Kaya Taner is the CEO and Co-Founder of AppLift – the mobile games marketing network. Previously he was Head of Publisher Relations at HitFox, responsible for building relationships with online game publishers. This year he presented at App Promotion Summit in Berlin on the subject of “Best Practices to Deliver LTV: Insights on Generating ROI-Positive Traffic for Free-to-Play Mobile Games”. The presentation drew on the data and insights AppLift have generated through their marketing activities for a number of mobile games clients. The talk covered the following areas:
- Tracking, Both Outside (Install Attribution) And Within The App (Post-Install Event Tracking)
- Best Practices To Deliver ROI-Positive Traffic
- ROI-Positive Traffic Buying Through Customer Lifetime Value (LTV) Optimization
We’re now able to share this presentation with you in a number of formats including video, audio/ podcast and the full transcript.
Insights On Generating ROI-Positive Traffic For Free-To-Play Mobile Games Video
Insights On Generating ROI-Positive Traffic For Free-To-Play Mobile Games Audio/Podcast
Insights On Generating ROI-Positive Traffic For Free-To-Play Mobile Games Presentation
Insights On Generating ROI-Positive Traffic For Free-To-Play Mobile Games Transcript
Yeah, great to be here and to talk about best practices to deliver lifetime value today. I’m going to share some insights on generating ROI-positive traffic from our experience of running it for free-to-play mobile games, but actually many of these insights, if not all of them, can be applied to any sort of other apps when you do user acquisition.
So basically, briefly to the background, I’m CEO and co-founder of Applift. We are a mobile games marketing platform so we drive millions of installs per month for top game publishers on the one side, and then monetize the mobile traffic of media partners that drive these installs. And I want to share today the insights that we acquired over the last year-and-a-half or so by running install campaigns for 150 plus mobile games. And in this presentation today, I’m going focus actually on three major parts.
So first of all, I’m going to talk to you about the concept and the status quo of user acquisition in the industry. Where do we actually stand? Then I’m going to tell you a little bit about how you should approach traffic buying in general, if you wanted to buy ROI-positive traffic. And last but not least, I give you a little timeline for an app launch, of how an app launch should look like.
So let’s come to the first point, the status quo of the industry. Yes, so over the past years we have seen three main waves of user acquisition. If you look at it around 2011, when there was not such sophisticated install attribution tracking solutions set in place, we had many, many campaigns. Basically the goal of app developers was, yeah, I want to come up to the top of the charts and I’m ready to pay a lump sum fee of it. You would pay something like, I don’t know, 25,000 Euros to get a top ten positioning and get a lot of installs through that. There was no really, no SDKs, and the main traffic source was either through app discovery services or incentivized traffic.
Then with adoption of more sophisticated tracking solutions, when you could actually start really tracking the installs, the cost model changed to cost per install model, so the CPI model emerged, and then the goal of app developers started to be, okay, I want to have a high level of installs at the lowest possible CPI. And the trend was I integrate many SDKs and to either incent plus non-incent campaigns.
Since the start of this year, and it’s actually a movement that we are very much as well trying to co-shape, it’s towards ROI-positive traffic buying. So it’s not, you don’t just buy a high volume, but you look what is the lifetime value of the user and determine is this lifetime value that I’m buying bigger than the return, than the CPI, to make sure that the costs are lower than what you are actually getting from this user to buy ROI-positive traffic.
So the pricing model in this case is different. It’s not a CPI-based model, it’s a CPI or adjusted CPI model. To just give you a comparison, imagine in the middle you have a model where you have ten traffic sources and you just say okay, I’m paying two dollars. It doesn’t matter for this traffic. Whereas here, in this newer model you would examine each different traffic channel. And even we, at Applift for example, we go deeper and go within each traffic channel at each sub-ID level, and determine what is the quality of the user I’m buying, and what is a reasonable CPI to pay for each of these channels.
So they’re going to be very different. And so the advertising format, it’s moving more into non-intrusive advertising, something that I’m going to give you an example on later in the presentation. So this is the current status quo.
Then I’m going to focus now, of course, on the ROI-positive, on the last bit, and give you a deep dive on how to do ROI-positive traffic buying. So there is a majorly three steps involved. One is, of course, to buy ROI-positive traffic, I need to be able to track that traffic, and I need to be able to not only to track the quantity but also the quality of what I’m buying.
