Following several years of testing and integration of machine learning algorithms to power media buying, Google announced that they will begin to phase out standalone app promotion campaigns for search, video, and display in favor of their Universal App Campaign (UAC) format. The UAC format uses machine learning to guide ad formats, bidding, and placement across Google properties. On October 16, all new app campaigns will run as UACs, and after November 15, the standalone app campaign formats will be unavailable.
Google launched UACs in late 2015, and apps were a natural fit given the end goals (initially app downloads) and the sources (Play Store, with iOS apps added later). Apps are also key to Google’s pivot to a mobile-first and AI-first world, as they say, with nearly 85% of time spent on phones in apps, according to eMarketer. UACs were not the first of the dynamic campaign formats, though they are the first to foundationally rely on machine learning to scale the management of message variation, ad units, and placements rather than primarily through manual setup and optimization. Creation of a UAC requires only assets (up to 4 lines of text, 5 videos, 10 images), settings such as language, locations, bids, and defining a campaign goal — either installs or in-app actions.
How it Works
According to Google, the focus on Universal formats for app campaigns is due both to performance and to help advertisers better manage the complexity of advertising campaigns through automation. Google cites numerous case studies, noting that the UAC format has consistently seen lower cost-per-install (CPI) and higher click-through-rates (CTR) than the standalone formats. A single Universal App Campaign scans Google properties and attempts to find the most valuable app users at the lowest cost.
For advertisers, the move to UACs for app promotion campaigns means added simplicity in management and the smart integration of machine learning algorithms at the expense of some visibility and control. While this may be daunting for advertisers that are accustomed to the more hands-on approach with standalone campaigns, the machine learning that powers UAC has proven its value from a performance perspective at a macro level from Google’s testing, and with Resolution brands. Optimization remains key to UACs and any other machine learning-powered advertising. While many of the “levers” of optimization become less granular, it’s no less important to “manage the algorithm” through proper budgeting, creative asset development and testing, and analyzing campaign and app-level performance to drive insight and continuous improvement. More broadly, UACs also allow campaign managers to allocate more time to thinking strategically about building the brand and app user base.
Currently, marketers have limited visibility to what drives the success of UAC. While channel performance is exposed in reporting (Google Search, Google Display Network or YouTube), the inputs and decisioning that the algorithms employ to make UAC highly effective remain relatively obscured to marketers, including creative performance, queries, audiences, sites and apps.
This is a notable, yet incremental step in the integration of machine learning into the way that digital advertising is managed through AdWords. Machine learning should be welcomed, and smartly integrated into workflow as data points that can power effective digital marketing continue to explode, and make the manual synthesis of this information and the ability to take action more difficult. While marketers might lose some control over, for instance, the keyword that drove the most installs or the creative + placement that generated the highest engagement, they gain the ability to focus more strategically on business goals and in optimizing the customer journey.
Part of this will come from an integrated approach between a brand’s UACs and their app analytics packages, as the algorithms that power UACs can optimize towards installs, post-install actions, or post-install value. Advertisers will also succeed by collaborating on an omnichannel app promotion and performance approach that focuses on strategy and business goals through creative planning and testing, budget flighting, and audiences. Advertisers will need to:
- Be patient: Machines require time to learn and find the most valuable consumers for your business. Do not let week 1 of UAC forecast the success of your app. Test for at least two weeks to allow the algorithms to properly learn.
- Expect change: If you only managed one Google tactic previously to drive installs, performance will change significantly. From scalability to engagement, the Universal App Campaign will effectively reconfigure how you reach consumers.
- Accept the unknown: Advertisers thrive on data and insights. While you lose some visibility, you gain efficiencies for both media and management. It is worth noting that app installs can still be driven in a more targeted fashion through other formats, such as App Extensions in search ads.
The mechanics of UAC allow advertisers and brands to focus on long-term business challenges and growth, rather than managing the scale of finding optimal placements, bids, audience, and message combinations. This technology is here, and marketers need to understand how to most effectively manage it to drive business results. Resolution will continue to partner with Google on the development of UACs and the intelligence that is exposed to drive insight on creative performance.