Two of the most important and educational points in a customer’s life cycle are when they decide to pay you, and when they cancel. If they’re switching to another provider, you can learn even more. This critical moment deserves your attention as it’s one of the best ways to learn where you really need to improve your product.

You should recognize from the get-go that people will cancel–it’s part of the deal. While it’s always disappointing when people decide to close their accounts, with the right approach you’ll see that cancellations come with the territory and aren’t worth getting torn up about. In fact, cancellations can represent one of the most tangible ways to get feedback and understand your product’s shortcomings.

People don’t just wake up one morning and decide to cancel their account. It’s usually something that happens over the course of weeks or even months. And there are usually signs that someone will cancel long before they actually do. For example, most of Sifter’s canceled accounts sat relatively unused for around a month beforehand. Of course, by the time an account isn’t being used, the customer has already moved on. The cancellation is just a formality at that point.

To minimize cancellations, don’t wait until after your customers have decided to leave. Trying to retain a customer after they’ve made their decision won’t work. If you want to keep your customers, look at your data to notice if they’re trailing off, then figure out how you can reach out to them before they decide to cancel.

When a customer is ready to cancel, they should be able to do so easily–it needs to be completely painless on their part. While not all customers cancel because of negative feelings, don’t get in their way at this point. Avoid drawn-out forms, approvals, or other impediments. If people want to cancel, let them. The worst anti-pattern I see around this issue is apps forcing people to email support to cancel. Not only is this not secure–anyone could forge someone else’s cancellation request–it places a burden on the customer at the worst possible moment.

I can think of only two rationalizations to take such an approach. The first is sheer laziness: it’s easier to drop a mailto: link on a page than it is to write code that will delete an account. But that doesn’t make sense to me because someone still has to push the buttons. So the code has to be written. The second, more nefarious reason is that companies may want to save the account either by making the person give up, or by having someone talk them into sticking around. As I already mentioned, this isn’t the time to try to save accounts. Neither excuse is reasonable.

The only exception I’ve seen is offering effectively instant cancellation through live chat right in the application. As long as there’s minimal friction and your response time is near instant, it can be a good balance of being helpful without getting too much in the way. Once customers have decided to leave, your best bet is to focus on understanding why.

In my experience, cancellations are more about the customer than the service. Yes, some people cancel because the product is buggy or the support was bad, but most cancel simply because they’re moving on to something different. Not necessarily better or worse–just different. Their needs have evolved, and they need something new.

With Sifter, the number one reason people canceled was because their projects ended and they no longer needed the account. They weren’t switching to a different product, they were just canceling something they no longer used. This is a downside of Sifter being so project-focused, but it’s a scenario we had very little control over. Short of helping teams win more projects, there’s nothing we could have done if they didn’t have enough projects to justify keeping their account.

The only way to understand cancellation is to explicitly ask your customers why they’re leaving. You can’t force them to tell you, but you can make it as easy as possible. Do everything you can to understand why people cancel without placing too much of a burden on them. The best approach is an exit survey. It shouldn’t be mandatory, but make it easy for people to share their thoughts. Too long, and people won’t bother; too short, and the feedback can be misleading.

There’s a balance to find between keeping your exit surveys brief enough so that people will fill them out, while thorough enough to provide useful insights. I iterated on the Sifter cancellation exit survey a fair amount to try to find a happy medium. I’ve found four pieces of information to be most helpful:

  1. How satisfied were you with the product?
  2. Would you use it again in the future?
  3. Would you recommend it to others?
  4. Why did you cancel?

These questions are brief enough to get a feel for whether you should be concerned about the cancellations without placing too great a burden on the person completing the form. Sifter had a few open-ended questions too, but the four above covered most of what we needed. By understanding cancellations, we were able to make the right decisions. Most importantly, it helped me avoid losing sleep at night from thinking that everyone who left hated Sifter.

One strategy to reduce cancellations is to offer customers the option to put their account on hold. In theory, this is a great tool. Having tried it with Sifter, however, I can’t recommend it without some caveats. For a brief time, we allowed customers to put their accounts on hold indefinitely–no questions asked–but we discovered that people simply placed their accounts on hold in lieu of canceling. Of all accounts placed on hold, only 15% returned.

Offering to put accounts on hold did little to reduce cancellations. It merely masked the fact that people were canceling. And that left us with even fewer insights into why those customers left. Offering holds isn’t a bad thing, but it’s important to understand the drawbacks. Indefinite holds without limitations are little more than an alternative to cancellation. It might seem as if those accounts have been saved, but that’s likely not the case. By limiting holds to three months or adding a small monthly holding fee, you can help ensure that customers exercising this option are genuinely putting their accounts on hold and not simply using it as a de facto cancellation.

Churn, Retention, and Reengaging Customers  Des Traynor of Intercom has written a great article about recognizing when a customer’s usage is trailing off and how you can reach out to them to increase your chances of retaining them.