What You Can Learn from Churn (with Founder and CPO of Churnkey, Scott Hurff)

Churn is far from a static metric that you can’t influence. Scott Hurff talks about how to create humanistic cancellation experiences and then learn from them so fewer customers have to cancel at all.

What You Can Learn from Churn (with Founder and CPO of Churnkey, Scott Hurff)
Do not index
Do not index
Scott and I talk about churn — how he approaches it at Churnkey, and how he hopes more companies will approach it in general. As he says, “It’s one of the most underrated sources of information and intelligence.” Creating a cancellation system not only helps customers feel like they’re being heard as you part ways but also gives you precious insight to get ahead of potential cancellations before it’s too late.
Video preview


(09:49) Creating Customer Profiles with Churn Data

Learning from churn can be easier than you think: a simple survey, prompted by cancellation, that encourages customers to share their experience. Of course, the questions in the survey need to be specific enough for insights to emerge. Having users opt into responses like “I no longer need this service” doesn’t help them feel heard or help you pinpoint what’s not working for them anymore.
Once you have enough responses to start finding patterns, you can create customer profiles and use them to design better marketing, activation, and retention experiences. The key is to build an understanding of who your best customers are, even if it's at the point of parting ways with them. Scott says:
If people are canceling, there's an intent there. So, what are the reasons that people cancel? And how are they responding to your cancellation system? 

Once you can collect that data, it’s about taking a signal from it, rolling it up into understanding who are the people who are just gonna go all out for you. Even if they leave, they still love you. So who’s that customer? How did you get them? What was their price point? How long were they subscribed? What was their customer support history? Do they have any parting advice?
If people are canceling, there's an intent there. So, what are the reasons that people cancel? And how are they responding to your cancellation system? Once you can collect that data, it’s about taking a signal from it, rolling it up into understanding who are the people who are just gonna go all out for you. Even if they leave, they still love you. So who’s that customer? How did you get them? What was their price point? How long were they subscribed? What was their customer support history? Do they have any parting advice?

(19:17) Using Churn to Refine Value Metrics

Facing churn head-on can be difficult, but where better to learn if your app was capable of providing value? Churn data can be particularly insightful, Scott says, if you run it through your value metrics. If you’re an invoicing company, for example, and your value metric is “number of invoices paid” — ask how many cancellers managed to get their invoices paid. Or why they couldn’t, or what they were hoping to do after that they couldn’t figure out.
Churn can also be a value-metric-reality-check. Does optimizing your value metric have a positive influence on churn? If not, it might not be so valuable to your users. They might have different expectations in mind:
The [churn responses] you wanna look for in particular are the “misalignment in expectation” ones, which means “I wanted to do X, but you gave me Z.” Or, “I wanted to do X and it worked out some of the time, but it wasn't consistent.” There could also be this phenomenon of misperceived value where you are providing it but you're not stating it in a way that is practical and easy to uncover and understand.
The [churn responses] you wanna look for in particular are the “misalignment in expectation” ones, which means “I wanted to do X, but you gave me Z.” Or, “I wanted to do X and it worked out some of the time, but it wasn't consistent.” There could also be this phenomenon of misperceived value where you are providing it but you're not stating it in a way that is practical and easy to uncover and understand.

(39:15) Key Questions to Ask of Your Churn Data

When it comes to extracting insights from churn data, here’s Scott’s list of questions to begin with:
I would start by asking:
I would start by asking:
  • [For a churn cohort] Were users from this cohort promised a common thing?
  • Were they promised it at a certain price point?
  • Did they all come from a specific source or similar sources?
  • Did they sign up at the same time?
  • Was there a recent spike/dip in churn? Why? Was there company activity that attracted people with similar characteristics?
  • [For a churn profile] Does this profile have particular traits that would make them price- or seat- or usage-sensitive?
  • Did they perform <the behavior that your value metric is tied to>?

