Archives For LTV

Last year, I published a series of posts on the importance of knowing your users’ or players’ lifetime value, the key components and how to impact them and techniques to increase the accuracy of your customer lifetime value (LTV) predictions. I intentionally did not publish a formula for calculating LTV—while it is always a factor of retention, monetization and virality—as it is different by product and there are many alternative ways to get to an accurate customer lifetime value. Prompted by an infographic that I came across (see below) I did want to go into some details of the mechanics of calculating LTV.

The first step is to obtain your key variable metrics as averages across all users. The ones I prefer are ARPDAU (average revenue per daily active user), day 1 retention (how many people who use or install your website, app or game come back the next day), day 30 retention and k-score (how many free/organic users does a user bring in).

The second step is to estimate your “constants,” as Kissmetrics refers to them in the infographic. The key constants you need to understand and estimate are average customer lifespan (and days played in that lifetime) and customer retention rate. Lifespan is how long the customer will remain a customer, which is derived and estimated from your retention data. For a game, you may want to start by assuming six months and then adjust that number as you acquire better data on your users. The second metric, customer retention rate (also often referred to as churn), is the percentage of customers that, given a certain period of time, will return when compared to an equal period of time. As the infographic shows, there are other constants you may want to incorporate into your model (and I have discussed some of these in previous posts) but the core of an accurate LTV prediction for a tech company is understanding your user’s lifespan and churn rate.

Once you have your data, both the variable and constant metrics, it is time to estimate your LTV. As I have written before, always keep in mind that your calculation will be a prediction and actually represents a range of expected values of a new user. Also, to generate actionable data, it is more important to understand LTV of users from different ad sources and cohorts. That said, once you have the data either aggregated or by source/cohort, you then build a function to calculate your customers’ lifetime value. The infographic shows several formulas based on the metrics their example (Starbucks) considers useful. If you have focused more on the metrics that I mention above, an oversimplification would be ARPDAU times number of session in the lifetime of a player (which will incorporate churn rate) times their k-score.

Key takeaways

  1. Step one: obtain metrics for your variable data (average of players), such as ARPDAU (average revenue per daily unique user), day 1 retention, day 30 retention, etc.
  2. Step two: determine the constant metrics, most important being customer lifetime and churn rate.
  3. At its simplest, your customer lifetime value will be ARPDAU times number of session in the lifetime of a player times their k-score.

++ Click Image to Enlarge ++
How To Calculate Customer Lifetime Value
Source: How To Calculate Lifetime Value

A few weeks ago, I posed a question on Quora about the differences between machine learning and predictive analytics. I was surprised at the number of people who started following the question and actually liked it (although it sounded rhetorical, I was hoping to understand machine learning better). A recent article in Fast Company about The New York Times, of all companies, did a great job of explaining the fundamental value of machine learning.

My interest in machine learning was ignited recently as it has become the hot buzzword in the Bay Area; some have argued if you add machine learning to your PowerPoint you can add a zero to the end your company’s valuation. While that claim is obviously an exaggeration, investors are among the savviest businessmen, so their interest in machine learning shows it is a crucial emerging space.

The article discusses Chris Wiggins, a biologist The New York Times just hired as its Chief Data Scientist. Wiggins’ mandate is to build and lead the Times’ machine learning team. In Fast Company’s interview with Wiggins, it became clear exactly what machine learning is, how it is different than predictive analytics and why it is important.

What is machine learning

According to Wiggins, “Machine learning sits at the intersection of data engineering and mathematical modeling. The thing that makes it different from statistics traditionally, is far more focus on building algorithms.”

University of Utah image for machine learning

Also, while statistics is traditionally focused on explaining data, machine learning is geared to building predictive models. When Netflix or Amazon make product recommendations to you, they are using machine learning to predict what you would be interesting in experiencing. Continue Reading…

I am very proud to announce I will be speaking at MIT’s CDOIQ (Chief Data Officer and Information Quality) Symposium in July about how data affects your LTV projections, and ways to improve the quality of your metrics. It’s going to be a great conference and I would love to meet up with anyone who will be there.
MIT Logo

Given the importance of analytics to social and mobile game companies (just see all my posts about LTV, performance marketing, virality, monetization, etc.), having the best business intelligence (BI) team is of central importance. Finding that talent, however, is not easy. I have been very lucky to work with some of the best BI talent throughout my career; they have made me look much smarter than I am. Not everyone will be as lucky as I have been. A recent article in the MIT Sloan Management Review provides great advice on predicting the performance of potential analysts.

Data Ninja

The article points out that the ideal analyst does not exist; the job description is looking for a “unicorn.” You should not be hiring for a laundry list of skills (e.g., “I need someone with R, SWRVE and Mixpanel expertise”) because the game industry is evolving so quickly most of those skills will soon be outdated. Instead, you should look for the curiosity to keep learning, rather than the skills themselves. Continue Reading…

Below is a presentation that I gave yesterday on lifetime value (LTV) to the portfolio companies of YetiZen. It covers the importance of LTV, key variables (monetization, virality and retention) and how to affect them, importance outside gaming, cohort analysis and the predictive nature of LTV. Other than the final section on uncertainty, which echoes my blog post on Tuesday, the presentation is largely consistent with the one posted earlier that I gave at Groundwork Labs a few months ago. Here is the one from last night: