Predictive analytics for your mobile app

Revenue forecasting is one of the most important problems that analysts and marketers solve to optimize the app strategy. The task can be formulated in many different ways: “How much money will we earn by the end of the year?”, “Which feature (A or B) will bring us more profit?”, “Does it make a commercial sense to enter a new market?”, and so on. No matter what the question is, it is always a look into the future, therefore, some inaccuracies may appear.

In this article Vasiliy Sabirov, Lead Analyst at devtodev, shares 5 proven tips to make accurate revenue forecasts.

Method 1. Mathematics only

There are some mathematical tools that allow to solve the forecasting task.


If you haven’t worked with trends and seasonality before, it might be worth reading this article first. I’d like just to add a few practical tips to it.

Polynomial trends can be tricky. They are very good – sometimes even better than all the other methods – at repeating available data. However, when it comes to forecasting, they tend to be quite stormy. Depending on the degree of the polynomial, the tail of the graph (the forecast itself) can bend in one direction or another. The higher the degree, the higher the flexibility of the graph and the probability that it will bend in the wrong direction.

The easiest way to understand the dynamics is… an ordinary linear trend! It simply shows whether your income goes up or down, and also specifies the speed of these changes. It is absolutely enough for understanding the direction of your development. However, it is not enough for making an accurate forecast.

Use data from relevant time period for making the prediction. Everything has a life cycle, online-projects are not an exception. Therefore, it would be pointless to make predictions with the help of the linear trend that includes all data from the very beginning of the project: there are simply too many things that have changed. It makes sense to specify several stages of your project, understand the reasons associated with going from one stage to another, and make the forecast based on the data from the last stage. Do you remember yourself 5 or 10 years ago? At that moment, not everyone would be able to imagine ourselves in 2017. This is an illustration of why it is so important to divide time series into segments.

Mobile App Analytics


Autoregression turns out to be a more accurate method than trend and seasonality. You are building a regression model of income based on income values for one, two, or N previous periods.
It allows you to reveal hidden patterns in data, which couldn’t be found with trends and seasonality. The more periods you use when building the regression, the longer the period of your forecast can be.
Let’s say, you have data about your revenue for each day of the previous year. Then you can make forecasts for each of the 30 days of the next month by adding 30 variables to the regression model.


Pay attention to the fact that ARMA & ARIMA models are the result of the development of the regression model. Actually, “AR” in their names stands for “Autoregressive”. “MA” stands for “moving average”, which means that these models go deeper into data and reveal its internal patterns better. It might be difficult to implement these models in Excel (though there are some add-ins for this purpose), but still possible. I recommend to use statistical tools, such as SPSS or Statistica. However, these recommendations are based on nothing more than my personal experience with these tools.

Usually ARMA and ARIMA allow to make more accurate predictions than simple autoregression. However, the growth of accuracy is not as drastic as it can be when comparing autoregression with trends and seasonality. So, for a quick forecast there is no real reason to spend your time on ARMA and ARIMA.

Mobile App Revenue Forecast


Regression is quite a universal method. It has an important advantage over time series: in case of time series you make predictions based only on revenue values for previous periods, whereas in regression model you also consider other metrics.
There can be several ways to calculate revenue. For example, revenue is the audience multiplied by ARPU (revenue from a user). The audience is a quantitative metric, it says a lot about the scale of the project and is influenced by traffic. Revenue from a user is a qualitative metric: it shows how willing your users are to pay. These metrics can and need to be considered and predicted separately, because they behave differently and are influenced by different factors.
Similar reasoning can be done by looking at another revenue formula: paying users multiplied by the revenue from a paying user (ARPU). Theoretically, it is possible to use regression in relation to any of the metrics that you have.

Here I can give you the following advice:

  • If it is possible (not always possible in in Excel), use only significant variables when building the model. If you input a hundred of metrics, it is not necessarily mean that all of them should be used in the final equation.
  • Input metrics should be maximally independent from each other and weakly correlated. Otherwise, there is a risk of getting unstable results (that will be good at repeating your input data, but will produce strange values when it comes to forecasting).
  • Know your residuals. If you studied regression at the university, you probably remember a terrible word “heteroskedasticity”. This is what we are going to talk about, and it’s not that scary as it may sound! If you have done everything right, then when looking at the graph with residuals, you will not be able to say anything: there will be an unpredictable random value with a mathematical expectation equal to zero. If you see some regularity (let’s say a sinusoid), then it might be the case that you came across heteroskedasticity. This means that you haven’t taken into account an additional logic by which data is distributed. In this case, you need to change the regression equation by adding an unaccounted equation (in our case – sinusoid) to it.
  • App Analytics Mobile Games


    Actually, time series and regression are not the only possible ways to forecast revenue. If you want, you can always build your own models designed specifically for your project.
    I will give you an example of the model that I like to build:

  • First, we can calculate how many users are on the first, second, third, etc. month in the project at the current moment.
  • Secondly, we can calculate the percentage of users who stay active on the second month, as well as the percentage of users who proceed from the second to the third month and so on.
  • Finally, we can calculate an average sum that a user, who is on the N-th month in the project, spends during this month. In other words, ARPU (Average Revenue Per User) for the month.
  • There is enough information to build the model: you know how your users proceed from one month to the next one and how much they pay. By the way, it is possible to make calculations on a yearly, weekly, or even daily basis (although, to be honest, I haven’t tried taking a day as a period myself). It can be any period that suits your needs – depending on how long users stay in your project.

