With the roll-out of Universal App Campaigns (UAC) on the Google Network, the routine of micro-managing ad accounts has taken its first gut punch. With the age of self-driving cars and fully autonomous journalists on the horizon, it was only inevitable for Google to give the job of actually making the ads to the robots.
Google has been pushing automation for years, so this encroachment should come as no surprise. And to no one else’s surprise, the efforts are working. According to Google, UACs are showing an average of 140% more conversions per dollar than other Google app promotion efforts. The machines aren’t just doing the job, they’re doing it better.
So where does this put us? The online advertisers? What efforts can we make for our clients if the job is being done for us, with no way to know how or why? How do we continue to push the envelope and make our clients happy (and money)?
Despite the amazing things Google’s algorithms can do, there are still things we can do to improve it. If you are unfamiliar with the basics, there are great materials out there, including these articles from Supermetrics and PPC Hero. If you are ready to start diving in, here are 3 ways to optimize your UACs:
1) Test Your Variables
While you must give the wheel to Google for most of your campaign adjustments, you can still step in and draw some insights from your data. One of the ways Google allows for manual adjustments is through their Ad Assets window. Once you input your content, and after enough data has been accrued, you will get a diagnosis of the top performing assets.
To gather comparative data, we recommend at least 3 pieces of content for analysis. You are given room to input 4 text descriptions, 20 images, 20 videos, and 20 rich media pieces (HTML 5, GIFS, etc.). You can use these pieces to see how they compare to each other. The catch? You don’t know which assets are being used in tandem or have the ability to break out individual content for testing… unless you make a new campaign for each new piece of content (We don’t recommend that).
The key is to limit your variables and test. You need a control group. If you have 20 individual images with differing copy or colors, how will you know what people are responding to? Is it the messaging? The cool image? What is a good conversion rate?
Text boxes are easy, use all 4 slots and see what works. For images and videos, try using the same content in different sizes. Here are the recommended image dimensions from Google:
Use one asset in all sizes and test to see which is converting better. Remember, some of the sites across the Google Network only accept one size format. Not having a certain dimension could mean missing an entire market! For video, Google recommends 15-30 second clips. See for yourself and play around with different play times. Once you have an idea of what is doing well, incorporate new content and set new benchmarks.
2) Change One Thing at a Time
As we discussed earlier, limiting your variables will give you a better feel for how your content is performing in the big picture. Once you have a nice mix of content, rotating poor performers out of the equation should yield better results over time. The urge will be to remove all the low performers quickly, but we would caution against it. One asset at a time.
As we said, we don’t have relationship insight into how the content is being paired. Certain text descriptions may be a top contributing factor to an image. Limit your content rotation. Google recommends waiting for 100 conversions to gather enough data, so if you follow this formula and have a lot of conversions and a ton of content, you may be doing this a few times a week.
For us, we find that doing our rotations once a week is a happy medium. We can usually get a few new images a month and the text descriptions are easily changed. Set a day each week to check in and make your adjustments.
One important piece to remember is to document all your performance data and the asset it is associated with. Once you remove a piece of content from UAC, you will no longer have access to that data. You will want to see how your new content is comparing to its previous version, so keeping a running document on each change is necessary to optimization.
3) Give it Space
At the end of the day, the algorithm will work itself out.
There will likely be pressure from your client or yourself to micromanage — try to step back. The algorithm takes time to pull in data and will learn how to adjust your bidding and content. If it isn’t performing well one day, see if the data normalizes over time. In the same vein, if you are seeing spikes in conversions don’t immediately up your budget. There will be changes out of your control, avoid seeing the forest for the trees.
Despite UACs not having a true spending cap (some of us don’t have unlimited budgets Google!), you still shouldn’t need to overreact. Google can now spend double your daily budget during peak traffic, but will never exceed a monthly limit (your daily budget x 30.4). Frantic adjustments to high spend days won’t change your monthly cost, so let the program do all your heavy lifting.
Slow and steady wins the race with Universal App Campaigns. With its “set and forget” functionality and the ability to continuously update, it is a powerful tool to utilize for app advertising. A benefit of being self-sustaining, UACs can be very forgiving. Don’t be afraid to try new things and see the results. If it’s bad, it will get phased out. Maybe you will learn something from our new robot overlords.