Oh, the confusion a poorly worded Google Help Center update can bring.
A recent update to the Google DV360 Help Center had some advertisers panicking that invalid impressions served through Google’s ad platform would effectively double starting in mid-July.
The update amplified concerns that increased bot activity online due to generative AI adoption is producing more bot-driven ad fraud for programmatic platforms to filter out.
But, according to Google, that’s not what it meant to say at all. A Google spokesperson attempted to clarify the update after AdExchanger reached out for comment.
Keeping it 100%
Before we get to the clarification, though, let’s explain the confusion.
Last week, Google published a support page update entitled “Coming soon: Changes to MRC accredited metrics in July 2025.”
The page said that, starting on July 14, “there will be a change in how invalid impressions are counted” in reporting for buyers who use Google’s DV360 DSP and its Campaign Manager 360 ad management and measurement interface.
According to the update, both “Invalid Begin to Render Impressions” and “General Invalid Traffic (GIVT) Begin to Render Impressions” will “increase by an estimated 100% starting July 14.”
“Invalid Begin to Render Impressions” is a Google term of art for ad impressions that are initially served to bots, but which Google later successfully identifies as invalid traffic (IVT). Advertisers are not billed when this happens.
Then there’s “General Invalid Traffic (GIVT) Begin to Render Impressions,” which is Google’s way of distinguishing declared bots that were initially exposed to ads before being caught. Unlike undeclared bots, which are often malicious, declared bots are typically benign. Think of the search crawlers that Google uses to index websites for search results.
AdExchanger Daily
Get our editors’ roundup delivered to your inbox every weekday.
Daily Roundup
Google’s reporting interface discloses how many of both types of invalid impressions were served during a given campaign as part of its compliance with its Media Rating Council (MRC) certification for IVT detection, the Google spokesperson said.
Easy enough to understand. But, naturally, the “will increase by an estimated 100%” language in the Help Center update led some in the ad industry to believe that Google was saying overall IVT activity detected by its ad platform is set to double.
A concerned source flagged the update to AdExchanger.
These are legitimate fears, as online activity from AI agents and other bots continue to ramp up. A recent study by security tech firm Thales claims bot activity now accounts for 51% of online traffic.
But, in this case, the fear is unfounded – according to Google, at least.
DV360’s increased IVT detection isn’t related to the rising bot traffic trend, the Google spokesperson told AdExchanger. It’s about Google changing which underlying data set DV360 is basing its campaign reporting on.
Reporting data
Here’s what’s happening.
Starting on July 14, DV360 will switch to using Google Ads data as its underlying data set for campaign reporting, rather than data derived from Campaign Manager 360, according to Google’s spokesperson.
As a result of the switch to aligning DV360 reporting with Google Ads data, “advertisers may see some duplicate impressions, which our spam filters are able to detect and filter,” the spokesperson said. These duplicate impressions will be flagged as IVT – thereby contributing to the inflated IVT numbers that Google anticipates it will soon be reporting.
Such duplicate impressions could be caused by accidental page reloads or other technical glitches, the spokesperson said. But these impressions “are not necessarily indicative of fraud,” they added.
According to the spokesperson, the increased IVT detection is also not indicative of increased bot activity on Google’s part – say, from increasing search crawler indexing or scraping more websites to train its Gemini AI model.
Also, while the Help Center update mentioned 100% increases in IVT detection, this actually reflects an increase from very small counts of impressions mistakenly served to IVT before being flagged, the spokesperson said.
Currently, an advertiser purchasing 1,000 ad impressions via DV360 is likely to see a single-digit number of impressions flagged as being served to IVT in their campaign report, the spokesperson said. For instance, they might just see three impressions flagged, they said. In that case, once DV360 switches to the Google Ads data set, a 100% increase would mean Google flagging six impressions as IVT due to impression duplication.
But regardless of whether DV360’s campaign reporting flags three impressions, six impressions or more as IVT, the advertiser would still get 1,000 valid impressions and not be billed for the IVT, the spokesperson said.
Clarification needed
Given the confusion its Help Center update caused, the spokesperson told AdExchanger that it plans to update its notice “in coming days to provide additional clarity.” The misleading Help Center page has been taken down, and the spokesperson said it will be completely rewritten.
For example, in addition to the ambiguously worded info on IVT and GIVT detection, the original Help Center update also included a note saying that “Inactive Impressions will increase by an estimated 30%.”
“Inactive Impressions” is another Google-ism for describing non-viewable impressions that load below the fold, for instance, or in the background of a page.
The note about Inactive Impressions increasing by 30% will likely be removed entirely from the revised Help Center update, “because, actually, that’s not happening,” the Google spokesperson said.
Glad we could help clear all that up!
