Home Platforms Advertisers Finally Win Key Transparency Concessions From Google’s Performance Max

Advertisers Finally Win Key Transparency Concessions From Google’s Performance Max

SHARE:
Comic: Black Boxes

Almost five years have passed since Google quietly introduced a beta product called Performance Max (PMax).

There are now more than one million advertisers using PMax, Google boasted in an announcement on Wednesday. But for many of those advertisers, it’s been five long years of screaming, flailing, cajoling and waving their arms about as they sought campaign reporting info from the black box.

Finally, it’s happening. Advertisers will soon be able to see around three of the most important transparency blockers in PMax: channel-level reporting, search terms reporting and creative asset reporting.

Google will open a beta program for channel-level reporting and will roll out the other updates for search terms and asset reporting over the course of the year, the company also announced on Wednesday.

Crossing the channel

What are these reports and why are they important?

Well, PMax serves ads across almost the entire Google media fleet (Search, YouTube, Gmail, Maps, Discover, Shopping Ads and the ad network). However, Google provided no visibility into where those ads served or how much had been spent on the placements that you can’t see. Advertisers also didn’t know what formats were used or which specific videos or product pages in a shopping campaign were performing well or poorly.

The channel-level reporting beta, which will launch after the Google Marketing Live event next month, shows granular stats on how many impressions were served to each channel in the PMax mix, Google Ads product director Tal Akabas told AdExchanger during a press roundtable prior to the announcement.

“You’ll be able to see the number of impressions, the number of clicks, the number of conversions from each of those channels, as well as cost,” Akabas said.

Getting the cost visibility is a key win, too, because even while PMax had cracked open slightly to reveal some ad placements and format types, it has never revealed the spend breakdowns between different channels.

For example, advertisers might have been able to glean some PMax ads serving to YouTube and other channels. But it was a jury-rigged and inconsistent way to reveal which video ads were served by PMax, because it’s hacked together by vendors rather than clearly reported by the Google platform.

The channel-level reporting in beta will disclose where those ads were served. And the new asset reporting shows which specific video assets were used in a PMax campaign, while retailers will be able to see which products from their catalogue were pulled by PMax, which is another important bit of information they had with their manual shopping campaigns.

Searched and found

Another long-running PMax transparency controversy is its search reporting. PMax’s pitch has partly been that Google AI can find new search keywords and pockets of growth. For instance, if a museum has great activities for kids, the PMax algo might sniff that out, steer search ads toward child care-related search terms or things to do for parents, write new ad copy and run a whole campaign along those lines. And the museum might not even know it’s happening.

One of the earliest PMax transparency concessions was for branded search. Advertisers were rightly horrified by PMax’s initial search tactics, because it bought branded search terms (the company’s own name and products) and thus could cheaply self-attribute purchases that were about to happen anyway.

Even with controls to block PMax from targeting certain terms, advertisers still didn’t have the same visibility as they do with their own standalone search campaigns. The new search terms reporting, though, which also goes live after Google Marketing Live, will create parity between the campaign reports an advertiser sees using their own single-channel Google Search campaigns, Akabas said.

Why not always?

PMax has had to negotiate an awkward growth spurt.

It needed to demonstrate to advertisers that Google’s machine learning could produce superior and reliable results based on the business’s stated outcomes. But to do so, it’s had to push advertisers into new channels, placements and formats they wouldn’t think of or might even explicitly avoid if the human ad buyer was at the controls.

“We really wanted Performance Max to be the place that advertisers could go to branch out from just buying by single channel and see how those channels could work together to deliver maximum performance,” Akabas said.

The lack of transparency with PMax has no privacy rationale, like how advertisers are losing visibility into cookies, device IDs and IP addresses in many channels. But Google must balance how much info to give advertisers without them ditching the product while seeing results to justify the spend.

“We didn’t want to release the granular stats without any explanation, because our worry was it would either be misinterpreted or advertisers wouldn’t know what to do with it,” as Akabas put it. PMax has focused instead on adding back reporting info “within the context of how it drives your business objectives and diagnostics, so that advertisers feel like they’re in control to some extent,” he said.

If these PMax updates remain on track, then before next year advertisers should have near parity between the channel-level information reported by PMax compared to their standalone, single-channel or human-controlled campaigns.

“Advertisers want to see both the raw data as well as the inputs [made by Google’s machine learning],” Akabas said. “We try to offer both where it makes sense.”

Tagged in:

Must Read

Ad Performance Hinges On Kicking Fragmentation's Butt

As performance takes center-stage in more advertising discussions, demands to solve fragmentation and cruddy measurement are reaching a fever pitch.

AdExchanger's Big Story podcast with journalistic insights on advertising, marketing and ad tech

AI Off The Rails

A word of caution to digital advertising companies, as they go all in on AI algorithms: They need to build these solutions with ownership, governance and accountability from the start – or AI could sink them with a single mistake.

square Headshot of Mohammad (Moe) Chughtai, global VP of strategy & partnerships at MiQ, against an orange and yellow gradient background

Better Attribution Makes Live Sports A Performance Play

To squeeze the most juice out of their live sports campaigns, many marketers are adopting programmatic buying and marketing mix modeling, both of which are also drawing more advertisers to the digital live sports cornucopia.

Privacy! Commerce! Connected TV! Read all about it. Subscribe to AdExchanger Newsletters

Roblox Opens Up Advertising To Kids Under 13

Roblox is making its under-13 audience available to advertisers for the first time. And it named youth-focused ad marketplace SuperAwesome as its exclusive advertising partner for under-13 users.

Comic: Header Bidding Rapper (Wrapper!)

Outgoing Prebid President Mike Racic On His Departure And The Org’s Next Act

Prebid is turning the page on what might be called its second chapter as the organization navigates some major changes in the digital advertising landscape and within its own ranks.

Meta is giving advertisers the ability to connect their third-party analytics tools directly to its ad platform via API.

How Apparel Brand Tuckernuck Devised The 'Why' Behind Its CTV Ad Performance

Performance CTV tech company Keynes launched an AI-powered platform. Tuckernuck says it can finally “pop open the hood” and see what’s working.