23:35 Applying End-Of-Lifecycle Lessons to the Early Lifecycle

Gathering data, and turning the data into insights are the first two steps of the “learning from churn” process. The final step is applying those insights so they result in better conversion and retention. Patterns that emerge from the churn data need to be tested and refined with new users, and early in the customer lifecycle because that’s when you lose the most people.
It would be amazing if you could connect your Churnkey data and your billing data, and just immediately understand, “Who are my best customers? What do they do? Who are the worst customers? What do they do?” And then go further up the activation process and look at acquisition sources, price points, payment methods, and whether they were discounted. These things just stack up over time. This kind of analysis shouldn't be an archeological operation.
It would be amazing if you could connect your Churnkey data and your billing data, and just immediately understand, “Who are my best customers? What do they do? Who are the worst customers? What do they do?” And then go further up the activation process and look at acquisition sources, price points, payment methods, and whether they were discounted. These things just stack up over time. This kind of analysis shouldn't be an archeological operation.


Want to read along?
Toggle Transcript
Yohann: Hi, I'm Yohann, and you are listening to the Self-Serve SaaS 'Cast. In this episode, I'm talking to Scott Hurff, the Founder and CPO of Churnkey, which is a platform that helps companies manage and learn from their "offboarding" process, as opposed to an onboarding process.
Scott is so passionate about churn. As he says, it's one of the most underrated sources of improvement and intelligence. I can't think of a more compassionate and knowledgeable person to teach us how to learn from churn. I felt like I was learning from a friend.
And it's so important to talk about churn, especially in a Self-Serve context, because the quality of your Self-Serve system basically comes down to the signals you gather from your users and customers across the customer lifecycle. Scott walks us through how you can create a more humanistic cancellation experience, and learn from churn so that fewer customers want to cancel at all. I hope you enjoy our conversation. Here we go.
Hey Scott, thanks so much for taking the time to talk to us today.
Scott: What's up, Yohann? Good to be here.
Yohann: I'm so excited to talk because not only am I curious about Churnkey's story, but I'm also insanely curious about Churnkey's insights when it comes to tackling churn. And I feel like we can dive into both. What do you say?
Scott: I'm down. Let's do it.
Yohann: Okay. So to begin with, let's talk about Churnkey's goals. Over the next year, walk me through some of the things that Churnkey is planning to achieve.
Scott: Yeah, so we have moved from voluntary churn reduction to involuntary churn or Dunning. And this is, um, you know, an industry that has been pretty static and constant over the past decade. And we're excited to take what we've learned from cancel flows and voluntary churn and apply it to that.
And we're still learning about Dunning and how to be AI-centric and data-driven and be very robust and dynamic here. But we think there's still a lot to do on the data side. We were very surprised about the amount of data you can collect from a simple... a very simple flow, right? So, it's why people are leaving your product, it's the plan they were on, how long they've been with you, the qualitative feedback they give you, all these sorts of data points that mix billing and activity and invoice payment, geography. You can build an incredible profile and a timeline of who your customer is, how they interact with your product, what they love, dislike, what their patterns are. And this isn't intended to be creepy, it's intended to be useful to both parties. So if I can build a profile of who my best customers are, I can serve them better, I can identify them more quickly and help them out. Maybe it's achieving a price equilibrium that, that makes more sense for both of us. Maybe it's recognizing seasonality in their product usage and their lives. And that's really rich, unique stuff that you can apply to certain cohorts. On the flip side, I can identify customers that maybe just aren't clicking, aren't worth going after, aren't worth paying acquisition costs for, and that makes you more efficient and more focused and maybe opens the door to more audiences that make more sense.
And so, all these things come together into understanding how to release product features and maybe new product lines that may be more beneficial to your company. And also helping you be more proactive about product development. So, if there's a backlog of stuff you've had, maybe you re reexamine it based on the data you're getting.
And I would encourage every product team, customer success team, marketing team, at minimum get a customer survey in place when people are canceling their subscription to understand and start collecting that data, why people want to leave. Not every breakup is negative, we have a lot of people who leave glowing reviews. And I'm talking about my customers as customers here, which is kind of a weird inception, leveling down, right?
So, it's really fun to, to examine the qualitative side. And we have an AI which is actually an AI, it's not fake. That's-
Yohann: That's become a thing you have to say now in 2023.
Scott: I know, "I promise it's real."
But we're analyzing sentiment, we're analyzing types of feedback, right? So we're auto categorizing at scale. Is this a testimonial? Is it a chargeback? Oops, gets scary. That could be a lot. Or is it, straight up a feature request and these are treasure troves of ideas that you don't have to do the work for in the CSV. So that's my long-winded answer to, to what's coming soon.
Yohann: Okay. Just for context, before I dive into that, is Churnkey a Self-Serve company? Like, how have you guys set things up? Is there a sales team and a bit of Self-Serve going on at the same time?
Scott: Yeah, depending on when you asked me that question, it would've been different because we've evolved and shifted and contracted and expanded various cases. So-
Yohann: Okay.

