Churn rate is one of the beasts that every SaaS founder has to fight with. The good or bad news about churn is that it’s neither a UX-only problem or a marketing problem or a customer success problem; it’s a company-wide problem that nobody can hide from, and everybody has to take responsibility.

Before I delve into the specifics of this case study, let me give you the backstory. A SaaS founder—who didn’t want his name to be revealed—found me on Linkedin and asked if I could help with his product’s above-average churn rate. After 4 months of working together, we managed to reduce churn in half. Here is how we did it.

Company: Popular B2B SaaS product that helps salespeople create beautiful proposals

Challenge: To find out why there was such a high churn and reduce it. Here are some product data:

Case Study Data Table.

Our approach

As you can see in the above table, the company that we helped had a high churn rate (>20%) before we take over. Even though this is a common problem among most SaaS companies, the solution is, most of the time, different. Also, in most cases, companies seek a dramatic decrease in churn in a timely manner.

However, you should know that when it comes to churn, small reductions accrue for significant gains. This means that a decrease of 1% or 2% can have a substantial impact on your bottom line. Having said that, we wanted to do the best we could, given the fact that the churn rate of our client was excessively high.

We decided that the best way to go was to conduct user research so that we can better understand what makes users churn. Is there a better way to understand why someone stopped using the product than asking them why they canceled right after they canceled?

Thus, our approach—even though simple—allowed us to identify the reasons why people churn and either stop using the product or find another solution. This type of user research is useful not only in terms of identifying why people cancel but also allows us to dive deep into what people need and try to give it to them through our product.

There are the steps that we followed for our user research.

Step 1: Predict customer churn

A churned customer is a user that decided to cancel his subscription with your tool. The hard truth about these customers is that, in most cases, they have come to that decision weeks or even months before you actually noticed it.

A way for us to predict this upcoming customer churn is to use activity churn metrics. Activity churn is the drop of product usage, in simple words, users stop using your product. That, if monitored, is the best signal of an upcoming customer and revenue churn that can be potentially prevented.

Photo by

The tools you can use to predict customer churn through activity churn metrics are multiple: Pendo, Mixpanel, Heap Analytics, Baremetrics, Chartmogul and many more but my favorites for this job are and

With the use of these tools, you can be served with smart signals whenever a customer is at risk of churning which is indicated by the slight drop in the frequency or breadth of usage he is having within your tool.

In this case, we used Brightback.

Photo from


Step 2: Prevent the preventable churn

In my experience, 30-40% of the reasons for which users churn are easily preventable. So, before we dive further into the qualitative analysis, there is a low hanging fruit that we need to grab. When you realise an account is going to churn, there are multiple tactics that you can use to prevent that such as:

  • Set up advisory calls
  • Extend the free trial or give a discount
  • Re-establish your value proposition
  • Pause the account for a smaller fee

In our case, we extended the free trial and set up quick web calls to help the user use the tool more effectively. That way, we managed, not only to prevent a 30% of the current churn but also collect super useful feedback that was passed back to the product team and me.

Home 1

On top of that, we also used Brightback to set up tailored cancellation pages.Through this feature, we served a personalised cancellation page with each user’s successes within our platform so far which re-established our value proposition. If a user decided to churn anyway, we served them with a feedback form and as soon as they filled it in, we started recruiting them for our Jobs-to-be-done customer research project that I am analysing further down the case study.


Step 3: Define the Main Questions

As said before, it is inevitable that some users will want to cancel and there is nothing you can do about it. After preventing a large percentage of churns, we had to go deeper. Therefore, we recruited 10 churned users to interview them.

Before the recruitment starts, we have spent some time preparing the project. To launch a user research project like that, we had to define some main things:

  1. What do we want to learn
  2. Which are the people that we want to ask

The obvious answer here is “why paying users to cancel.” But that question won’t get far. In most cases, people don’t tell the truth or even don’t know how to express the truth. So, to understand why paying users were cancelling, we had to answer some more in-depth questions and shape the timeline of their actions:

First Thought—This is the time they first thought they needed a solution like ours.

