Lifetime Value - The attempt to estimate the future
What is it?
The Lifetime Value (LTV) is the revenue a gamer is expected to generate throughout their lifespan, the time from the first to the last login. It is a key driver to marketing budget, proof of the power of customer loyalty and a great tool for forecasting growth. When talking about mobile app metrics, the LTV will always be discussed. Why is it so important? How does one calculate it? In this insight article we’ll go about answering these questions.
Why is it important?
LTV is often referred to as the metric to determine the success of an app or game. The LTV puts a cap on the marketing budget by providing a maximum price to spend on User Acquisition (UA). The Cost per Install (CPI) or Customer Acquisition Cost (CAC) are marketing expenses to acquire new users or players. If those players will bring in less revenue than they cost to acquire the venture will not be profitable: your studio will be making a loss. LTV minus CPI thus gives the profit margin. Here, it should be pointed out that LTV is a projection, an expectation while marketing costs tend to be present, real costs.
How to calculate it?
LTV is a function of two factors:
retention and monetisation
LTV = (1/churn) x Average Revenue Per User
The retention explains how often (or for how long) the average user will return to play the game, thus generating revenue for the developer. Monetisation is the strategy of revenue generation. The better the monetisation strategy, the more revenue the user will generate while playing. To improve LTV one would have to increase retention (get players to play longer/ remain more days) or improve the monetisation strategy (get more money out of playing time).
LTV is often seen as an equivalent to revenue, but this is wrong. LTV is a projection of future revenues. It projects expected revenues of an average user once they become active (download the game and open the app). Hence, the LTV could also be considered a function of probabilities, with the probability of a user to stop playing after a given number of days (which gives us the expected duration per user) and the probability of said user to spend money on the app on a given day (which gives us the expected revenue per user per day). Those two factors multiplied give us the expected revenue generated per user, the lifetime value.
Google stresses the importance of including all revenues (in-app purchases as well as ads) into this calculation which means CPM fluctuations must be considered as well. Further, the time frame should also be considered in the projection. The longer the time frame for the calculation the less accurate it will be. So what length should be chosen? Jacqueline Zenn of GameAnalytics suggests a length of 180 days as a good rule of thumb. Too short of a time period will not allow for the effect of a marketing campaign to take place, too long becomes unpredictable. So really, LTV is the expected revenue generated by a user by a given day. It might therefore be preferable to give a range, rather than an absolute value when talking about LTV or to give a confidence interval. However, while LTV is a good measure for the health of an app, a high LTV is no guarantee for success.
How does LTV as a marketing metric work?
Since LTV is central to any marketing strategy, it might be a good idea to approach it as one would approach a marketing analysis. By segmenting users into different groups (location, age group, platform) the most profitable of the segments can be identified and the marketing strategy targeted accordingly. Here, it is also important to differentiate between the different groups when it comes to CPI. Acquiring users in the US might be cheaper than in Japan but the LTV of players in Japan might be higher. This is where the Return on Investment (ROI) comes in: the difference between LTV and CPI. It is this difference, the profit margin, studios should focus on when developing a marketing strategy.
Just as with the LTV, CPI should be seen as a marketing metric. While we could derive the CPI from adding all User Acquisition expenses and dividing them by the number of new users, the resulting number tells us (too) little about the actual UA expenses. Rather, just as we did in LTV, CPI can be divided into fractions and evaluated to see where money is most effectively being spent (and on what audiences). Values without context, be it LTV, CPI, etc. are of little use.
How does LTV fit into a growth strategy?
As Eric Seufert explains in his article a central problem of growth strategy is that marketing costs are “born upfront” while the revenues generated per user slowly accumulate over a given time frame. Seufert uses the example of different cohorts, stacked together to give the revenue cash flow. On day one the studio is likely to make a loss which is no problem for larger studios, but difficult to maintain for small indies. Those will be targeting early break-even points, which may hinder the overall growth of their project. The LTV gives an absolute maximum that should not be crossed if a studio hopes to gain from an app. Most studios remain below that threshold, depending on how long they can wait for the return on their investment.
What are the problems with LTV?
If LTV sounds too good to be true at the moment, that’s probably because it is. While the concept of LTV is fairly simple to grasp (and an approximation easy enough to calculate) getting more exact measurements is difficult. Firstly, enough data needs to be collected to construct any reliable model. The data in the first three months may differ very much from data in the months 6-9, despite being the same time frame. Internal and external changes can make historical data unreliable in a fast-moving environment. Secondly, as the game evolves (improvements are made) the user base might evolve as well. The collected data thus offers misleading information. Lastly, Cost per Install (CPI) is hardly the only cost a company will face.
In response to facing all these problems (and some more) Jon Radoff (Games of Thrones Ascent, Star Trek Timelines) came up with a different way of calculating the LTV:
ARPU: Average revenue per user
CAC: Customer Acquisition Cost
Costs: Costs incurred servicing the customer (hosting, serves)
WACC: Weighted average cost of capital (time value of money)
In his equation, maintenance costs are immediately subtracted from the revenue generated by a user, then discounted by a factor to account for the discount in value of money over time. The resulting number must be greater than the CPI for the studio to make profit. Accounting for Cost of Capital in LTV calculations is a controversial topic but the general consensus is this: it depends on the period in which the flow will be active but, typically, this aspect can be neglected. Since mobile game apps often operate in shorter time frames (financially speaking) of maybe 180 days, the discounted time value of money will be so insignificant that it is not worth the added complications to the process. It is good to keep the time value of money in the back of the mind but there’s no need to over complicate matters.
What to takeaway?
Depending on the business model, type of game, internal resources, and availability of data one or another method of calculation may be preferable. For a small studio an approximation should be sufficient since the cost of calculating the exact LTV of different cohorts would be higher than the profit earned with better projections. LTV remains one of the most important metrics when working with mobile apps but by itself it’s not meaningful.