Why Churnkey Blends Self-Serve and Sales

Scott: We thought initially it would be a pure Self-Serve product at a relatively low price point. And what we discovered was that worked for some people, maybe starting out, or were maybe solopreneurs or indie hackers. But going beyond that, there were bigger teams who needed to... had opinions on various aspects of the implementation, had different credentials authenticating Stripe or Chargebee, for example, who had to implement the whole thing. And so we quickly realized that we were consulting all the time for bigger teams, and our price points didn't reflect that.
It didn't value our time.
And so, we quickly created a hybrid flow where basically were saying, look, "if you have the means to implement right now, and you're motivated,"... and not every case is cut and dry, right? We've had literally, Unicorns sign up on the phone. "Hey, we're sold. Let's do it. We've already implemented it right now, so can we go live?"
We prioritize and kind of grease the skids for Self-Serve, but what we found was that for that to be successful in our context... I spoke with Samuel, um, a couple weeks ago and he likened Self-Serve to a vending machine. And I won't go deep into the metaphorical implications there, but everyone gets it and for that to work for us, there needed to be the expectation of if something went wrong, if I had a question, if I was weighing some decision on implementation or building out my Dunning campaigns or whatnot, there was someone there who knew the best practices who had experience seeing things in their industry and could help immediately.
And so, we quickly built an infrastructure where we could respond quickly, act as consultants, lend knowledge, and there's varying layers of on-demand-ness there, right? You can't be everywhere all at once, to paraphrase an awesome movie. So, just by the promise of that presence, Self-Serve works way better. What I remember is people expecting there to be a week long email response time on implementation questions or whatever, and they just kind of lost interest and lost trust and momentum. But if we showed that we had expertise kind of behind the curtain that was real, they implemented very quickly and got up and running.
Yohann: Thanks so much for the context. I, this is one of the lines I wanted to draw because we are talking about churn and there's a difference between good practice in churn for everyone and the kind of stuff you are doing to battle churn at Churnkey. So, the context of how you've set things up at Churnkey is very helpful.
But before we dive into Churnkey, let's talk about the former. Let's talk about churn practice in general and the insights you have for Self-Serve companies who are trying to figure out churn as a problem because there's so much that you talked about in the beginning that I want to dig into from the ideal customer profile to how you find out which customers aren't a good fit to measuring in cohorts to customer surveys... I love all of it.
To begin with, I just wanna appreciate the fact that you mentioned billing along with all of the other stuff that people do inside a product to find value. A lot of people see that as two separate things. There's the "find value" and then there's the billing, which is someone else's problem. Or maybe it's your billing systems problem.
But when you think about monetization, it's all just one flow. That really came across when you were talking and I'd love to dig into that as well.
Scott: Awesome.

