Revenue Editors Picks

Google’s ad-count reset: cleaner numbers, no promise of more money

Ivan Massow 21 Oct 2026 · 6 min read

A red-shirted man holds a huge AD placard beside a blue-clad reader, forcing her to lean around it with her newspaper. Caption: A better reading experience. Signed Noah.

Cartoon · AI-generated

Google is changing the point at which an advertisement counts as delivered. Today’s download-based count can register an ad before it starts appearing on the page. The new rule counts it when it begins to appear.

Noah gives a 77% chance that, on 17 February 2027, Google’s operative documentation confirms this “begin-to-render” counting rule for display banners across web, mobile web and connected TV in both AdSense and Google Ad Manager.

My judgement: this is a useful move towards more meaningful advertising numbers, but an awkward reckoning for publishers whose sales promises and budgets rely on the old count. Buyers gain a measure closer to actual delivery. Publishers may report fewer ads delivered to the same audience, and face harder questions about campaigns that span the change. A higher reported advertising rate will not, by itself, mean they are earning more.

For readers, the benefit depends on what publishers do next. Fixing slow or wasteful ad delivery could improve the page. Loading everything earlier to protect a sales target could make it worse. The right response is to improve delivery and explain the figures, not cram in more advertising.

Advertisement · Noah

A better-looking rate can disguise a business standing still

An advertising “impression” is a counted delivery of an ad. CPM is the price or revenue per thousand impressions. Change which deliveries count and the rate can move even when the money does not.

Here is an illustrative calculation, not a measured publisher result. A publisher earns £1,000 from 100,000 counted impressions: a £10 CPM. Under a later counting rule, only 90,000 of those deliveries qualify. If revenue stays at £1,000, CPM rises to about £11.11. The dashboard looks better. The bank account is unchanged.

That is why this matters beyond the advertising department. A board could mistake a measurement change for stronger pricing. A sales team could promise more qualifying deliveries than it can supply. An editor could be asked to produce more stories to repair an apparent loss of audience which never happened.

The practical test is money earned from a comparable volume of reader visits, alongside delivery against actual campaign commitments. Keep the old and new counts together long enough to explain the difference.

Britain’s preparation has already started

Content Ignite’s James Hanslip said on 6 October that its team was reviewing configurations, delayed ad loading and refresh across its publisher network. This is preparation reported by a supplier; it is not a measured UK revenue loss. It does show that the February change is already becoming operational work.

The company’s August partnership with Ozone connects that work to the wider British market. Content Ignite named The Observer, Global, Haymarket and The Weather Channel among its partners. Ozone’s founding publishers were News UK, Guardian Media Group, Telegraph Media Group and Reach. Those relationships show the shared infrastructure available to publishers; they do not establish any named title’s exposure or readiness.

A smaller publisher using a managed service should ask its supplier for one specific thing: a comparison showing how much of its own advertising qualifies under each counting method, with the same audience and inventory. A network-wide reassurance is a poor substitute for that answer.

Google already advises comparing “Ad server impressions” with “Ad server begin to render impressions”, using a historical start after 12 August 2026. Additional Ad Exchange, yield-group and total begin-to-render measures are marked as coming soon. Ask which sales channels the comparison actually covers.

Do not confuse a likely switch with proof of its effects

The opening forecast concerns Google’s operative counting rules on a specified day. The four forecasts below ask a different question: whether particular effects will be demonstrated in qualifying public evidence by their deadlines. Noah considers each of those published-proof outcomes unlikely within its window. None establishes that the underlying effect is inevitable later.

Only 30% — By 30 June 2027, a qualifying public case will show at least one UK news or specialist publisher recording at least 10% fewer display impressions per 1,000 UK sessions under the new rule than under download-based counting. The comparison must use simultaneous counts or matched periods, with session totals differing by no more than 5%. The other 70% is the chance that no qualifying public case establishes that result by the deadline, not that no publisher experiences it.

Only 30% — By 31 December 2027, a qualifying public UK news or specialist publisher case will demonstrate the “better rate, same money” effect: display CPM up at least 10% after the switch, while net display-ad revenue per 1,000 UK sessions changes by less than 5% in either direction against a comparable earlier period. That is a 70% chance of no qualifying published demonstration by then. The arithmetic is straightforward; whether a public case meets all those conditions remains uncertain.

80% chance of no qualifying public demonstration — By 31 December 2027, Noah expects the specified evidence of easier billing reconciliation still to be absent. The positive forecast is 20%: a public UK display-campaign case shows at least a 25% relative reduction in the absolute billed-impression discrepancy rate after buyer and publisher both adopt the new counting point, using both parties’ records for the same campaign or a matched predecessor. That means a relative reduction, not 25 percentage points. Do not budget for a proven reconciliation benefit yet.

83% chance of no qualifying published study — By 30 September 2027, Noah expects no public technical study to meet this full test: at least 10 UK news or specialist domains; at least 5% of downloaded display ads failing to start appearing before their slot leaves the screen or the visit ends; and creative or multi-partner delivery latency identified as the primary cause. Its YES probability is 17%. This is a forecast about a tightly defined study, not a finding that 5% of UK ads already fail.

These are modelled probabilities dated 21 October 2026. Their exact contracts, evidence requirements and uncertainty ranges remain in the private claim audit. For an executive, the message is clear: prepare for the likely platform change; do not treat its possible financial consequences or promised operational benefits as established facts.

Fewer ads can still be good business

There is a useful example of a publisher-services company deliberately removing advertising rather than trying to preserve every impression. Ezoic said a change implemented on 4 August removed about 25% of impressions, with viewability reaching 87% and Prebid win CPM rising about 54%, while publisher earnings held.

Those are Ezoic’s figures, not an independently controlled UK result. The CPM figure concerns one auction measure, not the publisher’s entire advertising business. Its near-screen loading policy also goes beyond Google’s minimum rule. It does not settle either 30% forecast.

It nevertheless illustrates the commercial opportunity worth testing: fewer wasteful deliveries, a better opportunity for ads to be seen, and earnings that hold up. Protecting the old impression total is not necessarily the right objective.

Starting to appear is still a lower bar than being viewable. Google’s usual display viewability test requires at least half an ad’s pixels to remain visible for one continuous second, with a different threshold for very large ads. Neither measure proves that someone noticed the message. Sales teams should promise the benefit their evidence supports.

Three decisions to make before February

First, save a comparable baseline now. Record old and new impression counts, net advertising revenue and reader sessions for the same inventory and periods. Separate direct sales from programmatic sales, and mobile from desktop where material. Otherwise, February’s new number will have no honest history.

Second, review campaigns crossing the change. Agree which count determines delivery and billing, and how buyer and publisher will explain any difference. Google’s separate buyer-side notice gives February 2027 as the estimated billing change for Campaign Manager 360 and Display & Video 360; it does not establish the same exact day as the publisher transition.

Third, test delivery changes against the reading experience. Prioritise important placements and investigate slow creatives. Google’s own guidance distinguishes fetching an ad from rendering it and supports delaying less urgent placements. Bringing every ad forward can compete with the article for the reader’s device and connection.

Google says publishers need take no action for the counting transition itself. That does not remove the commercial decisions. My advice is to prepare the comparison and settle the contract questions now. Better measurement is good news when it helps publishers sell a better product. It becomes bad management when a prettier rate is mistaken for more money.

Sources

Source: Tomorrow’s Publisher