Google uses machine learning for ad frequency control, and third-party cookie tracking is completely cool
Google announced on Tuesday that it would launch an ad frequency control tool for non-dependent cookies for its DSP platform, Display and Video360 (DV360), in the coming weeks.The company also said it plans to add the tool to Google Ads in the future.
The tool uses machine learning to analyze traffic patterns when third-party cookies are available, and then builds models to predict traffic patterns without cookies.
"This allows us to estimate how users are accessing different media that use Google Ad Manager to push the same ads.When there are no third-party cookies, we are able to optimize the frequency with which ads are placed to our users.Raheul Srinivasan, Google's advertising privacy product manager, said in a statement.
However, industry reviews say this feature is a far from a real solution to the problem of limiting the frequency of online advertising.
On the one hand, this is the frequency limit within the Google system, not across platforms and devices, which is for marketersis the key issue.Also, it uses modeling data instead of tracking the number of ad exposures a user receives for true frequency control.
Simon Harris, head of sales for MayyHive europe, the Middle East and Africa, says: "Marketers tend to be sceptical about the products of the probability model , as ID analysis has always been , because they can be difficult to verify.”
Earlier this year, Apple unveiled a solution to the "attribution chaos" problem on Safari that, like Google, requires marketers to accept the platform's machine learning without seeing any user's individual level of data.
However, in browser environments such as Safari and Chrome, where machine learning is built into the DSP platform is useful because third-party cookie orientation is not available, Harris said, "I expect that more tools for probability model classes will be available in the future when we want to balance advertisers' need for addressability with consumer demand for privacy."
'Google will not integrate data signals such as IP addresses into the tool because it could have an impact on user orientation,' Srinivasan wrote.IP addresses are useful for frequency restrictions, but users do not have the right to opt-out clauses as they do cookies to protect privacy.
Source: MartechToday; adexchanger.com
Go to "Discovery" - "Take a look" browse "Friends are watching"
sent to have a look