Creating Customer Profiles with Churn Data

Yohann: I'm wondering where to even begin. Um, Let's start with customer profiles and how you recommend building these out. What is the approach you take when you say you're looking at patterns to build out a customer profile? How exactly do you do that?
Scott: Sure, yeah. Right now it starts with the feedback that customers are giving to our customers. Again, inception.. "customer's customers," it's all kind of confusing sometimes. For us, what we've learned, how we've structured things, that's where it, it starts and it is: what are the reasons that people cancel? And how are people responding to, for lack of a more sophisticated term, Dunning campaigns. And how are they reacting to all those operations? And those are, what we call Retention Operations and those, To be clear, like those are still, you could maybe consider those towards the end of a lifecycle, right?
If people are canceling, there's an intent there. If people don't care to update their payment info, there may be an intent there.
Yohann: Right, right.
Scott: And that's a reality you have to face, right?
Yohann: Yeah.
Scott: So we, we start with responses and it's quantitative, qualitative, and "are you updating your payment info? Do you value the product enough to want to keep paying for it?"
And, I always talk about how Dunning is really awkward. It's you're subscribed to a product that by all intentions you want to pay for, but you're being reminded that you're behind on payments and you gotta take an action.
And that's really annoying because it's something that is no fault of your own.
It may be becoming more common with pre-pay cards and virtual cards and there's all kinds of legislation around this coming out throughout the world. And, I think in India in particular, if a subscription is above a certain threshold, you have to approve it every month.
Yohann: Yeah. Very annoying.
Scott: Yeah. I bet for everyone, and maybe in certain cases that it isn't annoying, you're really thankful that you had to do that, cuz you're like, "wait, I don't remember . Signing up for that."
Yohann: Yep, that's true as well, yeah.
Scott: so there are benefits to it, but maybe, in the 80/20 rule, most of the time it's really frustrating.
So, it's accommodating all these edge cases, use cases, and taking a signal from it and rolling that up into "who are the people who are just gonna go all out for you?" And I mentioned it before, even if they leave, they still love you. So what's that customer? How did you get them? What was their price point? How long were they subscribed? What did they, what was their customer support history, if you want to connect your systems up? Do they have any parting advice? It's the equivalent of somebody in your life who was a great friend and maybe they move away and, or they go to a different school, it's someone who leaves your life by no fault of either of you. And then there's various different profiles of just people who are angry at you, that's the extreme, and you know, then there's everything in between, where someone came in with a very specific use case and then they're out. Someone came in and tried to find something and they didn't find it. And so every chance is an opportunity for improvement. And one of my favorites is actually on the Dunning side where, somebody will open the email, somebody will click the update card button, somebody will go to the form and they don't put in their new card.
Even if Apple, Google Pay, whatever equivalent are enabled, it's so easy. They don't do it. And they do, they open the email and get there multiple times. And we saw this pattern play out over at scale, and we were just like, man, there's gotta be something we can do to make someone deal with the inconvenience of this, right? From the customer's perspective, they're just being inconvenienced. And so we came up with this idea of what we call Dunning offers, it's a very original name I, I admit. So we were thinking, okay like, let's give some people — a customer — some flexibility.
You can discount the current invoice, you can discount a future invoice or invoices or both. And we're finding that there's a statistically significant increase in payment updates when there is an incentive for their time to do it.
Yohann: Okay. Okay.
Scott: That's an interesting segment there as well, right?
Yohann: Yeah. So there are things you can do to one, discover intent and two, to address that intent and make sure people do the inconvenient thing that they have to do towards the end of the customer lifecycle when people are about to churn.
Scott: Right.
Yohann: And when you do figure out this intent, you can go backwards in time and look at all of the things associated with this customer to build out a customer profile of "what are the activities that actually correlate with good customerhood?"
Scott: That is an amazing question because it is different throughout every one of our customers and that's one of the things we're working on, where my dream is... it would be amazing if you could connect your Churnkey data, your billing data, and just immediately understand, "who are my best customers what do they do? Who are the worst customers? What do they do?" And then go further up the process, the activation process, and look at, "okay, where do they come from were they discounted?"
All these crazy things that just stack up over time, and we're working with data teams in bigger companies directly with, like, real time data pipes and just plugging into their data lakes, and they're doing this analysis themselves. But my dream is everybody can do this kind of analysis and it shouldn't be an archeological operation.
Our whole model with Churnkey was... this was an internal tool for a SaaS that one of my partners had running, two of them had running, and their churn was awful and they tried everything with consultants and this and that and better onboarding copy and whatever. And then we both were talking about it and had this same idea at once and I was like, "Hey, I just tried to cancel Audible and they had this really weird flow that I had never seen before. And it was saying, stay on for these benefits." And my partner was like, "wow, man, I just tried to cancel Hulu and I'm still a customer cuz I paused it and I went through a similar flow."
And so, all that is to say that our pattern is taking tactics and strategies from big tech companies and making those user-centric, customer-friendly and democratizing them in a way. That's what we love to do. I hope to bring that to an understanding of data and of billing which you pointed out all the way to activity and histories and all that.
Yohann: And outcomes as well. So thinking about what the people who are leaving came to do, and did they do it? Are they leaving because they did it and now they're done and there's no need to use the product anymore, or they couldn't figure out how to do the thing that they came here to do? Or, if those are a spectrum, all of the options in between.
Scott: Right on. I mean, that's one of the biggest things we love to uncover because it can be a mystery unless you have robust, flow analysis in place and all these things. But even if you're just, even if you're asking why you are leaving at the end of the process, it's like breadcrumbs and whatever fairytale that breadcrumbs are in, I forget what that story is, but.
Yohann: Hansel and Gretel-
Scott: -they're all... yes. I'm ashamed of myself. Everything is a clue. Everything is a data point. Everything is something to consider whether or not it is a path to the truth or not, but it's better than nothing. And man, one of my, excuse me, one of my biggest frustrations is when someone cancels and they say, "oh, it didn't work." Or "I didn't get what I wanted." Or whatever. And you look at their usage patterns and their history, their activity, and it's like they, they skirted around, around the features that would've gotten them to the promise.
Yohann: Damn.
Scott: And that's so frustrating cuz it's... you were right there.
And so, obviously something was lost in translation. So I love identifying that, going through the pain, and then uncovering how do we fix this? Right? And so that's one thing that, that these flows can unlock.

