Even slight variations in involuntary churn can significantly impact revenues, so it’s critical to address — and easier than ever. Join this VB Live event to learn about the latest subscription technology, which leverages machine learning to reduce involuntary churn and boost monthly recurring revenue by an average of 9 percent.
Most of the time, the reason that companies bill on a subscription basis or set up subscriptions is for the benefits of recurring revenue, says Emma Clark, director of product at subscription commerce company Recurly. The clear advantage is reliable revenue every month, instead of the one-time transactional model where you’re focused on converting more transactions.
The drawback? A much lower rate of ongoing growth in revenue. And that makes involuntary churn, or the process by which a transaction declines, the major problem that subscription businesses face every month — and one that compounds.
“Involuntary churn is passive,” Clark says. “A lot of times customers don’t even know about it.”
They’ll receive an email from a business that says, your card was declined, but they may ignore it or not see the email in time, and then eventually the service shuts off, or the product stops coming in the mail, and you’ve lost the customer.
“We see that come up and eat away customers,” she says. “And if subscription businesses aren’t managing it correctly, it can eat away into their revenue a lot.”
The trouble is that most subscription businesses are instead focused on their decline rates — the percentage of recurring transactions that are initially declined — which hovers around 13 percent, Clark says. Meanwhile, they forget that preventing those declined transactions from turning into involuntary churn is the key piece.
“The transaction may decline, but because you have this ongoing relationship with the customer, there’s the opportunity to recover it before that decline turns into lost revenue,” she says. “That’s where we see machine learning play a role.”
And with decline management strategies powered by machine learning, subscription businesses have shown on average a 9 percent lift in revenue, Clark says.
It works because every decline is a bit different; every transaction amount may be different; and the products or services differ among companies. Customer profiles vary, as does the profile of each subscription business. There are differences in card types — Visa, Mastercard, American Express — gateways, industries, and more. Some of these factors are under the control of the customer, and some of them are out of their hands.
“All of those different factors turn into millions of different data points,” Clark says. “We use machine learning to help analyze those trends because there are so many different factors involved. It’s almost impossible to lay down a one-size-fits-all model. We use machine learning to analyze all of those different attributes of a declined transaction and then build a strategy to prevent that decline from turning into involuntary churn.”
For instance, a customer of a streaming media service is up for her monthly renewal, and the transaction declines because the bank has put a temporary hold on it, and banks across the world have different rules for how long they keep holds on a credit card.
They’re able to learn from these trends by looking at millions of data points across different types of declines over a long period of time, she says, and can then use those data points to surface the essential insight: for invoices with a similar profile to Emma, with that decline type, a subscription company is really successful when they try again three days after the initial decline.
In other words, they’re using historical data points to predict when it’s best to retry any specific transaction, or what they call the optimized future date for the highest likelihood of success.
They want to do smart retries, meaning they’re not going to retry a transaction every day, because there’s a cost associated with every time you retry a transaction, and that gets expensive for subscription businesses. It’s about using machine learning to optimize retrying the transaction with the bank and through the payment gateway to uncover the smartest opportunity to retry, maximize the likelihood that transaction will succeed, and never interrupt the customer’s service.
“We’ve all had that situation in which we realize that a decline from a subscription, a decline from the bank, turns into some service being turned off, and you didn’t even know about it,” Clark explains. “If a customer has to update information and figure out what’s going on, a lot of times that turns into them saying, it’s just not worth it. We don’t want that.”
Want to know more about how subscription businesses are making a positive impact on revenue with these and other decline management strategies? Join our latest VB Live event and you’ll learn how to start and where, plus get a first look at the latest Revenue Recovery Benchmarks, which reveal the powerful impact of machine learning.
Don’t miss out!
In this VB Live event, you’ll learn...
- The power of dynamic retry logic, optimized for each individual invoice
- The incremental lift that a well-designed dunning strategy can have on revenue
- The key metrics every subscription business should understand to prevent churn
- How to develop a comprehensive decline management and revenue recovery plan using proven strategies for successful transactions
- Emma Clark, Director of Product, Recurly
- Devin Brady, Data Scientist, Recurly
- Stewart Rogers, Analyst-at-Large, VentureBeat
- Rachael Brownell, Moderator, VentureBeat
Sponsored by Recurly