Testing New Networks for Your UA Campaigns: Simple Guidelines
Historically, many UA campaigns have relied heavily on social media to discover new users, and for good reason: Millions of people spend a lot of time on sites like Facebook. To acquire users, you have to catch them in the places they actually go.
But putting your entire UA budget in Facebook advertising may not be the wisest plan — it’s that whole “putting all your eggs in one basket” scenario. Social media is getting expensive, overly competitive, and less effective. Anyone running UA campaigns on Facebook — or any single network — should also consider testing other networks to find additional scale opportunities and to continue their growth.
So how do you test new ad networks and what do you need to test? Here’s a quick list of Do’s and Don’ts to help you sort it all out. (Note that they’ll vary slightly depending upon the kind of app you’re marketing.)
DO enter testing with an open mind. Every network and every partner is different. All of them want you to succeed and to continue to invest your budget with them. Listen to their recommendations, even if their approach to testing is very different from what you’re used to (or what you were expecting). Everything from CPIs to ROAS to number of installs can vary dramatically from network to network, but you won’t know what works or doesn’t until you test. So set aside a specific budget for testing, and enter each test without prejudice.
DO be transparent about your expectations with your new network partner. If you’re seeing a 30% return on ad spend (ROAS) over seven days with Facebook, share this information with them, so they can level-set with you. If their network is small, they may never be able to meet that kind of return, and they’ll be honest about that. Keep your expectations reasonable and try to remain unbiased and flexible. You may not be able to match Facebook’s return, but it doesn’t mean you won’t have successful campaigns.
DO allow enough budget to test adequately. Every network optimizes differently. Some algorithms require 500 users to convert to a certain event before they’re optimized, others may need 2000. While 2000 conversions may take longer and cost more, it’s likely to result in even better optimization. This means that you’ll need to allocate different budgets and different time periods for the networks you’re testing.
DO be patient with testing — and recognize that different apps require different tests. Conversions for different types of apps take different lengths of time. For example, if you’re promoting a chat app, conversions occur higher in the funnel, and you’ll collect data very quickly. And you’ll know relatively quickly whether your test is a success.
If the new ad network isn’t optimizing within six weeks or so, then there’s a problem. The ad net has the data so optimization should quickly follow.
However, a banking app might be optimizing toward a first deposit, so it will take longer to achieve that conversion and collect the necessary data. It may be that only two percent of users actually make a deposit, and that could take seven days. So not only will conversion take longer, you’ll have a smaller set of users to optimize against.
So it’s not necessarily about monitoring a test for a particular length of time, it’s really about reaching critical mass on your conversions so the network can optimize and work at its best. That could take one week or it could take six.
DO test different creatives and formats. If you have ads in various formats and styles, upload them all for testing. Most networks use AI and algorithms, and these will help determine which creatives perform best — and show them more frequently than lower-performing options.
DON’T set unreasonable expectations or make impossible demands. If your campaigns are driving 30% returns on Facebook, don’t expect a 40% return from your new partner. It’s possible and it would be nice, but it’s not fair to expect that kind of success — especially right out of the gate. By the same token, keep your CPI targets reasonable, as well.
DON’T expect results for free. You have to make an investment to see results. Don’t expect great returns if you’re only budgeting $100 per day for your tests.
DON’T continue tests that aren’t successful. If after a few weeks or months — or whatever time period you’ve defined — you’re not able to scale, then it’s time to take a careful look at the network you’re evaluating. If you’ve given transparent information and a competitive rate and your campaign still hasn’t grown, then the network may not be a fit. There are many levers you can adjust in a campaign, but scale is not one of them.
In some situations, you may achieve scale, but you’re not hitting your performance goals. Let’s assume you’ve allocated an appropriate test budget and gotten to the optimized campaign stage, and you’ve allowed the optimized campaign to run for the period of time recommended by the network. But you’re still not hitting your targets. In this situation, you need to have a frank and open conversation with your new partners. They will almost certainly try to convince you to stay with them. Think long and hard before you invest another dollar.
DON’T expect your tests to be ROI-positive. When you are testing new networks, make sure you set clear goals with your partners and go in with the understanding that your tests will not be ROI-positive. It’s not that they won’t ever be ROI-positive, it’s just that it takes more time to achieve that goal. So don’t panic if your initial tests don’t get there — there are other important metrics to consider early on. Profit shouldn’t be one of them yet. At this stage, the primary goal is to learn.
Again, the tests you run and the way you run them will depend heavily on the type of app you’re marketing, but these rules apply in most scenarios. The most important things to remember are to be completely open, honest, and transparent with your new partners, and to enter every test without prejudice. Think like a scientist: Keep your mind open as you monitor your results. You may be surprised by which networks perform best for you. Hopefully, you’ll continue to test and grow your marketing programs across new networks and channels and find plenty of new users along the way.