Using Churn to Refine Value Metrics

Yohann: Yeah, and these are such actionable insights. We've covered the customer profile and I think it's fantastic, but another area that I want to ask you about is the qual/quant distinction. So at the end, when you're asking people with a survey, for example, why they're leaving, and you get that qualitative data, you can find some patterns there over time and work on it. But what does the quantitative side of battling churn at the end with Revenue Operations look like?
Scott: Yeah it's the most powerful when you can segment-
Yohann: Right, I was just going to ask, are you looking at customers and cohorts?
Scott: Yeah, for sure. It's even more effective when companies can pass their key value metrics in.
Yohann: Okay.
Scott: And we can take that and really tailor everything from copy to offers to yeah, just all those things to to speak directly to the customer and say, "Hey, you've been with us for three years. You've made," I don't know, thinking of an example here, "you made 25 videos per month. And if that's something that's meaningful to your business, you better be sure it's gonna be meaningful in this process, in the off-boarding process."
And it has the net effect of one, speaking to either your worst or best, or in-between cohorts. Two, it's humanizing because you recognize, " hey, I see you," in a positive way, not a creepy way. "You've gotten value out of this, or maybe you haven't gotten value out of this. Let's find something that works for both of us. We would love you to stay around. We have value to provide. Here is how we can provide it."
So enabling that for every company that comes on is really important to us. And we think it's pretty unique in what's out there right now. And honestly, it's one of the most fun aspects of building these flows out for companies is that, "hey, look, I get to learn about your business. We get to talk to your customers at scale directly in a very personalized way. And your customers have a chance to feel recognized and seen and they get to respond, and not feel like they're just some thrown into some algorithm and chewed up or sent to a call center in, I don't know, Arizona and put on hold for 25 minutes, right?
Yohann: Right, right. So when you say key metrics, you're talking about usage metrics. Some activity inside the product that the company has identified is valuable to users.
Scott: Right. And, um, when I was talking with Samuel the other day he joked how this could be misconstrued, right? So, "oh, if someone makes three projects, they're gonna be successful, right?" The way we talk about this in SaaS is interesting because most of it's derived from crazy at-scale, viral social media companies. And the myth is still seeped in that too, right? Yeah. And I've heard you talk about this too, where, "wait, we're not adding four friends, it's completely different. We're taught that these actions are meaningful, but in reality, they're relatively meaningless in the world of B2B where there are many people who have input, there are legitimate actions that need to be taken. Maybe it's installing a script, maybe it's connecting a Stripe or Chargebee or Braintree. Those are significant steps to take. We do leave it up to every company though to add their, to know their key value metric, to push those through to us. And in some cases that might be asking a lot, right? Maybe they haven't figured it out.
Yohann: I'm realizing that there are two layers here. There's the layer of providing the right information to customers who are just about to leave, to either get them to stay or learn why they're leaving. And the other layer is at scale, learning about how you can acquire better customers and how you can connect your existing customers with value so that they don't even end up in this position to begin with.
Scott: Yeah. Yeah. I mean there are so many layers to this and it can make you your head spin with the amount of questions that arise, right?

