Google now counts a one-pixel view in Demand Gen. Which view-through settings should you keep?

KEY TAKEAWAYS

1. A view in Demand Gen now means one pixel on screen for any time. The old Display standard was half the ad for one continuous second.
2. View bidding covers video assets only and is on by default for new campaigns, so the setting is a choice to make at launch.
3. Work out the share of view credit that must be real for the campaign to hit target. In our hypothetical example it is 60%.
4. Do not compare Display view-through counts across October 2026, and do not migrate a Display campaign in the middle of a lift study.

On October 5, 2026, Search Engine Land reported that Google is changing what counts as a view for Display ads inside Demand Gen campaigns. The old test was Active View. The new one is a rendered impression, which means at least one pixel on screen for any length of time. It is the third change to view-through conversions in Demand Gen since June. Together they turn a reporting column into something Google bids on, bills on and switches on for you. This article separates the decisions.

What changed between June and October 2026?

A view-through conversion is a conversion from someone who was served an ad, did not click or engage, and converted later. Google's help page ranks credit in a fixed order of clicks, then engagements, then views, and each conversion is credited once.

The August items came from a Google email that a practitioner shared on LinkedIn, and Google said they would arrive over the coming months. Google's US help page now matches them. It says the feature is in open beta, applies to video assets only on YouTube, Display and the Discover feed, and is enabled by default on newly created campaigns. Its billing FAQ lists impression billing for YouTube, for the Display Network and for Discover in any campaign that bids on views.

One point of timing matters. The one-pixel rule is not new for bidding. The bidding help page already defines a view that way, and Thomas Eccel described the same threshold in an October 2025 post about the beta. What the October report adds is Display ads. According to the report, Google's stated aim is consistent view definitions across inventory. Google published no estimate of how many more conversions the new rule will credit.

How much lower is the new bar?

Google's viewability page gives the old standard. A display ad was viewable when at least half its area was on screen for one continuous second. Ads of 242,500 pixels or more needed 30%. The Media Rating Council set those thresholds in 2014, and its guideline explains the reason. A served impression may or may not have had an opportunity to be seen.

The size of the gap can be bounded. The 21st edition of the Integral Ad Science Media Quality Report, covered by Advanced Television in July 2026, put global display viewability at 67.9%. If every impression that failed the standard still showed one pixel, the pool of impressions able to earn view credit would grow by 1 divided by 0.679, or about 47%. That is our calculation and it is a ceiling. The benchmark covers the open web, not Google's inventory, and some failed impressions never reach the screen at all.

The added impressions are the weakest ones. They are slivers at the edge of a scroll. Any conversion credited to them is far more likely to be a purchase that was already on its way.

Do view-through conversions overstate or understate what the ads did?

The evidence points both ways, and the two sides are measuring different things.

The case that platforms undercount comes from Haus. In March 2025 it published results from 190 geo holdout tests of YouTube campaigns across 74 brands. Incremental lift in direct sales averaged 3.4 times what Google Ads reported. The tests covered conversion-optimised video campaigns, including Demand Gen. Haus added that the ratio can differ three or four times between similar brands, so it is not a benchmark.

The case that views overcount is older. Johnson, Lewis and Nubbemeyer ran a controlled retargeting experiment on the Google Display Network in 2014 with 566,377 users. The ads raised the number of buyers by 12.1%. Put the other way round, about 11 in every 100 exposed buyers were caused by the ads and 89 would have bought anyway. That is our arithmetic on their result, and it describes one campaign. A one-day window narrows the pool but does not remove the problem. Jon Loomer's position on Meta is similar. He accepts view-through numbers as context and says an account whose conversions are all view-through should disregard them.

A popular test on the other side is weaker than it looks. WordStream describes an agency that ran a blank banner against a real one. The real banner recorded 9 view-through conversions from 2,418,973 impressions. The blank one recorded none from 315,677. At the real banner's rate the blank ad would have produced about one. Zero is not evidence.

Both findings can hold. Haus measured video on YouTube, with a method that sees sales the platform cannot attribute, before view bidding existed. The October change concerns banners at one pixel. Video may be under-credited while a sliver of a banner is over-credited. Google's own help page is candid about the limit. It says view bidding "doesn't directly focus on incrementality". It also says the feature is expected to have no effect on conversion lift tests.

Which setting should you keep?

Google's own guidance supports the first row. The help page says view bidding may not suit advertisers who only care about click performance or who evaluate Demand Gen outside Google Ads. It also warns that the extra conversions will not appear in your backend systems.

The opposing view deserves a fair hearing. Google's product page argues that bidding on views gives a like-for-like comparison with social platforms, and Eccel expects faster learning from the extra signal. Both are reasonable for a video-first advertiser. Meta advertisers face a similar question, which we covered in our piece on Meta attribution settings. Parity is an argument for comparing platforms on the same basis. It is not proof that either number is right.

How much of the view credit has to be real?

One sum turns the argument into a threshold. These are our formulas and an invented account.

Required share = (spend / target CPA - click and engaged-view conversions) / view conversions.

A hypothetical campaign spends $9,000 in 30 days. The Ad event type segment shows 70 conversions under Clicks, 35 under Interactions and 75 under Impression. Google Ads reports 180 conversions at $50 each. Without views the cost is $85.71. The target is $60, which needs 150 conversions. The required share is (150 - 105) / 75, or 60%.

So this campaign meets its target only if six in ten view conversions were caused by the ads, and that assumes every click conversion was. Set that against the Display experiment above, or against Haus for video, and decide how comfortable you are. A result under zero means the campaign pays without views. A result over 100% means it cannot reach target on any reading.

