AppsFlyer together with AppAgent and Incipia has published a comprehensive guide on predictive LTV modeling which is a must-read for mobile marketers, UA managers and marketing analysts. We interviewed top experts from companies such as Rovio, Hutch Games, Wargaming, Joom, Wolt, Blinkist, Kiwi.com and Boombit to provide a comprehensive view on how LTV modeling differs for various business models.
What you will learn:
– What are the 3 main approaches to LTV predictions
– Methods for assessing marketing profitability with Excel
– How to predict in-app ad LTV
– What are the best practices by top experts
If you prefer watching a video, check my talk from MGS Berlin on How Do Real-Life Companies Use LTV predictions.
Predictive Modeling: Basic Concepts and Measurement Set-Up
Why Build Models?
There are numerous benefits to predictive modeling in mobile marketing. Knowing your typical user behaviour and the early milestones that separate users with high potential and users with low potential can be useful especially during an acquisition and re-engagement fronts.
So What Should I Measure?
To understand what you need to measure to get your predictions right and what’s not necessary, let’s briefly explore which data points are useful. The complete scope is in the full version of this article.
1) Legacy metrics such as CTR and CTI (low confidence in predicting profit, fastest availability)
2) Early indicator metrics such as CPI and retention rate (medium confidence in predicting profit, fast availability)
1) Tier 2 KPI confident predictors such as CAC (a great article on customer acquisition cost by my colleague Vit), CPA and many others (medium-high confidence in predicting profit, slow availability)
2) Tier 1 KPI confident predictors: early revenue (LTV) and consequent ROAS as an indication of long term success (high confidence in predicting profit, slowest availability)
Metrics are easier to calculate and mature much sooner than KPIs, which tend to take longer or involve complex formulas.
Pros and Cons of Different LTV-Based Predictive Models: Insights from Top Marketers
Building an LTV model to predict Return on Ad Spend (ROAS) could be overwhelming due to obvious differences in the way different types of apps retain and monetize users; just think of how distinct in-app purchase games, subscription-based apps and e-commerce businesses are. It’s clear that there cannot be a one-size-fits-all LTV model.
1) Retention-driven / ARPDAU Retention Model
Model a retention curve based on a couple of initial retention datapoints, then calculate the average number of active days per user (for Day 90, D180, etc.) and multiply that by an Average Revenue Per Daily Active User (ARPDAU) to get the predicted LTV. Good for high-retention apps.
Calculating a coefficient (D90 LTV / D3 LTV) from historical data and then for each cohort, applying this coefficient to multiply the real D3 LTV to get a D90 LTV prediction. Good for “Standard” types of apps including many game genres.
3) Behavior-driven / user-level predictions
Collecting a significant volume of data from app’s users (session and engagement data, purchases, geo / device type, etc.) and processing them using machine learning to define which actions or action combinations are the best “predictors” of a new user’s value. Good for any app with access to an experienced data science team, engineering resources and lots of data.
From MVP to complex models
A good way to start your prediction path is with a simple “Minimum Viable Product” (MVP). The idea is to verify initial assumptions, learn more about the data and gradually build a model. That cost/benefit ratio of more complex models is not necessarily better than simpler ones which was proven by the fact that companies confessed that they tend to stick to conceptually simple models.
Teams & Responsibilities
Ideally, there are two roles: an experienced analyst with an overreach to marketing that can advise on the strategy and tactical levels as well as decide which model should be used in addition to how it should be used; and a dedicated analyst which then “owns” LTV calculations and predictions on a day-to-day basis. Outsourcing can certainly jumpstart the process, especially if a company has limited knowledge of the topic. However, in the longer term, given the product is found to be viable and more advertising dollars are being spent on more networks, an internal team should take over.