Applying End-Of-Lifecycle Lessons to the Early Lifecycle

Yohann: Yeah. We've covered one of the layers. We've covered when someone wants to leave, how can you talk to them in a more humanistic way? How can you quantify the relationship that you've built together so that they're less likely to leave. How can you offer them a discount if price point is an issue? But I'm very excited to dig into the other layer in a little more depth as well. We've covered that in the sense of the customer profile and the quantitative insights. Like, you can learn things about the kind of customers you're getting from looking at the patterns that emerge when people are leaving. But just digging into that a little bit more, how do you take your churn cohorts and turn them into better retention? Does that make sense as a question?
Scott: Yeah. It's essentially backing out of the worst case scenario and moving, kind of jumping, to the beginning of the relationship, and giving yourself a new opportunity to make a better impression.
Yohann: Right, right. How does that typically work? Jumping from the end of the journey to the beginning?
Scott: Yeah, in practicality, it starts with the questions you're asking and if you're asking the right questions and you're not vague. So, so a lot of times, out of the box, we'll find, and we do kind of a sweep of our customer exit surveys, and there might be questions that are catch-alls where, I don't know, one, one that comes to mind is the answer "no longer needed." And, that can mean a lot of things. And we want to help them get as specific as possible about what went wrong in this relationship. It starts with, "okay, so we have some responses to the sentiment of our customer base, the ones who decided to leave. And then maybe you can compare that if you're doing customer research or you're doing NPS surveys or... you can kind of compare notes there and say, okay, the happiest people, where did they come from? What did they have in common? The people who canceled what did they have in common? Where did they come from? And then you can say, okay the people who left us, let's break that out. So there's technical issues, there were bugs, okay, those are things that we can bring up to engineering and are they legitimate? You can even go deeper and incentivize some people who left due to bugs to talk to you about what exactly happened. So that's a bucket.
Yohann: Yeah. Yeah. I love that idea.
Scott: Those are just kind of table stakes. The ones you wanna look for are like the misalignment in expectation, which means I wanted to do X, but you gave me Z. Or I wanted to do X and it worked out some of the time, but it wasn't consistent.
There could also be this phenomenon of misperceived value where you are providing it but you're not stating it in a way that is practical and easy to uncover and understand. And to be honest, that's something we struggled with in the beginning and that's why I bring it up, cuz it's near and dear to my heart, where our value metrics, one of our primary value metrics is boosted revenue and boosted revenue goes all the way down to the invoice level. And it is the customers who would not have paid you are paying you now because of us and we're measuring that.
Yohann: Wow. Okay.
Scott: And this pretty crazy stuff, I'm not a data guy, but it sounds really impressive, at least to me. But you know, what's your, what's the value metric there, right, that you're missing out on? Because you are delivering value, but no one's seeing it. So those are some takeaways on that front.

