BI For Experts

Churn Analysis: A guide for SaaS founders

In this post you'll learn how to conduct a churn analysis and how to use the insights to grow your business.

Churn rate is the single most important metric for determining the growth trajectory of a SaaS business.

A business that has a high churn rate will struggle to grow, no matter how good the company is at marketing it's products or services.

Imagine trying to fill a bucket with holes. A SaaS business with a high churn rate is fighting the same type of battle.

How to conduct a churn analysis?

Before you jump into your churn analysis you need to do some initial planning.

Step #1 - Defining churn

The first step in conducting a churn analysis is to define churn within your organization.

Churn can mean different things to different people so it's critical that you come up with specific definitions and then add these to your data dictionary.

My suggestion would be to start with the following terms:

Subscriber / Logo Churn

Defined as the count or percentage of individuals or businesses that stop paying you within a given cohort.

B2C companies should use the term subscriber churn, B2B companies should use the term logo churn (logos as in company logos).

Revenue Churn

Defined as the absolute value, or percentage of the total revenue that is lost as a result of subscribers or logos cancelling or downgrading their subscriptions.

Step #2 - Planning the churn analysis

Now that you've defined churn you can start planning your analysis.

The first decision you need to make is to decide which question or questions you want to try to answer in your analysis.

Below are some examples of the types of questions you may want to tackle in your churn analysis:

  • What is my subscriber / logo churn rate?
  • How many of my subscribers are still paying me after 6 months?
  • Do users that signed up to our quarterly plan churn at a higher, or lower rate, than users that signed up to our monthly plan?
  • Do subscribers from non-English speaking countries churn at a higher rate than subscribers from English speaking countries?
  • Which of my account managers are doing the best job of lowering logo churn?

For my example we are going to start with the basics.

I'm going to show you how to create a visualization that will help us analyze our subscriber / logo and revenue churn. Once we have the high level picture we can start slicing and dicing our data to answer more granular questions.

To simplify things a bit I'm going to make the following assumptions:

  • We are a B2B SaaS business that sells monthly subscriptions only.
  • There is only one plan which costs $50 a month.
  • The company only works with one currency, US dollars.
  • The company started offering subscriptions from the 1st January 2019.

Don't worry, I'm still going to show you how to tackle quarterly or annual subscriptions. You'll also soon see that the methodology I share below can be used to calculate churn for a B2C business.

Step #3 - Prepare your subscription data

This step in the process is by far the most tricky. If you don't have any experience with SQL or at the very least with Excel, you are going to struggle.

The data set you need to build in order to conduct a churn analysis needs to have the following structure:

Each row in our data set needs to represent a successful transaction.

You want to have the following columns in your data set:

  • Id of the transaction (typically transaction or charge Id)
  • The timestamp of the transaction
  • The value of the transaction (if you work with multiple currencies you will need to pick one and normalize your values to that one currency)
  • The internal id of the individual / business who paid you (typically user Id or account Id)
  • The date when the individual or business first became a subscriber
  • The number of months between 1st subscription creation date, and the transaction ("datediff").

Your data set needs to look like the dummy data set shown below:

churn analysis data set

Step #4 - Build your cohort tables and retention curve

Once you've collected all the relevant data and structured it as I've shown above, you can easily pivot the data it into the following cohort table.

churn analysis cohort table

In the screenshot above you have a cohort table showing the number "logos" that become subscribers in a given month (cohort), and how many paid you over a 6 month period. Column zero is your starting point. It contains all the subscribers that started in that month.

This view is helpful but a better way to view this data is by looking at each cell in each row as a percentage of total. When you do that you get the following table.

churn analysis cohort table percentage of total

By adding conditional formatting to the table, we can now clearly see where we stand in terms of logo retention and churn.

We can see that we are getting worse at retaining our accounts since there is more red in the bottom right hand corner of the table. We used to retain 75%+ in month 6 (column 5) but now are under 70%.

In my example I only showed 6 month retention but ideally you want to expand that out as far as your data allows so you can see as full of a picture as possible.

The next step in the process is to calculate your average retention rate and create a retention curve graph.

Based on my dummy data, my 6-month retention curve would look like the graphic below.

retention curve example

The main take away from this graph is that after 6 months we've lost on average a quarter of our subscribers.

If it is easier for you to see the churn curve, which is the inverse of retention, then simply calculate 1 minus the retention rate for each period.

churn curve example

In the case of a business which offers quarterly or annual subscription options, the approach is exactly the same. The only differences are:

  • Longer lag time - More time needs to pass in order to have enough data points. A subscriber who pays quarterly only has 4 data points in a 12 month period.
  • When preparing your data you need to add an extra field which stipulates that a user is on a monthly, quarterly or annual subscription. Build tables and line graphs for each subscription type, or add a filter which makes it easy to view one category at a time. The tables will look different for each category unless you add additional logic so that the count of each payment is the same for all categories.

Step #5 - Start answering your churn-related questions

The visualizations you've created at this point will give you a high-level understanding of your retention and churn rates.

You might be in shock at these numbers but at least you now know where you stand.

Having a benchmark for your churn rate is the first step in improving your churn rate.

In the example above I calculated logo churn. You can repeat the process for revenue by summing up the total amount generated from transactions by subscribers in each cohort. Since everyone paid in column zero, this will be your starting amount.

As more subscribers cancel, you collect less money from each cohort over time.

If you want to calculate the churn rate of specific subscriber segments, like US companies vs. rest of world, or the performance of certain account managers, then you'll need to enrich your initial data to include these additional dimensions.

Now instead of pivoting with subscription creation month as your grouping, you can group by whichever dimension you care about. You could even group by 2 or more dimensions to see if churn rates have improved for certain segments over time.

