Customer Retention – a key objective for the Product team

In my opinion, every product should have retention as one of the underlying Objectives (or health objective) for the product team, by default.

So, how do you measure retention?

The most commonly used criteria for measuring retention is number of users that are active in N day/week/month interval, where N is the usage interval.

In order to compute this metric, you need to determine two things first – your products critical event and usage interval.

A products critical event is the action that the user performs that shows that they are truly active on an application – e.g. booking a hotel room, booking a flight, completing an online course, logging an incident ticket etc.

The usage interval is a measure of the interval between two consecutive occurrences of the critical event – is it a daily or weekly or monthly used product?

How will you find out the usage interval time period N? If you don’t already have a good intuition about it - then you will need to draw a graph of the time interval of users repeating the critical event once. You draw a cumulative percentage graph of this, and find the inflection point. The graph will look something like this –

You can see, in this case, that the inflection point is at around day 9 after which the percentage of second users tapers off (at around 90%). So, we can assume in this case that N interval for measuring retention is 9 days.

Based on that you will compute -– number of users that are coming back to your app/application at multiple 9 day usage intervals progression. The retention graph over 1st , 2nd, … usage interval will measure how many users are coming back to the app between 0 and 9 days and then 9 and 18 days and so on for 5-6 intervals. This graph will be used for finding your core user groups.

User groups

An important consideration while formulating your retention strategy is to segregate users within your user base in the following brackets:

1) New user – the users in the first few usage intervals, before they form a habit of using the product. It takes a few months of use to form a habit

2) Current user – all users in any interval other than the new user, i.e, they are continuing to use the product. In a few months we can consider them as habitual users

3) Dormant user – the user that has missed the last few intervals. You can define the number of intervals before you label a user has dormant by studying usage characteristics of churn.

4) Resurrected user – the user that has missed one or more intervals, but has come back to the product again.

Based on the percent of users within each of these groups, as well as the rate and direction in which these are changing, you will have to determine the retention strategy to follow.

Additionally, there are user groups or cohorts based on user characteristics also needs to be created in order to understand the retention characteristics of the most important groups.

The benefits that drive retention – product strategy meets data analytics

Knowing the retention criteria only helps you to understand the retention trends among various user groups. It is not an actionable input in itself since it does not help identity the action that are needed to retain customer or increase user base.

What we need to find is what benefits is the customer looking for, in order to drive the retention strategy.

There are two ways of finding out the benefits sought, and both of these should be used by any Product Manager :

1) Finding the key benefits sought by talking directly to the customer – there is no shortcut to this process. The Product Manager has to meet the customer and find out the pain-points and benefits sought. There is a lot of posts on this, so I do not want to delve into this further.

2) Finding the behavior characteristics of the user from the usage data, and correlating this with the retention trends among users

Finding the behavior characteristics from the usage data is an intricate, but rewarding task. A product manager or data analyst has to spend hours scouring through the data to find these relations.

For example - Facebook found out that users who added at least 7 friends in the first 10 days, was more likely to retained for long. A behavior and an associated time frame are two inputs for a behavior criteria.

There are software tools, such as Amplitude or Pendo, that can be used for formulating and discovering key usage characteristics for your products.

Steps to discover your most important behavior characteristics are -

1) Find out your habitual users, core users and power users that are using your product regularly, more than anyone else

2) Follow the usage flow in order to figure out what they are using the product for, and find what these users have in common in terms of behavior – what are they doing or using

3) From the behavior, find out which ones show high retention or are highly correlated to high retention current users as opposed to dormant users

4) Fine tune the behavior criteria and time period to the retention.

With this you will have your hypothesis on your habit forming behavior. You will need to track this for some more time to confirm it.

Here are some example behavior characteristics that you should look at based on your game plan –

Stickiness – number of times the user utilizes the product within a certain time period.

Session metrics – amount of time the user is spending at a time with the product.

Once you have your habit forming behavior – your product strategy will have to find ways to make this behavior easier to find and use in the product.