Google Universal App Campaigns: All your burning questions answered
If you want to advertise with Google, now is a great time to read up on Universal app campaigns, (or UAC) and get all of your questions answered on the mobile ad campaign optimization service.
We loved having Lee Jones, Director of Global App Install Strategy at Google at Mobile Spree in San Francisco so much, we invited him back for a longer fireside chat with Adjust’s own Global Director of Product, Katie Madding. To listen to their entire conversation, watch the video or read along below.
Katie: You started with AdMob in 2010. What have you seen change in the past seven years on mobile, where there’s kind of constantly this creative destruction at play?
Lee: Well, a lot has changed. AdMob was the first to put ads on the iPhone. We went around trying to convince marketers were apps they were worth caring about. People’s strategies were geared toward getting you to the top of the App Store categories that mattered, which at the time you could calculate pretty easily. Over time, it evolved. People realized that apps were driving meaningful business and in the last few years, Facebook has helped to give people much more control with targeting demographics. In terms of where we are today, there’s no question that apps are extremely strategic for all forms of marketers. We’ve moved the conversation to ‘how do we make apps work for your business?’; it’s no longer just about installs, but about the actions people take and the proxies for value that should matter. We can now accommodate that in a campaign with UAC, where you can set a value for an in-app action and actually optimize towards that. We think that honing in on the actions that matter up front is the way forward.
Katie: Can you explain to us what UAC is?
Lee: Sure. Universal App Campaigns (or UAC) is our primary campaign type for driving installs or events that happen after the install. So Universal App Campaigns run across YouTube, AdMob, Search on Google.com, Search in the Android Playstore, and in display ads in the Android Playstore. On iOS we run these on AdMob and YouTube only, as well as the Google Search App.
You can optimize in three different ways, either via cost per install or cost per action, or targeted return on ad spend (ROAS). Once you set that objective, we will find users most likely to complete that action across all of those different properties. So one of the values of UAC is that we can enter what used to be five separate auctions all at once. It’s very powerful in that sense. Compared to the legacy campaigns we used to run, where you only ran search or YouTube, UAC runs 140 percent more conversions per dollar. So we are seeing a lot of efficiencies with UAC.
Katie: Wow, that is pretty incredible. A lot of the things you mentioned there are heavy on data. So what does Google think about Big Data and what kind of role does that play into what UAC has become?
Lee: There are a few ways to think about it. We’ve figured out that with machine learning we can make some better decisions and drive greater efficiency on our own tools, especially when you think about doing that across all of the channel auctions I just mentioned. When it comes to customer data, we also believe there’s a lot of information that you have as a marketer that’s better than what we have, and so UAC has changed some of the controls a little bit. One of the key inputs is actually which in-app events that you care about, how you value those events and set your bids accordingly. So the marketer data here is almost equally important in terms of driving outcomes and it’s the combination of the two where we match the actions you care about with our user base across the entire marketing funnel. It can get more nuanced depending on the type of business you have, but all you really need is ten in-app events per day, essentially, from unique users, to take advantage of that bidding.
We use this idea of one plus one equals three - your data plus our data is that much better, and we weren’t really able to do that before, so that’s what we’re most excited about with UAC.
Katie: I think it’s important to highlight that these ten users of an event or a specific event are all you need to really get set up, which really isn’t a lot. I have one more question about UAC; in terms of uniqueness, Google has a ton of inventory you can choose from, and a lot of our mutual clients have been out there trying to figure out where they’re performing best. In terms of UAC’s unique offering, what do you think that is?
Lee: The unique offering is that it’s dynamic and it will adjust in real time. So maybe you thought that search was your best channel. If something happens in search or a user stops searching for a certain keyword, UAC would adapt and look for users elsewhere. That’s the value in this; it’s dynamic in nature and reduces the lag time in optimization to real time; you don’t have to worry about some of the fluctuations that occur in user behavior.
An example I’ll give here is that my mother-in-law is the top-ranked Angry Birds player in the world, and she spends a lot on in-app purchases. She’s not your typical demographic. I won’t give away her age, but if you were doing just demographic targeting, you’d miss out on someone like her. So UAC means you can transcend demo targeting and really find those pockets of value.