A seven-step method

1. Find the setting. In each Demand Gen campaign, open settings and look under conversions optimisation for the option to include view-through conversions. Note its state. The US and Australian help pages disagreed on the default when we read them on October 8, 2026, so do not assume.
2. Check the window. Under Goals, Conversions, Summary, edit each primary conversion action. The view-through default is 1 day. Changes apply only going forward.
3. Segment 30 days. Use Segments, Conversions, Ad event type. Record Clicks, Interactions and Impression.
4. Run the sum. Calculate the required share for each campaign.
5. Record a Display baseline. Save the view-through conversions column for Display placements with today's date. Expect a rise that reflects the definition, not the ads.
6. Test. Conversion Lift based on users supports Demand Gen, and Google says it ignores attribution windows and compares all conversions in the exposed and held-back groups. Use the incremental cost per acquisition from that study as the yardstick. This is the same discipline we suggested for AI Max settings.
7. Decide at launch. Add the view setting to the checklist for every new campaign.

What if you are about to migrate a Display campaign?

Google is folding standalone Display campaigns into Demand Gen. Its help page says a migration tool began rolling out in June 2026, that new Display campaigns will later be created only in Demand Gen, and that remaining campaigns will then move automatically. It gives no dates for the last two stages. Search Engine Journal's Brooke Osmundson advised against waiting for the automatic move.

Three details affect view-through reading. The tool carries over up to 42 days of history, so a migrated campaign's record can straddle two view definitions. A migrated campaign cannot be reverted. And Google says not to use the tool during an active lift study, because the new campaign will not be linked to it.

None of the sources we read says whether Display campaigns that have not yet migrated are covered by the October change. Treat that as open until Google documents it.

Where this stops

Google's notice for the October change is not a public page that we could read, so the wording here relies on Search Engine Land. The general help page on view-through conversions carried no dated note about it on October 8. The viewability ceiling uses an industry benchmark, and the worked example is invented.

The habit is the one we apply to creative tests in how many ad concepts a budget can test. Know what a number can prove before it moves money. If measurement you can defend is the problem you are working on, the waitlist is open.

Questions

What changed for Demand Gen view-through conversions in October 2026?

Search Engine Land reported on October 5, 2026 that Display ads inside Demand Gen campaigns are moving from the Active View standard to a rendered impression. A view now counts when at least one pixel is on screen for any length of time. The change arrives automatically over the following weeks.

Is view-through conversion bidding on by default in Demand Gen?

It depends on when the campaign was built and which help page you read. Google's US help page says newly created Demand Gen campaigns have it enabled by default. The Australian version of the same page still said disabled when we read both on October 8, 2026. Existing campaigns keep their setting. Check the box in each campaign.

What view-through window should a Demand Gen conversion action use?

One day. That is Google's default for view-through windows and its stated recommendation for primary view-through conversions. The setting allows up to 30 days, and a longer window raises the reported count without changing what the ads did. Changes to a window apply only to future conversions.

How do I find view-through conversions inside the Conversions column?

Segment the campaign table by Conversions, then Ad event type. Google's help pages say click conversions show as Clicks, engaged-view conversions as Interactions and view-through conversions used for bidding as Impression. The separate view-through conversions column lists all of them, whether or not they feed bidding.

Sources

1. Search Engine Land, Google changes view-through conversion reporting for Demand Gen Display ads searchengineland.com
2. Google Ads Help, About view-through conversion optimized bidding for Demand Gen campaigns support.google.com
3. Google Ads Help (en-AU), About view-through conversion optimised bidding for Demand Gen campaigns support.google.com
4. Google Ads Help, About view-through conversions support.google.com
5. Google Ads Help, Understanding viewability and Active View reporting metrics support.google.com
6. Media Rating Council, Viewable Ad Impression Measurement Guidelines iab.com
7. Search Engine Land, Google updates Demand Gen view-through conversion optimization searchengineland.com
8. PPC News Feed, Google updates Demand Gen VTC optimization ppcnewsfeed.com
9. OneMetrik, Demand Gen VTC optimization goes video-first onemetrik.com
10. Search Engine Land, Google Ads shifts Demand Gen billing to CPM for some Discover campaigns searchengineland.com
11. Google Ads Help, Demand Gen Frequently Asked Questions support.google.com
12. Google Ads Help, About conversion windows support.google.com
13. Google Ads Help, About Engaged-view conversions support.google.com
14. Google Ads Help, Google Display Ads campaigns have a new home in Demand Gen support.google.com
15. Search Engine Journal, Google Is Retiring Standalone Display Campaigns In Favor Of Demand Gen searchenginejournal.com
16. Google Ads Help, Comparing lift types support.google.com
17. Google Ads Help, Understand your Conversion Lift based on users measurement data support.google.com
18. Accelerate with Google, View-through conversion (VTC) optimization business.google.com
19. Search Engine Journal, Google Adds View-Through Conversion Optimization To Demand Gen searchenginejournal.com
20. Thomas Eccel, Demand Gen View-through Conversion Optimization thomaseccel.com
21. Haus, Do YouTube Ads Perform? Lessons From 190 Incrementality Tests haus.io
22. Johnson, Lewis and Nubbemeyer, Ghost Ads, Improving the Economics of Measuring Ad Effectiveness conference.nber.org
23. WordStream, Can We Trust View-Through Conversions? wordstream.com
24. Jon Loomer Digital, Do View-Through Conversions Matter? jonloomer.com
25. Advanced Television, Report. Video viewability outperforms display advanced-television.com

NEXT STEP

See where your brand stands.

Your Digital Authority Score, who AI recommends for your target questions, and the placements that close each gap. Your first report is free.

Get your free report →