ads moving into Google Ads, as noted by Search Engine Land. Read the two stories together and the direction is unmistakable: more automation, more verification, and more of the category rules living inside Google rather than inside your brief.

Gambling, finance, healthcare, and now government documents#

Here is the reassuring part, and it is genuinely reassuring: this pattern is not new, it is simply new to this category. Every restricted vertical folkfox works in has already walked this trail, brushing through the same thicket of registers, referees and requests. The shape of the walk is consistent enough to plan against rather than react to, and the Google Ads policy arriving in October is the fourth verse of a very familiar song.

Gambling went first. Google's gambling and games policy allows gambling-related ads only where the advertiser holds the proper certification, and it requires that ads target approved countries, carry responsible gambling information on the landing page, and never target minors. That combination, certification plus geography plus a mandatory on-page disclosure, is the exact template now arriving for government documents. We wrote about the downstream consequences in how regulators moved from reading the ad to reading the funnel.

Financial services followed the same path. Google's financial products and services policy requires compliance with state and local regulations for every location an advertiser targets, and states that a verification process is required to advertise financial services in some locations. For complex speculative products the bar climbs again: the advertiser must be a licensed provider or aggregator with an account approved by Google. Anyone running FinTech marketing recognises the choreography immediately.

Healthcare completes the set. Google's healthcare and medicines policy requires advertisers to apply before serving ads for prescription drug services and requires pharmaceutical manufacturers to be certified, with country-specific registers doing the vetting: the General Pharmaceutical Council in the United Kingdom, LegitScript and the National Association of Boards of Pharmacy in the United States. Poland is stricter still, with online pharmacy promotion not permitted at all. Our healthcare marketing work starts with the register, never with the creative.

Why Google moved, and why it will not move back#

The harm this policy addresses is old, documented and expensive. The Advertising Standards Authority has pursued copycat websites for years, holding that a company should not imply it is affiliated with, or offering, an official service when that is not the case, while conceding it has no direct power to stop such sites operating altogether. The UK government funded the fight directly back in 2014, when the then Consumer Minister Jenny Willott committed an additional 120,000 pounds to the National Trading Standards Board so it could investigate these websites, per GOV.UK.

Rival platforms reached the same conclusion, and reached it first. Microsoft Advertising's third-party government services pilot, announced in March 2023, requires review and pre-approval plus proof of delegation for each product or service in the applicable state, and launched limited to the United States and to three categories: recreational passes and licences, vehicle registration and title services, and vital records. Regulation is pushing the same way. The European Commission's Digital Services Act guidance requires ads to be clearly labelled with information about who placed them, obliges very large platforms to maintain a public advertisement repository, and bans targeting built on protected characteristics or aimed at minors.

One last thing, and it is the thing most likely to save you. The Google Ads policy landscape rewards the advertiser who reads the primary page rather than the summary of it. Trade coverage is fast and useful, and we cite it here gratefully, but the certification decision will be made against Google's wording, not against a headline. Read the policy. Screenshot the policy. Diary the policy for a re-read the week before enforcement.

For anyone whose whole category runs on permission, from casinos to clinics to crypto exchanges, this is simply the ordinary weather. Restricted category advertising has never been about clever copy. It has always been about who is willing to do the dull, documented, deeply unfashionable work six weeks before it is needed, and that work is what a conversation with folkfox tends to start with. If your category is heading the same way, our brand strategy and Web3 marketing teams have run this exact drill more times than anyone should have to.

Frequently asked questions#