Historically, ProfitWell focused on the middle chunk of a customer’s lifecycle: after they’ve already been acquired, are still a paying customer, and haven’t moved toward ending their subscription. Developing a product that targeted cancellations had always been something that we discussed on the product team, but early on we struggled to justify how it would fit into our product offering at the time, and whether or not we could develop a unique enough vision to stand out in the market and deliver value. The market for cancellation flow products leans toward a red ocean, i.e. it’s an existing market with many competitors — hence our hesitation.

However, there came a point in time when we shifted our perspective to focus on the entire customer’s journey, both because new data came in from one of our core products (Price Intelligently) that suggested there was quite a lot of churn we could tackle using the surplus of data we’ve gathered over the years and because we wanted to make things a little more interesting for ourselves.

Through our research and early iteration, we defined three ultimate goals for the project:

» Balance a smooth, straightforward user experience with enough friction that businesses don’t easily lose customers

» Design a white-labeled widget that can elegantly fit inside any brand, without comprising on style and composition

» Salvage as many customers as possible with curated (& effective!) offers

In order to carve out our own space in the market, it was important to understand where we might stack up against our competitors, what our competitors offered that was highly valuable to their users, and what holes we could possibly fill in and use as leverage to gain traction in the market. We gathered a list of 12 direct and indirect competitors that offered a competing or similar product and measured them against what features and integrations they offered, what their competitive edge might be, how they priced, and who their target customers were.

It quickly became clear that while there were already quite a lot of companies building this type of product, none of them did it really well, and it came down to this insight: everyone has exit surveys, but very few companies also offer robust churn insights and/or curated salvage offers.

For our first round of user research, we wanted to gauge interest from current ProfitWell customers, as they would be the first people to test an early version of this product and the first users we would “sell” it to, i.e. ask them to try it out as a round of beta testers. In total, we polled 13 “friendly” companies (i.e. companies we had strong relationships with, who we could rely on to give honest and thorough feedback) and asked them a series of questions aimed at understanding what their current cancellation flow entailed, what was essential to that flow, what might be missing, and if they would let us take it over through adopting this product. We also showed them a very basic exit survey from an early round of UX exploration and ideation as a pulse check.

Some key takeaways from this round of feedback were:
‍
» Users were, overall, not that excited about exit surveys and churn insights themselves because these are something that anyone and everyone can easily build.

» What was exciting was the idea of offering salvage offers mid-flow based on those insights because not everyone knows how to act on survey responses in an effective and efficient way.

» Most companies ask their customers why they are leaving in an exit survey, but not many have figured out a way to implement and scale salvage offers prior to the point of cancellation.

From that research, we were able to carve out 3 main user personas:

Because the cancellation process can be quite delicate in nature (you’re potentially dealing with a disgruntled customer who has little time and little patience to deal with anything superfluous), the UX had to be quick, painless, and straightforward, while leaving a lasting positive impression on the customer.

While there are quite a few companies building out products like this, no one is doing this really well. I tested many cancellation flow products and explored many cancellation experiences by signing up for brands and services and cancelling my subscriptions with them, and through that testing found that many of the products out their were cumbersome and difficult to navigate in term of their UX, and on the visual side of things were just not that great to look at — it seemed like most of these companies didn’t really care about the user experience of customers who were cancelling, which seemed like a huge mistake and a missed opportunity to leave a lasting impression with a soon-to-be former customer (who might one day come back).

This research helped define my approach in designing a flow that would fill the hole in the market for something simple, sleek, and easy to navigate. Here are some of the bad cancellation flows I found:

Out of the research came three must-haves for the cancellation flow's user experience:

» This experience should be as frictionless as possible, while also acting as a barrier between the customer and the cancel button. There should be no friction for the customer in choosing their reasoning for cancelling, but enough friction between thinking they want to cancel and actually clicking "cancel" that they have time (and a compelling enough offer) to decide to stick around.

» Ultimately, the mission here is to get data on why people are cancelling, so we should be making sure the experience for the customers providing that data is as pleasant and painless as possible. If the customer is already pissed or frustrated or ready to leave, are they willing to do all of this reading to give the company information that likely only helps the company?

» We shouldn't be making them do extra work nor take a test before they leave. We should aim for a limited effort, high reward experience for customers.Since this is potentially (but hopefully not) the customer's last impression of the company, this experience should be extremely delightful. We're putting this in front of unhappy customers, the last thing we want is to further increase their level of frustration.