So I’m going to dive a little bit into the tracking set up. Second step, of course, before you buy traffic, you’ve heard it many times today I guess, is to do app store optimization, right? And last but not least, I’m going to focus on where to buy traffic and how to correctly optimize this traffic to make sure you’re buying and you’re doing it the right thing.
So in terms of tracking set up, we can see basically, I think you could draw a line in the middle of these two boxes and say the line is the install. So everything before the install to track the install itself is campaign tracking. So you have to make sure, okay, when I buy traffic, I need to know exactly if I get an install, where does, which channel did this traffic come from?
But in app tracking, what happens after the install actually? What sort of KPIs? I need to track retention rates, etcetera. And for these sort of in-app events, there is also some other players like Flurry, Google Analytics, and many more. The important thing is that you have a common identifier that link the two, so that you know, okay, which user drives what kind of quality inside of my app. And know also where this user actually came from.
So, let’s focus on the in-game tracking. What kind of KPIs do I need to have a look at to make sure I do ROI-positive traffic buying? Basically, the KPIs say you can sum them up in four blocks. So first you have retention KPIs. These are day one retention rates, day 7, day 28, day 30, whatever you choose. These are the most common ones. Then you have the number of daily active users or the number of daily new users that you’re getting.
Then you should track engagement KPIs. For example, the average time a user spends inside of your app when he opens the app, or each session basically. Or like total daily sessions divided by the daily active users. Then you have the virality factor, which is the K factor. It’s the viral growth rate per user. So if you buy one user, and he invites another user, you get actually two users, right? So the virality you should measure is, one is actually trackable. It’s, for example, Facebook shares, if you invite other users, etcetera, and one is non-trackable, if you are at the top of the charts you will get many more organic downloads, which you cannot directly pinpoint to a specific user that was causing that virality.
And then last but not least, and most importantly for the survival or for the monetization of the app, are the monetization KPIs. Typical KPIs here, average revenue per user or average revenue per paying user, or the monetization rate, which is simply the percentage of users, the overall users that do in-app purchases.
So important here, if you want to do ROI-positive traffic buying, actually the monetization KPIs are the most important, however most of you will know that monetization, that there is a time lag between the moment when I buy a user and the moment he pays inside of my game, my app, whatever applies to you. And this time lag, you can’t actually wait. If I have to wait two weeks or a week to know, is that user spending money, I am already buying a lot of traffic, so we need early proxies. We need early indicators.
If I take the day one retention rate, this will already and see if one traffic cohort where I’m buying from out performs others, I will already have an early indicator that it’s likely that this traffic cohort will have a better quality also in terms of monetization than another one. So it’s important to focus on a few early proxies to make sure you track those.
Then if I just reassemble these in terms of the user life cycle, so you see here the user comes in, I buy the user, you’ve got the earlier proxies like retention engagement, which then leads to monetization and the virality factor where users refer other users, and therefore new users coming in.
So this is the tricks that you should track, and the tracking set up in general. App store optimization, yeah, I wanted to go a little deeper here but I think it doesn’t make sense, so I changed it because there is a lot of presentations already here. For those of you who are interested, come talk to us. We also have an e-book on our website where you can learn more about how to do the app store optimization right.
What I want to mention here is before you spend money, make sure the app store appearance and the key words, etcetera, is correct, because you don’t want to spend a large amount of money and then your images are not nice, or people just, they get out of the funnel because of your app store appearance. I think this is important.
Then last but not least, how do I buy traffic and how should I optimize it? The traffic-buying part. So, this is very important to first understand what are the right traffic sources. I mean, you have in-app traffic, you have for games or other apps as well, in-game traffic, and then you have mobile web, or there might even be other channels. Like, for example, Wooga had a campaign, offline campaign, in South Korea in the subway, etcetera, with a QR code. These might all be valid channels depending on what you want to achieve.
After you find the right user groups, you should address the users appropriately. What are the best performing ways to reach this audience? In this case, what you should look at is, you should buy performance based, of course always on a CPI or a CPA basis. And you should optimize on an ECPI basis. Effective CPI is something that I will come to in a minute. For those of you who don’t know what I mean, we’ll look into that in a second.
