Before Performance Max, There Was Google's App Campaigns, Muddying Transparency

Automation is taking hold of the ad industry, and marketers are grappling with the idea of balancing transparency with effectiveness, spurred by interest in Google’s year-old tool Performance Max. And if history is to be believed, performance often wins. But that doesn't mean marketers are happy about it. Back in 2017, Google told mobile advertisers who run campaigns optimized to drive app installs that they had to use one tool to purchase across this menagerie of media. Machine learning would make all the decisions as to where each creative should go, who to target and with what keywords. Five years later, the mobile tool, now called App Campaigns, has drawn criticism for forcing advertisers to purchase unwanted inventory, offering inadequate transparency and control, and a perceived overuse of branded search, which some marketers think is inefficient, four sources told Adweek. Still, most said that ad performance is often impressive, alleviating some ire. “For some, it works. For others it doesn’t,” said Thomas Petit, an independent app development consultant. “In both cases, you just put money [in] and hope for the best.” The industry’s appraisal of App Campaigns is particularly relevant as automation becomes a pillar of Google’s strategy for advertisers. Last year, Google rolled out Performance Max globally, which automates campaigns across media types and is seen by some as App Campaigns’ web counterpart. AdExchanger detailed marketers’ gripes with Pmax, as its abbreviated, which mirror longer-standing complaints with App Campaigns. And, ahead of an ad downturn, marketers will gravitate further towards solutions that prove performance. “Everyone in the mobile apps space saw the Pmax stuff coming from a mile away, because [Google] already did it in the apps world,” said Adam Lovallo, founder of digital agency Thesis, which advises app clients, and co-founder of MAU, an annual conference for mobile app professionals. As Privacy Sandbox Comes to Mobile, Queries on Privacy and Effectiveness Remain A Google spokesperson said the company is leaning into automation to simplify marketers’ jobs in an increasingly complex media landscape, and that while Apps Campaigns and Pmax are both born out of this philosophy, they have separate and different product roadmaps. “With machine learning, App Campaigns are able to analyze hundreds of millions of data signal combinations to make the smartest decisions for advertisers,” said David Mitby, senior director of product management, Google Ads, “such as which creative assets would perform best and how much to bid.” Too much branded search One of advertisers’ biggest complaints about App Campaigns is that they get no control over what keywords are advertised on in mobile searches, nor transparency over where the algorithm eventually decides to send their money. Some complaints revolve around when marketers end up spending significant sums on branded search, which includes the company's name. There is debate in the industry about the utility of the practice, with some arguing that it wastes dollars on clicks a brand would have won anyway without advertising since someone is already using the company name in the search. “It’s really unclear how accretive branded search is, fundamentally,” Lovallo said. Google makes it extremely difficult for advertisers to turn off branded search for App Campaigns, which is still the only way mobile advertisers can run app install campaigns on Google. The only way is for marketers to call their Google rep, since there is no option in the ad-buying interface, and often the reps try to discourage the decision, Lovallo and Petit said. Advertisers are getting basically bamboozled.Adam Lovallo, founder of digital agency Thesis Google has the incentive to keep branded search in the mix even if some advertisers hate it, Lovallo and Petit said, because it's cheap and performs well, juicing the results of App Campaigns overall. On certain campaigns Lovallo has worked on where brands omitted branded search terms, App Campaigns performed significantly worse, he said. “There are, without a doubt, a lot of advertisers are getting basically bamboozled and do not realize branded search is driving the performance,” Lovallo said. The Google spokesperson said that Apps Campaigns do not optimize for any specific inventory, including branded search. Lack of control In addition to branded search, marketers worry that their ads will appear next to unsavory content since part of the inventory App Campaigns draws from is Google’s app ad network AdMob, which includes a long tail of publishers. “Their AdMob inventory has some legit parts and tons of very, very poor traffic, including shady apps with ad formats that are questionable,” Petit said. The Google spokesperson noted that the company has policies to ensure its apps are high quality, including a new app review process, and that marketers have access to brand safety and suitability controls and placement reports. [App Campaigns] is a black box, but it’s achieving results.Catherine Reich, senior director of paid search at Tinuiti But marketers said the lack of transparency, especially compared to the older process of setting up manual campaigns, omits valuable data: Insights into keywords can be proxies for campaigns on other channels, such as those with influencers, and knowing which creative worked on Google inventory can inform decisions for the rest of the media plan. For all marketers’ gripes, many find that App Campaigns' machine learning works, in part by opening up inventory that may have previously been too difficult to set up manually as separate campaigns. Google Is Backing Cookieless ID Solution Seller-Defined Audiences "[App Campaigns] is pretty much a black box, but it’s achieving results,” said Catherine Reich, senior director of paid search at Tinuiti. “I think the results outweigh the difficulties with lack of insights.” Still, automation and transparency aren’t inherently a zero-sum game. “There is nothing that prevents full automation to show where and how ads were placed and their results, except the interest for Google to hide this so results look great to people who don’t want to ask how or why, nor value the insight,” said Petit.