Churnkey's North Star Value Metric

Yohann: Tell me a little bit about how you came up with the value metric of Boosted Revenue. I see, I mean, it's very obvious, to even someone from the outside, like me, how this metric aligns what users are hoping to get out of Churnkey and what Churnkey is hoping to do for users. What's the story behind Boosted Revenue?
Scott: Yeah, this is one of many areas that we wanna show value and churn rate to me is a good metric to track, but it's also very nebulous. You can slice and dice it and cut it up in various different ways. But to me, the real effect comes down to how much more revenue do we have this month that we wouldn't otherwise have.
And that becomes really real to me because that's a salary, that's a marketing budget, right? And so that's one of our goals with metrics is to humanize it and personify it. How practical can we be? And so that's when we settled on early on, after um, shuffling in the dark and trying to figure it out.
We were like, "oh, well LTV was extended by 20%" but what does that mean on a massive scale, right? Like, how do I interpret that for my business? So that was one. And the other is our insights. There's the financial aspect, but there's also the benefit of the feedback. And giving that to the right people. So Customer Success, Product, even Engineering for certain cases. A lot of times feedback gets filtered through support and it's very piecemeal and it is usually negative. It can create a perception of, "wow, we, we are really terrible at what we do." Or in some cases maybe there's massive praise for response to a problem. But then there's still the problem and there's still that context of, "we're still behind."
So we wanted to highlight feedback in a way that was more neutral. It wasn't gathered by really any party. And yes, you could argue it's at the point of cancellation. They're already in a certain mindset. But what we found is that they're not.
Yes, you get angry cancellers and that's fine. That's always gonna be a part of life, right? Even in when I built consumer products that were free, that were video products, I got angry feedback that was very angry. And that's what we call the internet, right? So it, it's the monetary side and it's the product improvement engine that we try and enable.
And we have teams who, like I said, pipe this into their data lake and they make it a part of their weekly cadence of their team meeting where they are looking at the raw data from Churnkey, the reasons people are canceling — are their week over week differences? They're trying to find the patterns and how they can get better, and that's ultimately what we want to do.
Yohann: Right, right. I love that what you do for your customers' customers kind of mirrors what you do for your customers in terms of the two layers here, which are one, getting fewer people to cancel at that moment of cancellation, which results in real revenue. And the other layer being taking that information you have at the end of the journey and finding ways to apply it to the beginning of the journey so that you've got better retention across the customer base so fewer people make it to this cancellation point.
Scott: Yeah I like how you put that in the sense that, this kind of just explodes out into all parts of the funnel, and, and like along every path that, that teams are working to... ultimately, just want to have happy customers. You wanna provide the value worth it being paid for and you want to figure out where you're falling short. And that's what we try and do.
Yohann: Right. That's my biggest takeaway actually from this conversation is that churn can be your biggest learning opportunity if you approach it in the right way.
Scott: Yes. And I love that you put it that way. This is one of the biggest challenges we had starting out where it was a combination, I think, two factors where one, some operators and founders just see churn as quote unquote bad across the board. I mean, and it is it's painful. You don't wanna be rejected. No one does.
Yohann: Yeah.
Scott: And then the second bit was this is a dark pattern. You are making it harder for people to cancel something they legitimately don't want anymore. And any obstacle, perceived or real, I would say real obstacles are bad, but perceived obstacles are also bad.
And it's not like we're, I made a joke today and this is a persistent joke internally, where there was some meme website where the cancel button would just run away from your mouse cursor and, you chase it all around the screen, like never click it, right?
Yohann: Gosh, can you imagine?
Scott: I mean, it would, it would be kind of funny for a while, right? You would definitely remember it. But a real life equivalent for me was, I live in Los Angeles, I went to a flagship LA Fitness when I was, I dunno, when I first moved here, did the whole song and dance with, "oh, we'll give you a discount to sign up, but you have to sign a contract and all these things." And I'm just thinking, dude, all I wanna do is just show up at your gym and I'll pay you 30 bucks a month. But when I moved, to cancel I had to prove that I was moving,
Yohann: Oh gosh.
Scott: -which is very invasive. But here's the kicker. I had to fax the proof and the cancellation form to another time zone within a two-hour window.
Yohann: Whoa.
Scott: Are they turning off the fax machine after that time? And who has a fax machine anyway, right?
Yohann: Yeah, this is real life Cancel button running away from you.
Scott: Exactly. Exactly. And so, obviously I feel a lot of pain being compared to that sort of thing, right? So we've always been and I've said this before on, on here, but user-centric, customer-friendly, everything is meant to get people the answers they want. Maybe they want a new price, maybe the company needs to realize that they need to offer a new price. And both data points are very valuable. And that's just one example.
Yohann: Yeah, I totally understand what you're saying. It's not about coercing people to stay, it's about one, understanding why they want to leave, and two, if there are any misunderstandings in that relationship you want to clear it up in case the misunderstanding is causing the leaving to begin
Scott: For sure. And someone listening might be like "that's desperate" and "you should know already." But products change all the time. Customer sentiment, as we're seeing with the economy. Layoffs, all these things, interest rates, they change over time. Maybe you expand into a new locale and you don't know what you're doing. And churn's terrible. You need to know what you're doing very quickly, right?
Yohann: Yeah. And to the person who says, you should know already, I would say you don't, which is why it's so important to learn from churn so that you can stop misunderstandings like this from happening at all. This is how you learn to "know already."
Scott: Yeah. I like how you put that because churn is, I think, one of the most underrated sources of improvement and intelligence. And where else do we learn as humans if it's not through failure? If we categorize churn as failure, which I think for the most part we, we can all agree that's generally maybe the category. We learn from our mistakes. We learn from what doesn't go right and we improve, and that's just part of life and the cycle. And there need to be some... some structures around that to, to make it more useful.
Yohann: So I can't believe we're still covering all of the concepts you laid out in the beginning, but just to bring up one more. You mentioned that churn is also a way of finding customers that aren't a good fit. I think a good way to broach the subject would be how this has worked at Churnkey.
So at Churnkey, how did you discover which customers weren't really a good fit? And I think this is particularly relevant because you went from pure Self-Serve to a hybrid model, and balancing both of those meant that you would have to have made some hard decisions about which customers are good fit and which are not.
So I'm assuming you used churn data to come up with a profile of which customers aren't a good fit. And I'm curious about what happened next? How did you use that data to make things more efficient at Churnkey?
Scott: So, I'm going to be really annoying and tell you that we have negative churn-
Yeah, a nd that our net revenue retention is, in the triple digits. And one reason I think is because we had the benefit and the luxury, beginning our first customer acquisition operations by pursuing founders who had the same struggles we were and were operating at similar scales or slightly higher.
And we've taken that profile and expanded it to SaaS operators and realizing that all the things we do are broken up at bigger companies. And it's pretty easy just to talk to that. But I will tell you, literally I can think of four examples over the past two, three years.
It came down to perceived value or an implementation that just didn't ever get done. And so you could argue that the latter goes back to perceived value that you, they didn't prioritize implementation, right?
So, I'm very sensitive to that by communicating on an ongoing basis what we are doing for you when you're sleeping, when you're working, when you're thinking about us, when you're not thinking about us, when you're logged in, when you're not logged in, we are always there protecting your subscriptions and your revenue and also your customers in a way, because we're looking out for them. We're looking at how they can be better served too. And doing that in a way that is meaningful, not intrusive, positive, beneficial, for me has been an ongoing struggle, but it's something we can definitely do better.
Yohann: Okay. I have a different way to word the question, and if you'd prefer to talk about this, not in Churnkey's context, but you know, like, churnkey's customer's context, that's something we can do as well. But the different way to word the question is, did coming up with a "not-ideal" customer profile as opposed to an ideal customer profile, did coming up with a not-ideal customer profile help you get to negative churn?
Scott: I am trying to think back to those conversations and um, we weren't deliberately saying, " these people aren't our ideal customer."

