Take it from an Expert: Top Advertisers Offer Tips & Insights on The AppsFlyer Performance Index
For almost three years, The AppsFlyer Performance Index has been the industry standard report card on the performance of mobile media channels (want the inside story behind it? Read it here). We’re excited to share that Edition V has already been downloaded thousands of times in just over two weeks. But among this large group of advertisers, each has a different take on the index, with different uses, and different actionable insights applied.
To get the advertiser perspective on the index, we talked with three top UA pros: Fabien-Pierre Nicolas, Head of Growth / Marketing (US) at SmartNews (25+ million readers in 100+ countries), Thomas Petit, Growth Team Member at 8fit (10+ million users), and last but not least Saikala Sultanova, Director of User Acquisition at Ubisoft (massive hits like Rayman, Just Dance, and Assassin’s Creed, to name just a few). The three offer their insights on the index and the mobile media landscape as a whole — with tips on what has been working well for them.
Expert advice, and lots of it, straight ahead!
AppsFlyer: In what way is the Index valuable to you and to the industry?
Fabien: What motivates me to always jump on AppsFlyer’s Index as soon as they are out is that they provide a clear incentive for the ecosystem to get better through the “Power Ranking”, which factors not only volume but also value in the form of retention and average sessions per user — key elements in the SmartNews business model which is centered on engagement given our ad-driven model.
Another key value in the report is insight on markets your app is not yet present at, and a starting point for selecting media partners with which to launch your UA activity in a specific market. For example, we’re considering testing Android in India and thanks to the Index, I will go beyond Facebook and Google, and include Twitter and AppLovin as well.
This is especially useful if you are considering some of Asia’s top markets like Japan, India or Indonesia, where you will need to use local key players. In Japan, SmartNews and Yahoo Japan! are the #3 and #4 publishers for non-gaming iOS — this will enable you to expand beyond Twitter and Facebook far more efficiently than working with some of the Western video networks for example.
When it comes to your more mature markets, the index is a great for a sanity check and ensuring that you’re not missing out on a successful channel. Due to my gaming roots and multiple tests, I thought Twitter could not deliver significant value, but when I realized that their non-gaming performance was solid in the index, I included them in Smartnews media plan. It’s currently one of our top channels from a quality perspective and provides solid volume. Thank you AppsFlyer! 🙂
Thomas: First of all, the ROI index! Every app out there has different goals, different way of monetizing, etc. so it’s hard to compare with peers, especially considering the sensitivity of revenue and cost data. Therefore, it’s a great move from AppsFlyer to uncover a part of this critical yet largely ‘uncommented on’ datapoint. Kudos!
That said, I couldn’t help noticing that the ROI index was only able to use data from 1/6 of all apps (800/5500). It seems most developers don’t record cost and revenue properly. I want to believe more actually measure their ROI.
Second, segmented data! Many benchmarks only share global data without much info on OS, geo and category. In this case, anyone can relate to something closer to their business environment. Wishlist for next year: add specific categories, and split also the ROI index 🙂
Saikala: I would recommend using the power ranking – it offers significant value. The combination of quality and quantity is really important as far as what advertisers need to focus on. Having said that, I would also encourage everyone to test for yourself to prove whether it works for your app.
AF: What do you think of Facebook and Google’s domination?
Fabien: As a performance marketer and mobile marketing leader, I like to keep my mix as diversified as possible and not be dependent on any specific channel. This enables me to be in a situation where arbitrages are possible. I don’t have a specific issue with Facebook and Google’s domination as long as they stay “customer-centric” by providing us with the right outcome to our campaigns.
To be more specific, I love how Facebook innovates in their platforms — offering A/B testing of creative/targeting for example — but I dislike the lack of transparency when it comes to Facebook Audience Network (FAN) targeting or splitting FAN campaigns from Facebook/Instagram campaigns. And when it comes to Google, their programmatic creatives and level of support are terrific, but their reversal of placement level transparency following the mandatory switch to UAC is disappointing. I hope it’s temporary.
Thomas: Facebook and Google are well ahead and this doesn’t come as a surprise, they are said to own about 2/3 of US ad spend by now, and most growth. It’s just as you commented: “The two giants are dominating the mobile ad space like never before”. It would be enlightening to see how the Facebook feed, Instagram and the FAN compare. And next year how the aggregate of Search, Youtube, Display compares with all UAC. I can’t wait to see volume vs retention vs ROI on this “new” channel.
Talking about direct sources, the picture is slightly different for Apple due to OS/geo limitations, but even more so for Twitter, which seemed to do very well for non-gaming apps but not that well for gaming.
Saikala: Well , I think they are dominating for sure. However, Facebook is dominating on all fronts including e-Commerce, Gaming and everything else. They made UA on Facebook approachable and effective, offering a really good knowledge base to educate advertisers. Facebook can be a handful to manage though, as it has a high ad fatigue rate when it comes to creatives. It is still one of the top quality networks for most of us in the mobile gaming vertical.
Google on the other hand made it easy with UAC, but it currently seems to work best for e-Commerce, retail and other similar verticals. For gaming apps it can still be challenging to make it work outside the casual games genre, our own activity now is still work in progress. Wishlist for Google UAC: give us leavers to control volume of ad spend that we want to allocate to each placement i.e. UAC campaign = XXX USD (Search 60% + YouTube 20% + Display 10%) or whatever the advertiser decides :).
AF: What about the performance of video ad networks? How important are they in your mix?
Fabien: We love video ad networks, they are often the 2nd or 3rd best type of ads in quality after search and social. Whether it’s AdColony, Vungle, AppLovin or Unity Ads, I have worked with them since 2012-2013 and always saw a solid quality, for games and other apps as well. Video is working well for SmartNews on both Android and iOS.
My only regret is often the lack of targeting options by age / gender as well as the relative lack of investment in their tech platform among some of these networks. For example, if you cannot run more than one creative per campaign, it becomes very inefficient to optimize (for the ad publisher and advertiser alike).
Saikala: Video networks are solid, they can deliver volume and quality at scale with some investment as far as UA operations are concerned. Unity Ads, AppLovin, AdColony and Vungle usually work well for gaming apps across all kinds of genres. Although AdColony did really well in the index, for us it’s a bit different: we actually see performance that’s not on par with the past. We believe this changed when they launched their new ads for brands.
I would say that after Facebook and Google, video networks would be the next in priority for most games in terms of budget allocation. Sometimes for casual games, video networks are even more important in volume than Facebook & Google.
AF: What’s your take on the success of Apple Search Ads?
Fabien: Our account team at Apple Search Ads is great to work with so I’m happy they’re doing well. One point to make about this product is to ensure that all UA managers have the same understanding of the space: if you’re buying your own keyword, you are essentially paying for your organic users. If enough people are doing it at scale, it could lead to a bias in the blended quality AppsFlyer reports for this channel. And it’s combined with iOS 11 which gives about 35% to 50% of the search screen to Apple Search Ads placement = a powerful incentive to spend.
Thomas: The new kid on the block made its way up to the #1 position in the ROI index.While this is very impressive, I was actually a bit surprised, not at the high ARPU (+30%) but with the CPI datapoint (40% below market average). In my experience, Search Ads inventory has become very competitive to catch now, I wish it was the case in my vertical 🙂
Another very interesting fact is that Search Ads are extremely impressive in the Power Ranking, but less so on the volume index, which signals a much better retention and session score. Only in North America did Apple make it to the podium on volume. Let’s see how this evolves as Apple opens up new geos…
Saikala: It’s obvious that Apple Search Ads are in an advantageous position of serving ads in their own store 🙂 I think it was a smart move from Apple to launch search ads with the “Search Match” option, which brings ease of use, identifying new keywords and ease of spending. I would love to get more volume from those ads but I assume it’s only a matter of time when Apple opens more and more countries for Search Ads.
To Fabien’s point on paying for organics, while bidding on your own brand/app name. I guess this is a lot to do with how you think about it. Some gaming companies practice “brand retention” techniques. For example: you have an awesome app delivering amazing ROI, this is when bidding on your own brand/app name can be considered as retention. You could block that ad space from your competitors bidding on it for example ;).
AF: How do you mix local vs. global networks/media sources?
Fabien: Our global team here in San Francisco does not market outside of countries like the US, UK and India, where Western channels have a great reach. But when it comes to Japan, you can see in AppsFlyer’s Index that local channels are needed — for example Nend, Yahoo Japan!, LINE Games or… SmartNews (#3 right after Facebook and Twitter in the iOS non-gaming power ranking). The same goes for China and Southeast Asian countries.
Saikala: Local networks are mostly relevant in Russia, Japan or China (mainland). The main reason UA experts use local networks is due to a lack of sufficient inventory on Western networks.
Russia is normally better on Instagram, but it is important to test Target.my.com inventory (under the mail.ru umbrella which owns all 15 Russian speaking social networks, be it VK, my world, classmates, etc. These are used by most Russian speakers.
In Japan, media partners such as Nend, iMobile, Zucks, Five, Tap One, Adways, Line, Yahoo and Smart-C are an important part of the mix. However, Twitter is their Facebook and it is the biggest and best in Japan. I have learned that it’s all about Japanese culture and the importance of anonymity when it comes to using social networks. Japanese are more open to using Twitter because they prefer not to use their real names.
AF: Do you see the media space consolidating?
Fabien: While we’ve seen some consolidation, we mostly filter out bad players because of high rate of fraud or a lack of sufficient transparency and solid tools. The right balance seems to be around 10-15 channels per platform, giving you access to about 90% of mobile users in our markets (with some level of duplication in terms of format or category).
Saikala: Definitely there is room for consolidation, at least in the gaming vertical. We used to run with several dozen networks, but now we use fewer partners, depending on our budgets of course. The consolidation itself is mainly trending towards the quality networks. I would also cluster and consolidate networks by bidding tech and creative i.e. self-service, video, playable, affiliate and DSPs.
With the aim to consolidate better, it would be really useful to see networks become more transparent and send advertisers the app names, cost per app name and cost per device id to name a few from our wish list. I know a handful of networks that already offer this, which is great. We hope more will follow.
Thomas: When looking outside of the Facebook/Google/Apple triumvirate, I see about 20 networks concentrating the volume, and even less when it comes to games. There are hundreds of companies in this space, but as time passes, it seems the market is getting more concentrated, as Eric Seufert predicted last year. I would even argue that among the top rankings, we’re starting to see the same main few actors, mostly video ad networks. I believe the competition between the likes of AdColony, Vungle, AppLovin, Unity or Chartboost will get very interesting to watch.
Overall the AppsFlyer index is a fantastic source of insights for app marketers and decision-makers, surfacing the key trends in mobile media, and it’s certainly one to watch over time to see how the battle to get advertiser dollars evolves.