Product Data Metrics: Expected Revenue Per Install (ERPI)

“Data just doesn’t sit in our mind as much as stories do. More importantly, stories have emotions that data doesn’t.  Emotions get people to do all kinds of things, good and bad.” – Dan Ariely

How do we tell accurate stories around data? In a mobile subscription business, we have plenty of metrics to tell stories. For example – install to trial, convert to pay, churn and lifetime value, NPS, and leading indicators everywhere. The metrics bounce around week-to-week, month to month and quarter-to-quarter. 

Hopefully, your product team is making meaningful investments into each area of the funnel. Most importantly, you want to ensure that these numbers are all going up. But the reality is, while some go up – some stay flat and some go down. In addition, the nature of your audience and traffic will also impact your funnel development. Intent can be very high and your trial start rates go up, but satisfaction may not be met, and churn goes up as well.

How do we measure the overall output of our work? How are we really doing? I devised Expected Revenue Per Install (ERPI) to have a simple metric to answer these questions month over month. I wanted to understand where we should be making product investments, and how to measure the overall health of a mobile subscription app. And most importantly, to tell the right stories about our data.

What is ERPI?

At a high level, ERPI is easy to understand. It resets the classic Customer Acquisition Cost (CAC) to Lifetime Value (LTV) ratio back to the point of an app install. You can calculate ERPI by taking your new subscribers in period, multiplied by the estimated average lifetime value of your subscribers, divided by your total install in period.

New subscribers in period * Estimated LTV ÷ Total Installs in period.

Let’s take an example: your app has 4,500 new subscribers in March. The estimated lifetime value of those subscribers is $70. 

For the sake of this example – you have no free trial period and 50,000 installs in March. Your app’s ERPI is 4,500 x $70 ÷ 50,000 = $6.30 ERPI. 

ERPI also lines up perfectly with your Cost Per Install (CPI). As a result, your teams can easily understand how profitable they are on spend.

Why an install?

An install is the first measure of true intent by a new user. In addition, you gain a deeper level of intent than a website visit.  We browse websites casually, and we browse apps on our devices casually, but installing an app onto your device is a more personal experience. And while CAC may happen much further down the funnel, the product team’s responsibility is likely heavily focused on the user experience from the point of install.

So how do we calculate ERPI?

We know the high-level calculation, but how and where do we get this data, to ensure ERPI is accurate?

We’re going to step through this calculation using what I like to call the “startup” method. Automating the calculating of ERPI will allow you to more truly cohort your users, as well as let you see ERPI on a weekly or even daily basis. But that’s for more mature teams (and a different article).

To start the ERPI calculation, you need to take the length of the free trial into account. When we pull data on our new subscribers “in period” and our total installs “in period,” we need to ensure they are from the same cohort. If you have a 30-day free trial, and you are measuring your ERPI for December, then you need to measure against your app installs in November. 

For this example, we are also presuming your users start trials pretty quickly after install. If 90% of them start trials within 24 hours of install, you don’t have to account for this timing, but if your user experience involves a heavy free usage period before a subscription is required, you may have to do some more involved cohort reporting. If you have a trial of one week, it’s fine to pull your new subscriber data and your total install data from the same month.

While new subscribers and app installs may be easy data to obtain, lifetime value is more difficult to calculate.

While the cancellation rate can give you directional data on your subscription retention, it is susceptible to changes in traffic volume. As a result, it doesn’t give you the whole picture of how your members retain over time. To get a better picture, we need to use cohort churn. This will help in projecting an estimated lifetime value for the current cohort.

Cohort churn involves looking at each month that your members have been subscribers. Furthermore, this process will pull data on what percentage of members do not make a payment in the subsequent month. Month one churn is going to naturally have the highest volume of traffic, but it will also have the highest churn, as members are still deciding whether there is a real utility to your product. 

Thus, month one churn is the percentage of members who do not make a payment in their second month and churn out. This number can range depending on your product’s maturity, industry and pricing. Ideally, it’s less than 20 percent (although I’ve seen it much higher!).

In my next article, you’ll see ERPI in action and I’ll go through how to use the data to enhance your product performance.

 

Click here to read our latest blog posts

About the speaker
Britt Myers GameClub, COO Member

An experienced and creative entrepreneur and product leader, Britt Myers has developed an impressive resume of business successes in media and technology production. In 2014, Myers partnered with Stephanie Dua as co-founder and Chief Product Officer of ed-tech startup Homer. Homer is the #1 Learn-To-Read program powered by your child’s interests; an educational app for iOS and web that teaches a child to read and develops crucial early childhood cognitive skills.