Katie: So now we wanted to move into the really great case studies we’ve had with our mutual clients who have used UAC. Taking the plunge to lose control is obviously step one, because with UAC it’s a click of a button. So let’s take a quick look at Go Euro and see how they did it. In this example, they show us how automation can improve marketing in internationally-growing companies. Lee, maybe you can tell us more about how GoEuro used UAC?
Lee: Sure! GoEuro, as you said, is a travel booking engine. They had a couple different goals with UAC; first, they wanted to expand into new markets across Europe. They wanted to break into some new areas. Secondly, they realized that one of the benefits of their product is repeat usage, so they wanted to find users who were likely to book more than once, or book a plane ride and a hotel. That was the value they configured in their Universal App Campaigns. They wanted people who would download the app and purchase at least once in the first month. And by sharing that information with us, we were able to increase the number of valuable users and ultimately profit for them.
Katie: Definitely. I think we saw that with this campaign. GoEuro was able to expand into four new markets, which is pretty incredible. Typically this takes tons of time and research and figuring out localization strategies. So I’m curious - does UAC work better in particular regions, or certain demographics? Have you seen any disparities there?
Lee: Well, the only market we have a challenge in is China, because we’re not there domestically. The benefit of UAC is that it might look different depending on the market you’re in. Or the category! In gaming, we see people watching a ton of gaming-themed YouTube videos, so YouTube becomes a very different kind of asset for gaming companies looking to drive more users. So we can heavy-up on YouTube automatically.
Katie: Very cool. Moving on, we’ve got Viber. Obviously, Viber has a massive existing user base, so we wanted to focus on UAC and downstream engagement. So acquisition a few years ago was all about driving installs, more and more installs. But a lot of clients have switched their positions to not just finding new users, but the right users who will engage. In terms of that, how does UAC has an offering set for those clients who want to drive downstream engagements?
Lee: Using the Viber example, they told us they cared about people who were likely to complete a phone call, and it was even more valuable if they called the same people regularly. So they used Universal App Campaigns target CPA bidding strategy and were focused on people who had a higher retention rate. Overall, this drove installs with a 21 percent retention rate and a lower CPA. I think Adjust and Google work together well when we think about the right metric and price it accordingly. Sometimes it takes a little while to design the right event or the right bid for that, and that’s when you need to work closely with Adjust.
Katie: In terms of Adjust and how we set up our integration, we always recommend taking the top five things you’d want a user to do the week after the install and make events out of those that we would be able to track for you. We can send all of these events directly to Google, which would power your UAC campaigns, especially if you want to double up on some of those in-app actions. One thing I’m a little curious about is how you see UAC and remarketing. What are your ideas on what that will be?
Lee: Yes, that’s something we’re thinking about. Could you get the same efficiency if we were able to remarket across everything? We did just recently launch app dynamic remarketing, so for people who have a product catalogue feed, that’s available today. Without giving away too much, I can tell you we are thinking about similar types of campaign strategies for remarketing in apps. That could be a nice integration with UAC too…
Katie: That could be cool; we’d be very excited about that. So getting away from CPI and moving into CPA, which is what Viber used here, what do you recommend when it comes to bidding on target CPAs and what have you seen perform the best?
Lee: Good question! Back to what we talked about earlier, I have seen people make the choice to work with an event that’s too far down the funnel. That’s when it’s helpful to work with Adjust on sending not only multiple events so we can experiment, but finding the right balance of value and volume. The other thing that’s worked well is that a lot of customers have ROAS goals, so another area to work on with Adjust is on translating some of your BI goals into a bid for UAC. Those are more of the sophisticated things to think about it.
The final thing I’ll say is that if we’re optimizing for an event that happens after a download, this can take time. Make sure you allow for conversion delay and looking at your campaign results 14 or 30 days after the campaign has ended is a really good best practice. It’s another area to work with Adjust, to understand the pipeline of events and volume.