Your business needs a churn dashboard

I highly recommend that you invest in an automated reporting solution such as Tableau, PowerBI, or event Google Data Studio so you can quickly view your retention and churn metrics.

Churn is such an important metric for the business that I believe every decision maker in the company should have access. The best way to do this is by implementing a cloud-based reporting solution.

There are of course a number of considerations involved in adopting a sophisticated reporting solution but thankfully the costs, learning curves and access to skilled human resources in data visualization is improving all the time.

The process of building a churn dashboard is very similar to the process I covered in this blog post but you'll need to either invest in learning a reporting tool, or hire someone that can set up the dashboard for you.

How to reduce churn

Congrats, you now know your churn rate and have a basic dashboard in place that can help you keep tabs on this critical metric.

You've rounded up everyone in the company and set a goal of reducing churn by 20% over the next 6 months. You're excited and everyone is with you but you don't have a clue where to start.

Don't worry, the rest of this blog post is dedicated to helping you build out a comprehensive action plan for reducing churn.

Assume you know nothing

This might be tough to accept but trust me, you'll be better off assuming you know nothing about your subscribers, especially why they cancel.

In order to start understanding your subscribers you need to start collecting qualitative information from them. You can do this in a number of ways.

Below are just some ideas:

  • Set up a cancellation survey.
  • Send out surveys to users when they reach certain milestones in their user journey (they will be more likely to respond at this point).
  • As the founder, you should commit to chatting with 5 customers a week over Skype / Zoom. Use Calendly and add a widget within your app to make it easy for subscribers to schedule time to chat with you.
  • Meet customers face-to-face when visiting a new city on business.
  • Add a step in your onboarding process in which your account managers ask strategic questions to your subscribers.
  • Add more questions to your "did support do a good job" survey.
  • Launch a beta program so high engaging customers can opt in to try new features. When new features are released get your product managers to get on the phone and survey these engaged customers. Expand the questioning beyond the scope of the product.

Collect all this information in Google Sheets and block off time to go through it. The more data you collect the more trends you'll identify.

Focus on getting to the "wow" moment

Most great products have a "wow" moment, that moment when the user sees the value you offer with their own eyes. It's a moment of enjoyment and one to be remembered for the user.

You have to understand what that wow factor is for your product or service and help users get to that point quickly.

If your business is self service then this has to happen seamlessly within the product. In the case of a more complex platform the responsibility shifts to your customer success team.

Your account managers (also often called success managers), need to understand why the business signed a contract with you and what value they expect from your product or service. If they don't you probably have a poor process and a disconnect between sales and customer success.

If the account manager knows what value the subscriber expects then it is the main responsibility of the AM to help the subscriber receive that value as quickly as possible. This typically takes the shape of an onboarding call, and subsequent follow ups.

There are some amazing platforms out there like Gainsight and Totango which are designed to help account managers manage these responsibilities. If you are an operations-focused company and have the budget I'd highly recommend adopting one of these solutions.

Invest in a retention team

Many companies have a team dedicated to reducing churn. Members of these teams are somewhere between success managers and sales reps.

There primary responsibility is to actively reduce churn and ideally even help the company achieve negative net churn (make more from existing users than you lose).

The retention team can also help you collect useful information on why subscribers are cancelling and feed this information back to other departments.

Understand the competitive landscape and adapt as needed

One of the biggest drivers of churn are cheaper or better alternatives.

Think about this in your own life. You might go every week to the same hamburger joint until you find that a better one has opened up just around the corner.

If you're doing a good job in your market others will pay attention and competition will grow. If these competitors start offering a cheaper and/or better solution then your customers will jump ship.

In order to counter this you need to constantly be improving your solution and keep an eye on the market landscape.

Some businesses hire a VP of Strategy who is responsible for surveying and analyzing the market landscape in order to protect the company against changes in the market. If you aren't at that point yet then focus on building out a great offering and outwork your competitors.

There are many ways to create a competitive advantage and they don't all involve massive capital or other rare resources.

Invest in business operations and become proactive instead of reactive

One of the biggest impacts you can make to your business is a shift in mindset from a company that is reactive, to one that is proactive.

To reach that point a lot of variables have to be molded into place so you have a well-oiled machine that operates efficiently.

One of the first steps to get to that point is to invest in business operations, in particular business intelligence. My BI competency status framework is a good place to start.

Once you're organization is collecting operational data which covers the entire customer life cycle, and can make that data accessible to your customer success teams, you are ready to make the shift.

I previously mentioned Gainsight as a great solution to help you become more proactive. Tableau can also be used to provide self-service business intelligence capabilities to your account managers and support staff.

Once the infrastructure is in place the next step is to adapt your internal processes to leverage the new infrastructure.

Account managers should be trained to use the data available to them to proactively help users receive more value and have an overall better experience.

Each AM should have time blocked off each week to go through their accounts and identify ways that these accounts can get more value.

Below are some examples:

  • The AM notices that one of his accounts opened a support ticket a week ago and no one has responded. The AM can forward the ticket to a senior support rep so it gets taken care of immediately.
  • The AM notices that one of his accounts stopped using a key feature. The AM can reach out and ask why the account stopped using the feature and forward the feedback to a relevant product manager.
  • The AM notices that one of his accounts has invited a new team member to their account. The AM can reach out and introduce himself to the new contact and offer to conduct a demo to the new contact.
  • The AM notices that one of his accounts has never used a certain feature. After some investigation it looks like the subscriber would really benefit from that feature. The AM can now write a detailed email to the main point of contact explaining how the feature would benefit the subscriber's business.

If you made it this far, thanks for reading and if you have any questions on churn then please comment below and I'll happily help you out.