Deep Dive: Cancellation Flow Best Practices
My favorite types of projects for any stage of company have the following attributes:
- Doesn’t add a lot of complexity to the product or codebase
- Doesn’t involve a lot of product discovery work
- Fast to ship & can be rolled out incrementally
- Has a solid impact
- Improves the compounding over time
Setting up best practices in your cancellation flow meeting all of these criteria.
Improving it will impact 100% of your existing users (at some point), it doesn’t overly complicate your codebase or product and the best practices are pretty clear.
I’ll get into the details below, but this tactic effectively works like this:
- Understand why your users are canceling
- Ask them in a multiple choice quiz using their own language
- Depending on their reason, make them an offer to see if you can win them back.
Why This Is So Important
The “smart” way of building a product is to focus on the 20% of effort that drives 80% of the value.
While this is good advice in almost all scenarios, I would argue that lowering churn is the exception.
You should be doing every (reasonable & legal) project to fight cancellation as early as you can in you company's history.
Every month that you are missing best practices in this area means that you are churning users you’ll never get back.
Subscription products take so long to grow, that you don’t want it to take any longer than it has to.
A handy rule of thumb to know in the subscription world is that 1 divided by your month/month churn rate will calculate your average months of retention.
So if you churn 20% of your users on average every month, then 1 / .02 = 5. This means your average user will last 5 months.
This is shown in the graph below, which has 2 core takeaways.
- Every time you cut your churn rate in half, you double retention. If you double retention, you effectively double revenue (eventually).
- It becomes harder and harder to decrease churn the lower it gets, however the payoffs get larger and larger.
This shows a business that acquires 100 users in January and shows you how many are left at the end of December based on the churn rate.
While it may “feel” like there is not a huge difference between a 7% churn rate and a 3% churn rate, there is a big difference.
- If you churn 7% of your users every month, you will lose 55% of your users within 12 months
- If you churn 3% of your users every month, you will lose 29% of your users within 12 months.
How This Helps Reduce Churn
Fighting churn breaks down into the following categories:
- Voluntary Churn - these are users who lose interest in your product and cancel it.
- Involuntary Churn - These are technical issues that cause someone to lose access to your product. These are typically payment-based problems, but not always.
In this article, we’re covering voluntary churn, specifically the tactics that are shown to work once someone has already decided to cancel (see here for involuntary churn tactics).
The best way of reducing your churn, in the long run, is to better activate your users earlier in their journey. However, while you are working on that, you need tactics that you can implement now.
You can think of these tactics as a set of band-aids. They don’t fix the core problems, but they do make it better.
If you can fix the core problems, you should. However, you are bleeding users right now and you should still use bandaids while you fix other things.
Distilling Best Practices In This Area
If you look at most major subscription companies, they are all using a set of tactics in this area that works like this.
- Collect the universe of reasons that their users cancel
- Ask the users why they are canceling via a multiple-choice survey
- Depending on what they choose, try to appease their “objection” and defer cancelation
Churnkey, which is a vendor in this space, just released a great benchmarking report on this set of tactics, and the results match what I have seen in my consulting work.
All of the normal caveats apply to benchmarking reports here:
- This comes from their customer base obviously, which has the bias of companies who knew they needed to bring in a tool to lower churn.
- These are “averages” which doesn’t mean it necessarily what you should target
- The effectiveness of this tactic will highly depend on the use case your product is in and why people are leaving
All of that said, however, if you can reduce your churn by ~25% on average with a sprint or two of engineering work, that’s a great project to work on.
Very, very few other projects will be that “easy” to do and improve your business that much.
This chart shows the relative effectiveness of each offer type in Churnkey’s dataset.
This is helpful data to factor in, however, the core thing to keep in mind is that the better that you can actually solve a user’s reason for leaving, the more effective each tactic will be.
When I was at Codecademy, we had a lot of success by letting the user “pause” their paid accounts. This is because learning to program is intense and takes up a lot of your free time.
A common reason that users were cancelation reason for us was that “life got busy”, so allowing users to pause really helped the core problem.
Contrast that with a product like Netflix, where letting users pause might still drop cancelation a bit, but it's really about letting them save money. I would guess that this is less effective.
Case Studies
Note: I am slowly building out a database of great examples of best practices, you can see those here and filter for “cancellation flow”.
CLEAR
CLEAR is a travel product that allows you to go through security in the airport faster. My guess would be that their leading cause of cancellation would be “I am not traveling soon”.
To combat this, they offer to push your next (annual) billing date out 2 months.
When I went to cancel this product, this tactic worked on me. The 2-month extension pushed it past the US Christmas holiday when my family and I were flying to see grandparents.
Youtube TV
YouTube TV basically allows you to watch US cable TV channels through your YouTube account.
They likely face a series of objections from price, to lack of usage (if someone signed up to watch sports or a series that is over), to technical reasons.
Because of that, they use a variety of win-back methods (click the link below to see all the tactics) depending on what the user clicks.
The most interesting part of this flow is the variable length pause. I would guess that this is done for users who are canceling until their favorite sport or TV series is back on.
ZOOM
You probably know what Zoom is. Because of the wide variety of pricing options and their high traffic volume, they are likely testing multiple win-back strategies depending on the tier you are in.
I use a pretty basic Zoom account as I only have a handful of meetings per month.
Zoom likely knows that and is trying to combat the lack of usage by giving me a cheaper, lower-grade offer (which isn’t on their pricing page) and also offering me to pause.
So What Do You Do With This Information?
If you don’t have anything at all setup, I would approach this like this in two steps.
1. Collect the Universe of Reasons that People Cancel
The better that you can understand why people are leaving, the better that you will be able to win those users back.
It's really tempting to just throw up a generic survey in your cancellation flow, however, cancellation data is notoriously inaccurate.
If a user doesn’t see the real reason they are leaving, they will just click a random reason on the way out.
You can either start by giving users an open-ended question (e.g. “Why are you canceling?”) and then start to cluster the answers into categories, or you can interview the users who are leaving.
2. Create a Multiple Choice Quiz from This Data
Pretty self-explanatory. Here are a few things to keep in mind.
Just a note that if your product is built to solve a temporary problem (learning something, building a habit, etc), then remember to include a “happy” reason that people are leaving.
If you are selling them a weight loss subscription and your product worked, which helped your canceling user lose weight, then they need to be able to indicate this in the survey.
Additionally, I’d make sure this data is being stored in a way that’s useful and easy to access.
If this data only feeds a CSV download that most people on your team can’t access, then it's not that helpful. Send it to a dashboarding tool and review it regularly.
3. Make a Win Back Offer Based on the Reason
Depending on the core reason that people are leaving (and again this depends a lot on your product), you get 1 or maybe 2 shots at trying to win them back.
As mentioned above, you’ll have the best results the more that the offer you’re making is positioned to address their core objection.
e.g. if they don’t need your product anymore, they are still not going to need your product at 20% off.
The key to this feature is iterating your way through it and experimenting with the exact offers you make.