Key Questions to Ask of Your Churn Data

Yohann: Okay, what do you recommend people do with the information that's coming out of churn that gives them some indications about who they should be going after versus who they shouldn't? Because, I can see how that would make things more efficient, not to niche down, but really to speak to the people who can get the most value out of your product. I'm just wondering if you have any insight into how that process takes place, going from churn data to more efficient operations, and messaging, and so on.
Scott: I would like to know, were people who churn the most, the cohorts, were they promised a common thing? Were they promised it at a certain price point? Did they all come from a specific source or similar sources? Did they sign up at the same time? Was there a spike from, and this goes back to the source thing, but you know, was there some effort that attracted people with similar characteristics? Then you can go further, okay, so this profile, do they have particular traits that would make them price sensitive or sensitive to things like... maybe it's an enterprise plan and teammates aren't using it, and then they leave. So what are those value metrics that are uniquely yours that you can carry through and kind of connect a through line. So those are some initial things I would start out with, if that helps.
Yohann: Yeah, that does, that, does that. That kind of puts a "how" to this whole thing. So it goes back to customer profiles, I guess. When you're building customer profiles, you look at indicators that make a good customer and indicators that make a not so good customer and you go after acquiring more of the good fit customers, essentially.
Scott: Yeah, exactly. I mean, it sounds really basic but unraveling across departments and data sources and silos, it, it can actually be kind of hard. Sometimes you don't have access to, let's say the invoice history. That can be really important because they were, let's say they were discounted the first six months, and then they're at full price and then they're not as willing to pay because their perceived value doesn't match what they're paying, right?
Are you taking customers for granted at certain price points, or are you relying too heavily on discounting initially? There are all these assumptions that can easily... or all these tactics and strategies that can easily get lost across teams and departments and whatnot.
Yohann: Right, right. Those core fundamental assumptions, the more time you spend with those, the better all of the execution stuff that happens later is going to be, and that execution stuff happens either way, whether you question the core fundamentals or not.
I think we've covered so much important territory. Just from how to approach churn and learn from it to some of the insights you've given us on how you approach churn at Churnkey. I think both have been really valuable and I think anyone listening could get a lot out of either.
Scott: You're a generous host. That's very kind of you. I will take the win. I appreciate it, appreciate being on. So it's been awesome.
Yohann: Thanks so much, Scott. Thanks so much for joining us.
Scott: Great to be here. Appreciate it.

Episode Discussion

Have thoughts to share? Join the episode discussion here.


Written by

Yohann Kunders
Yohann Kunders

Co-founder: Self-Serve SaaS, prev Airbase and Chargebee