This article was originally published on Apptweak’s blog and can be found here
Although the concept of customer lifetime value (LTV) for mobile users has gained a lot of attention in the past couple of years, there is still a lot of confusion and misunderstandings around both LTV’s definition and how it should be used for mobile campaigns. Let’s try and debunk the myths, set expectations straight and see how LTV can be effectively leveraged for mobile advertising.
Let’s start off with the reassuring feeling that a majority of players within the mobile industry are leveraging LTV for their user acquisition efforts, are thinking of doing it, or at the very least have heard of the term. This is partly due to a general shift in mentalities and expectations in the industry towards ROI-positive media buying, as we have now completely entered the third wave of mobile marketing. Mobile marketers now understand the value of having insights on what’s happening after the install, all along the user lifecycle, and how it impacts overall ROI down the line.
Customer Lifetime Value Is Not About the Value
What is lifetime value? It is a measure of revenue per user, just as ARPU, with the major difference that ARPU is calculated over a specific period of time – say one month – while LTV spans over the entire lifetime of users, which makes it extremely cohort-dependent.
Lifetime value varies greatly across users, from those who will only open the app once to your heaviest users. For free-to-play mobile games, a 5% monetisation rate, meaning that only 5 players out of 100 spend money on virtual items, is generally considered excellent. Even for the users who do eventually monetize, spending behaviors can be very different. For instance, some users will only start spending money one month into using the app.
You can find many resources online on how to calculate lifetime value. There are basically two ways to compute it, using functions with aggregates of monetization and retention metrics. In the case of retention, its conceptual equation is rather simple:
The issue with getting LTV through this equation is that, for the purpose of mobile advertising, it is utterly useless.
LTV is historical value which can only be obtained backwards, which in turn means that it is impossible as such to compute the actual lifetime value of newly-acquired users.
However, when running mobile advertising campaigns, you quickly need to determine the quality of cohorts of users (depending on the granularity of your tracking) in order to tweak and optimize your budget allocation across different channels. You can use LTV to assess how much you can spend on one user, but can’t rely on it as a live metric to guide your acquisition efforts.
This is where the main confusion comes from: lifetime value optimization in mobile advertising is not about the value itself.
How Can LTV Help Mobile App Advertisers?
How can LTV then be leveraged to optimize mobile campaigns if it can’t be calculated as such? It can be approximated by tracking and analyzing specific user events happening shortly after the install. These events are usually calledproxies and can be sorted along the usual categories of the user lifecycle:
- Monetization events. Ex: first purchase, amount of first purchase
- Virality events: Ex: number of Facebook invites sent. Virality boosts organic downloads, which in turn increases overall campaign ROI.
- Engagement events. Ex: tutorial completion, level 10 completion
- Retention event. Ex: day-2 retention.
Some solutions offer to approximate the actual face value of LTV for new users through feeding these proxies to machine-learning predictive algorithms. However, for the purposes of optimizing advertising campaigns, no need to go that far. Analyzing these proxies for each acquisition channel relatively to each other is sufficient to determine which channels perform better and then adapt your spend accordingly.
The main challenge for you, as an app advertiser, is to determine which proxies are most relevant for your app. The hardest part is managing the trade-off between the strength by a particular proxy and the amount of time before it becomes relevant. For instance, the first-purchase is an extremely strong signal of the likelihood that a user will spend a significant amount of money in the app. However it can take a while until it actually happens, decreasing its utility for campaign optimization. Conversely, day-1 retention may be a rapid indicator, but it’s a rather weak one.
At AppLift we see a great diversity of proxies across apps, games, categories and genres. We advise our advertisers to think about which proxies to use as early as possible in the conception of the app, as it will impact their entire acquisition strategy.
Which proxies are you using to optimize your advertising campaigns?