Event One (Passive Looking)—This is the time they passively start considering to switch to another solution.

Event Two (Active Looking)—This is the time they actively start looking. That’s typically connected with a real-life event.

Buying/Cancel Moment—This is the time they consider all of their options and decide to “fire” a product or service over another one.

Consumption—This is the time they are actually using our product or service.

Post-consumption feeling—This is the time they decide if they liked or disliked their experience.

As you can see from the timeline above, the time we actually see the churn in our product analytics doesn’t match the time the user started thinking of switching. The problem was there for weeks or even months until the user decided to stop her/ his subscription. If we just ask them (why did they cancel) they will most probably give us answers like that:

“It was too expensive.”

“I don’t need it anymore”

“Other… ”

As it’s obvious, none of the above answers is accurate, helpful, and satisfying. To understand what made people churn, we needed to unpack the feelings, thoughts, and emotions they had throughout all this period from the first thought until the moment they cancelled.

Step 4: The Recruitment

The part of recruiting—finding users who cancelled and are willing to get on a call—is never easy. This happens for two main reasons:

  1. These users are (usually) unhappy since they cancelled in the first place
  2. They are not willing to spend 30-45 minutes speaking to a stranger

To overcome this challenge, we need to make sure that we give them a psychological incentive and, sometimes, even a financial one. So, for every user that churned and gave us feedback through Brightback, we used the following screen-message:

We are devastated to see you go!

Let us buy you a year subscription to any streaming service you want (Netflix, Amazon Prime, Hulu, or any other service you like!)

We are really sorry that you decided to cancel your account. I can assure you that we are working day and night to offer a product that meets your needs. For us to do that effectively, it’s important that we deeply understand what we did wrong.

Therefore, I would like to ask you if you would be willing to participate in a 30-45min interview with our Head of Product. As a small thank you, we will buy you a year subscription to any streaming service you are using.

You can schedule the call here:{url}”

By providing an incentive, we managed to recruit almost ten people in less than 1.5 weeks. These ten people landed decided to schedule a call and—partly because of the incentive we offered, partly because they wanted to talk about the product—give an answer to the questions that we wanted to ask.

Step 5: The interviews

Running Jobs to be Done (JTBD) interviews for churned users is not an easy task. It requires experience, fast-thinking, and a positive attitude. I will walk you through one of the interviews so that you can replicate the process that we’ve used:

Pre-interview preparation

  1. Prepare a notebook and a pencil.
  2. Test your sound and camera in the tool you are using for the interviews (e.g. Zoom, Whereby, Skype).
  3. Write down a few of the most critical questions. You probably won’t need them all, but it’s always good to be prepared.


  1. Be positive.
  2. Ask for brutal honesty; say something like, “don’t be afraid of hurting our feelings.”
  3. Explain in advance that some questions might sound odd. Tell them that it will be fun, describe it like “a documentary of their story with the product.”
  4. Always ask if there is a hard stop on your discussion.
  5. Let them know this will take 30-45 minutes.


1. When did you cancel, why? (We want to let them complain at the beginning of the interview to get it out of our way.)

2. Where were you, what were you doing, what was the weather? (We want to set the scene.)

3. When did you initially sign up for the tool? Which other tools have you checked or evaluated?

4. How did you find it?

5. Why did you search for it in the first place?

6. What led you into searching for a solution like x

7. What was your experience with x the first time you used it?

8. What made you think it’s not for you? When did this happen?

9. When was the last time you used it? What happened then?

10. How long did you wait until cancelling?

11. What are you using now?

Following this process, we managed to uncover some vital information that would prove really helpful to our client. To illustrate my findings better, here is the timeline for one of the users I got in an interview with:

First Thought—He has pivoted his service into something different. The thought of recreating all the proposals from scratch was a big deal.

Event One (Passive Looking)—He has heard about products similar to ours and googled the term “agency proposal builder.” He kept our website’s URL on his Evernote for future use.