    This model helps you to plan traffic inflow. You simply need to increase the number of new users for a particular month.

    Method 2. Traffic pay-off period

    It sometimes happens that projects on their early stages highly depend on new users. This means that a project is profitable only if new users are constantly coming to it. That is why all the tips above are useless if you don’t know when and how much traffic you will attract.
    Thus, it would be great if you could build a curve with a cumulative income from your traffic by days: how much money on average a user brings during the first day, the first, second, or third week, the first month, and so on.

    Average Revenue Per User

    This is that value, the limit of which is LTV. When you know your cumulative income, you can forecast revenue more accurately based on when and how many users you’ve got, as well as calculate the payback of your traffic.

    Method 3. Expert forecasts

    The following tip is the most relevant for situations when you are planning some changes in the project that may have a big impact on revenue. This doesn’t mean, however, that you can’t use it when the impact is less noticeable.
    Say, your company is preparing to release some new content, features or a new subscription type, or anything else. It is a good thing to first ask people that are involved into the preparation (project manager, game designer, producer, marketer) about the impact these changes will have on the revenue. As an analyst, you will be able to make a reliable forecast by combining their expert evaluation and your accurate calculations.
    It is possible to do as follows: every month you ask certain groups of your colleagues about their expectations of the revenue for the next month. By accumulating data about their evaluations and actual revenue, you will be able to sort their opinions by relevance (and, maybe, even exclude some groups later).
    For example, you may notice that the producer tends to overstate the expected income, while marketers are too modest in their evaluations. The truth is somewhere in the middle, and previously collected data will show you where exactly. In addition, it is a very good way to learn about the planned changes first-handedly.

    Add game elements to the process! It will Increase your colleagues’ involvement.
    Sure, you don’t need my explanations on how to bet. It is important that if you play this game on a regular basis with key team members, it will increase understanding of the prospected revenue, and will allow everyone to find a logical connection between their actions and an indicator’s value.
    Moreover, this game helps to understand why your forecast didn’t coincide with the actual revenue: maybe, you haven’t considered all necessary factors?

    Method 4. Analyze all the changes

    If you want to better understand your project, audience, and revenue structure, I can recommend you to analyze all changes from the very beginning of the product’s life. For this purpose, try to keep a log of every published feature to be able to make more accurate and informed decisions.
    There will be a moment when you notice that some changes in your product considerably increase revenue and related metrics, while others don’t bring anything. As a result, you will be able to forecast revenue from a new feature more accurately by comparing it to the similar changes in the past.
    Actually, in this context it is worth recalling the Bayes’ theorem. By going deeper into his methods, you will inevitably realize that there are good news and bad news.
    Good news. Taking into consideration all “but” while forecasting revenue, makes the forecast much more accurate.
    Bad news. It is not possible to take into consideration all “but”.
    This is, however, not a reason to despair, you just need to take into account all factors that you can come up with and formalize. While the set of factors will grow, the complexity of the forecast will grow as well, and its accuracy will increase.
    Let’s say you are making a forecast for a football match. Your first assessment is an emotional one: “Barcelona will beat Real Madrid 3:1”. For some reason Barcelona doesn’t win and you start analyzing why. The forecast consists of many factors: the current position of teams in the table, the history of their matches, the factor of home advantage, injuries of the players, motivation, and so on. After a while, your forecasts will become much more accurate. Basically, you are training your own neural network: by analyzing mistakes, you add more factors into your forecast and make it more accurate in the end.

    Forecasting is an interactive process. Practice, evaluate and look at the results in detail – and you advance your skills in making forecasts and gradually become an expert.

    Method 5. Combine all the methods

    Not sure which method would work best for you? If this article was a test, the right answer would be “all of the above”. There are lots of methods and their combinations, and I don’t want to limit you in which one to choose.

    If you are curious, there is a combination that I use myself and that think it works well:
    forecast audience and ARPU (Average Revenue Per User) separately;

  • use time sequences;
  • adjust your audience forecast based on data about traffic inflow;
  • adjust your ARPU forecast based on planned changes in the project (with the help of expert evaluation);
    multiply audience and ARPU forecasts to get the revenue forecast;
  • it is possible to make corrections to the revenue forecast based on the expert evaluation of planned changes (in this case, make sure that all the changes are analyzed in detail), and on the custom model that describes how users proceed from one month to another.
  • Hope these tips will find a practical application in your project, and wish you accurate forecasts and growing revenue!

    5 methods to forecast your mobile app revenue
    5 (100%) 11 votes