Methods for Assessing Mobile Marketing Profitability with Excel
Excel is more powerful than you think. By using a scatter plot and bit of algebra, you can turn an Excel trendline equation into a powerful tool for identifying early on the point at which your marketing campaigns prove they are likely to turn a profit. Here is a very quick go-through-guide:The first step is to ensure you have enough Week 0 and 6-month data points. A rule of thumb for Week 0 ROAS-based predictions, you should shoot for at least 60 pairs of Week 0 and 6-month ROAS observations. The second step is to split your data set into two groups, one for training and one for prediction. Place the lion’s share of the data (~80%) in the training group. The third step is to use a scatter plot to graph the data, with the Week 0 ROAS on the x-axis and 6-month ROAS on the y-axis. Add a trendline and add the equation and R-squared settings. Step four involves using the y = mx + b linear equation to solve for the equation’s x value (Week 0 ROAS) when the y value (6-month ROAS) is 100%. The answer to the question of how to predict profit at 6-month is that your ROAS must be greater than x in the first week. Step five is where you use your prediction segment of the full data set to assess how well your model was able to predict actual outcomes. This can be assessed using the mean absolute percentage error (MAPE). You may even try out a few more trendlines (exponential, logarithmic, etc).
Adding Another Piece to the Pie: Predicting In-App Ad LTV
In-app advertising (IAA) has become increasingly popular, accounting for at least 30% of app revenue in 2018. Even developers who had been completely reliant on in-app purchases (IAP) have started monetizing with ads.
Ideally, marketers would be able to understand the nominal value of every single impression; that would practically make it a “purchase”. After gathering sufficient data, we’d be able to create prediction models similar to in-app purchases. However, in the real world it’s not that simple because of the volume and structure of provided revenue data. To list a number of issues – not one source of ads leads to different eCPM, some ad networks pay for different actions (install, click) etc.
Among interviewed gaming app marketers, none had this actually figured out to a stage they’d be happy using it. Calculating user-level in-app ad LTV at the level of precision we’re used to having with in-app purchase LTV will likely continue to be a challenge, at least in the near future.
It leads to another approach called “The Contribution Method”. Contribution margins work by converting a channel’s contribution to overall user behavior into that channel’s earning margin from the overall ad revenue generated by all users. The theory is that the more a channel’s acquired users generate actions in an app, the more influential and deserving that channel’s hand in claiming credit for advertising revenues from those users.
Best Practices for Building Mobile Marketing Prediction Models
When building data models or systems that are used to guide significant decisions, it’s not only important to build the best system possible but also to perform ongoing testing to ensure its effectiveness. For both purposes, make sure that you continuously feed your profit prediction model to keep it trained on the most relevant data. Also, put in enough effort to choose the right KPI for predicting profitability. For better accuracy, try to segment your users into more homogenous groups and remember to factor time.
If you are interested in learning more on how to calculate customer acquisition cost that is closely connected LTV in terms of business economics, check the article on “How To Calculate Customer Acquisition Cost (CAC) for your mobile application” by Vit Volsicka.
Teemu Rautiainen, Elif Buyukcan, Kasim Zorlu and Leonard Seffer from Rovio Entertainment
Alexandra Lomakina from Joom
Fredrik Lucander from Wolt
Matej Lancaric from Boombit (formerly at Pixel Federation)
Gessica Bicego from Blinkist
Anna Yukhtenko and Tim Mannveille from Hutch Games
Martin Jelínek from AppAgent
Gabe Kwakyi from Incipia
Shani Rosenfelder from AppsFlyer
Predictive Modeling for App Marketers: The Complete Guide https://www.appsflyer.com/resources/gaming/predictive-modeling-app-marketers-guide/basic-concepts-measurement/
📕 Learn more about industry insights and best practices by signing up for our newsletter here or by bookmarking our blog.
🤝 Get help with the launch of your app or game, growth strategy, app and marketing analytics, ASO, user acquisition, video and playable ads production by reaching us at peter at appagent dot co
👀 Follow us also at LinkedIn, Twitter, Slideshare, YouTube, Facebook.