Index:

• 5 things to keep in mind when you start looking into data 

• All About Churn Users?

• Connecting Google Data Studio 

• Vanity Metrics - Ignorance at its best

• Data Visualization Techniques  (Word-Cloud)

All About Churn Users?

In this blog post we will discuss about churn rate - what is it, how it's calculated, good churn (yes you read that right!) and ways to get churned users back.

Terminology :  

User : a person who is just browsing your site/shop/blog

Customer : A person who actually performed a conversion i.e. bought/subscribed/signedup

What is churn ?

Churned Users are the users who are no longer associated with you, in other words the users who have stopped using your product.

Why should we worry about churn?

Churn rate provides insights on whether you are performing as per the expectations of your users or not.

Also, majority of spends in a company goes into Customer Acquisition (CAC). Churned Customers are those who already bought from you once and hence are high intent users i.e. they need the product that is why they bought it in first place. Also, you have all the required data to reach out to them (their personal information like email id, phone number etc.). They are just one text/email away, which are more cost-effective marketing channels to acquire customers when compared to other channels like SEM and offline marketing.

How to calculate churn?

In order to calculate churn rate, firstly we need the number of churned customers. 

For a subscription-based company, the calculation is pretty straightforward - the customers who end their subscription or do not renew their subscriptions in the future are churned. But for a product based company, a user/customer does not have any time-bound window for making a purchase, i.e. a purchase can happen at any point of time. In such cases, you need to look into the repeat purchase behavior of the customers, take a 95% confidence interval and decide a time period post which you will consider those customers as churned.

Basic Formula for churn rate :

Churn Rate = Total Customers Churned / Total Customers at the beginning of the obervation period 

Here, Total Churned Customers is the number of customers who were churned during the observation period.

For eg,

A Subscription based company is looking to calculate Jan month's Churn Rate.

Total number of Customers it has on 1st Jan is 1000. Out of these 1000 customers, 50 get churned by the end of the month.

During Jan, company acquires 100 new customers, but out of these 100, 10 also churned by month end.

So total number of churned customers = 50+10 = 60

Table:

According to the basic definition, Churn Rate will be = 60/1000 = 6.0 %

There is an asterisk attached to the above definition as it does not take into account many crucial factors. For example, churn rate includes churned customers out of newly acquired customers but does not consider number of new customers in total customers calculation.

A better way to calculate Churn Rate is described in this recurly blog(https://blog.recurly.com/2014/08/better-way-to-calculate-your-churn-rate) written by Devin. According to him, we must calculate Churn Rate on Per Customer Day basis. 

In the above example, 1000 customers at the start of month, each has 31 days, and we will assume that each new customer acquired will have 0.5 of it. So total number of customer days  = 1000 * 31+100 * 31 * 0.5 = 32550

Total Churned Customers = 50+10 = 60

Churn per customer day = 60/32550 = 18.43%

Churn Rate Jan = Churn per customer day * 31 = 57.14 %

Even the above solution has an asterisk attached to it, as it assumes that all the new customers are acquired at a constant rate, which is why we use 0.5 for the calculation of customer days. But that's not true in reality, some newly acquired customers get more time to churn (users who acquired in first half of the month) when compared to others (users acquired on second half of the month). 

So, what is the best way to visualize churn ? The answer is Cohorts.Create rolling day window cohorts for each month!

How to create and read Cohorts is topic for a whole new post, will elaborate it in upcoming posts.

At the end, whichever definition you choose, make sure you keep all your stakeholders informed. Make that definition standard across organisation to avoid confusion. Different departments may end up with different churn rates for the same month, if the definitions are different.

Good Churn?

One of the purpose of this post is to enlighten you about the fact that not all churn is bad, some of it is actually good churn.

Wait, how can losing my customers be good? Well losing the bad ones is! This brings us to the question of who are "bad customers", rather the "Not-So-Good Customers". Every customer is different, you cannot satisfy everyone's need because every person's requirement is different. You must be aware of your product's vision and what your target segment is, and act upon their suggestions, not of every other user or customer. The customer is surely not always right - and your product may deviate from its core vision if you pay heed to a small group.

How to get'em back?

Try to find out why did your customers churn. There can be multiple reasons - like the product was not what they expected, or probably your website/app is not user friendly, delivery was delayed, need for inclusion of some more features, may be they are "Not So Good Customers" etc. Any Web Analytics Tools and Customer Surveys are the best ways to get the answer to this! Once you know why your customers churned, you can run reactivation campaigns to get them back - send out emails/sms, lure them via coupon codes, share upcoming feature releases and much more.

Again, this is not a one-time activity. Monitor this on a regular basis and DO NOT STOP LOOKING!