Then also consider definitely non-incentivized buying. Incentivized traffic, while you shoot up in the app store might make sense for a selected amount of apps with mass appeal, if you get a high number of organic installs. But generally if you have an app that has not such a mass appeal or so you will just spend money and it’s forlorn, because the users that you buy then organically won’t drive you any value.
And then third, insure that your app is presented in the best possible way. So, there is different sorts of advertising. I have an example here. This is something we did for a partner. Here you have a, this is a traffic partner of ours. It’s a French celebrity app. And then on the left you see a channel jeux, it’s games in French. And actually, when the user clicks on there, he sees those gifts. And this is already the advertising. But it’s so tailored to the app itself that it doesn’t come across as advertising. And then you get this gifts screen. You can scratch it open with your finger, and underneath only is the advertising for the app.So this guy is basically getting his app presented in a very non-intrusive manner, which means that the funnel and the quality of users he will get is much better than what he would get from a banner advertising for example, and also much less spammier, which will give him an additional brand effect. And then on the other side also for this publisher, this worked very well. He achieved ECPMs of 33 US dollars which then also is, of course, very nice for the ones who are advertising your apps inside of their apps.
So, what is important, we talked about it in the beginning, is for ROI-positive traffic, is that the customer lifetime value is actually bigger than the cost of acquisition you have. However, the CPI itself may not adequately reflect your real costs. But the eCPI will. And the eCPI basically, it takes into account all costs induced by the acquisition. There is certain factors that drive the eCPI up, and certain factors that will drive the eCPI down. I will give an example of each.
So for things that drive the CPI up, for example if you do a free promotion. You give away five dollars in virtual currency with the acquisition. Imagine that. So there is a cost attached these five dollars that you give away for free. So theoretically, that drives your CPI that you’re paying up. On the other side, and maybe more importantly, is the K factor.
So, let me explain that through a simple example. Let’s say you’re paying two dollars for a user, or for your user cohort on average, for each user you’re paying two dollars. Each of these users spends, let’s say, $1.50 inside of your app over his lifetime. So two dollars is bigger than $1.50. In this case the CPI would be bigger than the CLV, so it’s not ROI-positive.
Now imagine that this particular user cohort is very, very viral. Imagine they have a K factor virality of 50 percent, meaning that every user that you buy will bring you half a new user. So, if you take that into account, actually, you would be paying two dollars not for one user, but for 1.5 users. So if you take two divided by 1.5, the effective CPI you’re paying is $1.33, which is lower than the $1.50 that you’re making. So in this case, you would be ROI-positive through an effective CPI.
Coming to an end, then the left side of the equation, the CLV, what is the customer lifetime value mathematical formula looks more complex than it is. You just take the retention rate of this user cohort you buy so, how many people come back on day one, how much do they spend on day one, plus how many people come back on day two times how much they spend, etcetera, etcetera, etcetera, etcetera. You sum them up and then you have the CLV.
That’s so much for ROI-positive traffic buying. Now what this means in terms of a timeline for your app, basically four weeks before submission of an app, you should have all install tracking solutions integrated. Remember campaign tracking as well as in-game tracking solutions or in-app tracking solutions.
Then roughly two weeks before launch, you should set up the campaigns with the traffic partners you’re going to work with. You should prepare the creator, set up the tracking, and basically when the game is live, you should optimize the different traffic channels and sources based on ECPI and estimate the CLV and see is the CPI bigger or smaller, or the ECPI bigger or smaller than the estimated customer lifetime value from this, and according to that adjust how much you’re willing to pay for each traffic channel.
Just as a backup slight, maybe we have here, like for example a table where you can see this is a, where we did this. You see a CPI. You see a K factor, which is the virality factor. Then you see the ECPI. We got the estimated CLVs of these from the partners we were working with in this case. And then the red ones are the ones where the CLS is lower than the ECPI , meaning it’s non-ROI-positive. And the green fields are ROI-positive, so we would scale them up.
Thank you very much.
Thanks to Kaya for doing such an interesting talk. You can find out more about App Promotion Summit here