Event Two (Active Looking)—He landed a big client and wanted to impress him. He felt there was no space for failure.

Buying/Cancel Moment—He free trialled our service. It was challenging to use, but the end result was good. He was desperate to deliver something impressive, so he upgraded the service to unlock some features he wanted. He didn’t think the product was well designed and easy to use but had no alternative at this point in time.

Post-consumption feeling—He decided to give the tool another try but had no new client inquiries for two months. After being charged two more times, he left a note to himself to cancel the subscription. The amount was small, but he has to limit the monthly fees he is paying to useless subscriptions. He has now ordered a professional redesign of his word templates.

As you can see, these answers draw a clear picture of how the user felt before, while, and after using the product. We have done nine more interviews for this specific client, but if we use this one as an example, we can extract some pretty useful insights:

Insight #1: The customer was willing to pay more for a professional design that would be delivered in word format but would be as easy as what he was used to so far. We need to “copy” the UX of other apps that are massively used for proposals.

Insight #2: The thought of cancelling happened two months before, at the time that he first used the product. By that time, he stopped using the product. We should have reached him by then when we noticed the activity churn.

Insight #3: Several UX malpractices gave him a negative first-time experience. We need to prepare a smoother onboarding process. Also, until we are ready to redesign the UX, we need to have a customer support representative reaching out just in case customers need help.

Insight #4: Designing beautiful proposals was a job we were doing for him, but we failed to do it easily. We need to focus less on adding new templates and more on the UX.

Organizing all these notes might become overwhelming. What we did was start with a simple notebook while we were recording the calls. Then, after each call, we filled these two slides:


The first slide is where we keep notes on the forces that incentivised the action. The second one is a simple timeline we are using to put these events in chronological order. At the end of the research project, we were able to create an overall deck with the most common observations and any commonalities in the behaviours.

In the next step of the process—the fourth step—we analyzed the results and started running experiments based on what we’ve learned.  

Step 6: The Experiments

After having gathered all the insights, it was the time for experimentation. The Founder, the Head of Product, the Head of Customer Success, the Head of Marketing, and myself locked ourselves in a meeting room. We heard some of the interviews and studied my insights. Each one of us was keeping notes with ideas that could improve the situation.

When we finished listening and reading, we all shared our ideas. Everybody could share any idea, but there was only one rule: No idea should be that difficult to implement that it would take more than a week. After a couple of hours, we ended up having a spreadsheet like this:


As you can see, we ranked the ideas based on 3 factors: expected impact, effort and priority for our business. That helped us create an order for our experiments. After 3 months of running experiments that were aiming to address the insights we collected from the interviews, we had some serious progress.

The outcome

The results of this work were magnificent and were not just limited to the reduction of churn rate for that company. After these 4 months we managed:

  1. To reduce the churn in half
  2. To help the product team prioritize effectively
  3. To create an evidence-based optimization loop that just made things clear to everybody
  4. To build empathy for the whole team
  5. To regain peace of mind for the CEO
  6. To save the company from any future mistakes, they would do that would cost literally hundreds of thousands or even the company itself.

Final Thought

As I’ve mentioned in past articles, running JTBD interviews in the context of doing user research for your product is never easy. However, as I hope it is evident by now, it can be very useful in helping you understand what goes wrong with your product.

Of course, this doesn’t mean that you should start building every feature or fixing every bug asked by your customers. But, when something occurs more than one time and becomes a pattern, it’s something you need to consider.

User research and JTBD interviews can help you identify those patterns and take actions that will allow you to improve important SaaS metrics like churn, sometimes even by 50%.


About the Author: Aggelos M

Aggelos M
I have over 8 years experience in Digital and Growth marketing. Currently, running Growth Sandwich, a London based Product Growth Lab. Before turning in Growth Marketing and Product Growth, I had the chance to pass by head marketing and head digital positions in Athens and London, work with numerous tech startups but also build and run companies. In my spare time, I am consulting ambitious startups about their Growth and Business strategy.