There’s a layer of complexity here in that this is technically a B2B product that has B2C use cases. We’re selling something to other businesses that ends up getting used by their customers, which creates a unique situation to design for because the incentives between the two groups are not necessarily aligned: customers want to cancel and businesses want them not to cancel. This gets more interesting when you bring in cancellation insights and salvage offers into the mix. Think about a situation in which a user wants to cancel because they think the price is too high — in that situation the user’s incentive is to save money, and the company’s incentive is to prevent that customer from churning; a salvage offer of a 20% discount would align the incentives between the two groups. The question though is, how do you elegantly surface up these offers to the customer?

Early on we defined our ideal flow for the cancellation process in order to help companies easily categorize cancellation reasons, as well as gather as much data as possible without causing so much friction in the cancellation process that their customers get upset. We needed to strike the right balance here between optimizing for our users (the companies) and their users (the customers). In order to do that, with our cancellation flow, not only do we ask the customers why they’re leaving, but we also ask them what they liked about the product/service in order to play off the psychological phenomenon of the nostalgia effect and help better tee the customer up to want to stick around once they’re presented with a salvage offer.

What's effective about the two questions (cancellation reason and incentive to stay) is that you can see what people are unhappy with, but also if they're unhappy with that, are they willing to live with it given something else is satisfied. You will be able to tell what is actually important to a given customer, as well as get one layer deeper into what you can do to prevent future churn. Having this second question will make the actual 'Churn Insights' more compelling.

Cancellation reasons can vary quite a bit across companies — for example, one of our customers, Aisle Planner, wants to know if people cancel just because it's slow wedding season. There are only a few salvage offers we can offer:

» Discount
» Pause subscription
» Swap plans
» Contact support (via calendar link or email)

Beyond those, salvage offers may need to be a lot more curated based on third party context that we do not have. The second question in the survey helps provide companies with the custom context that they want. If companies are going to use our tool to gather information on and stop cancellations, they want to make sure their CS teams are set up for success (aided with as much information as possible) in case we fail to salvage any given customer.

Once we had the foundation of what we wanted to build defined, I dove in deep into the design process. This project is a perfect example of the notion that simple doesn’t necessarily equate to easy. What seems like it should be relatively straightforward and easy to design, sometimes just isn’t. I went through countless iterations of the design on this project, trying to get it just right. It’s a tall order designing something that is generic enough to fit into a lot of different brands, without sacrificing on quality or aesthetics.Those countless iterations took the product in many different directions, which we relentlessly put in front of our customers to get feedback on whether or not it would work for them. This proved to be immensely valuable (maybe as expected) in drilling down into the most simple and minimalistic design I could develop. Below, I’ll walk you through some of the different directions I tried.

At the start of the design process, I defined a few must-haves that I knew would help enhance the effectiveness of the cancellation flow, based on my own experience, feedback from our current customers, and research in the market. These help provide more structure to the UI design of the cancellation flow, and were as follows:

» The options should be as visual as possible, without losing integrity or making things confusing.

» Text should be limited to reduce the amount of work we’re asking the cancelling users to do for us. If they’re already frustrated with the product, we shouldn’t make them do more work.

» A progress bar is necessary. Signaling to the user that there is an end in sight and that there are only a limited number of steps will hopefully ease their frustration level to some extent.

Early on, I thought that making the flow very visual and relying on icons and graphics for the answers would make the flow easier for customers to digest. However, this presented a few challenges in that some of the answers were concepts that were difficult to convey visually with a simple flat image, and this ended up making it incredibly hard to white label the form because the icons would be too distinct.

Another option I worked through was an attempt at making a really simple user experience, also meant to reduce cognitive load for the customer by only asking them very direction questions with two answers to choose from: yes or no.

The last big direction I tried was ultimately the direction we ended up choosing for the final version. I eliminated the icons, reduced the text, and limited the flow to 2 main questions which could lead to 1-2 salvage offers (depending on if the user accepted the first attempt or not), and then led to a feedback form and final confirmation screen.

Ultimately, I also decided to remove the company’s logo f the top of the form because it seemed to indicate that this was not in fact part of the site. Or better explained as an example, if Spotify built their own cancellation flow that lived on would they put the Spotify logo at the top of the form? Probably not.

One of the key aspects of this product was that it would be able to live inside the contexts of many different brands, which meant that it had to be simple and generic enough that it didn’t immediately stand out as different from the context in which it lived. I accomplished this by streamlining the design as much as possible, picking a web-safe commonly used font (Helvetica), and allowing  our customers to individualize the product slightly with custom accent and text colors.

The final result was a clean and simple form that allowed for minimal customization, while sitting seamlessly inside of our customer's products.