Guides and reports

Leveraging the payments community to make subscriptions unstoppable

We built Auto Rescue with a little help from our friends at Pinterest, Babbel, and GoHenry.

Gill Wells  ·  Adyen
4 December, 2020
 ·  6 minutes
Leveraging the payments community to make subscriptions unstoppable

Most of my purchases used to be one-offs. From the quiet suburbs of Harare where I spent my youth trading bricks of grubby Zimbabwean dollars with street hawkers for the latest blockbuster DVD, to London life in my twenties where I experimented with DVD delivery from Lovefilm. It’s a far cry from life today. Now, most of my purchases are automated and the service uninterrupted. My music comes from Spotify, films from one of three streaming services, and groceries from Hello Fresh. I don’t even own my bicycle (essential for life in Amsterdam), I pay for its rental and service fromSwapfiets.

As someone who is so reliant on subscriptions in my day-to-day life, it's one of the joys of my job that I get to spend the vast majority of my time helping businesses to keep the revenue rolling in.

Growth of the subscription economy

No other trend has changed how we shop more than the subscription economy. Once the preserve of magazine aficionados and milk guzzlers, subscriptions are now available for everything from scuba gear to sunglasses, software to shaving kits. The subscription economy has grown morethan 300%in the last seven years alone.2019 was the first yearwhere more Americans actually streamed more online content than they watched traditional TV. The coronavirus has accelerated this growth, with18% of companiesseeing their subscription growth rate accelerate.

The challenges for subscription businesses

Subscriptions are affordable because they assume a long customer lifetime. If customers decide to cancel, merchants find themselves running an unprofitable business. Speak with any subscription merchant and they'll tell you failed subscription renewals are the bane of their existence. Transactions fail for a range of reasons but the number one reason is insufficient funds. In fact, 72% of declined transactions are due to insufficient funds or ‘do not honor’ codes. Remember, renewal of your product or service isn’t the only payment coming out of your customer's bank account each month. On top of their numerous subscriptions, they may have rent or a mortgage to pay, as well as direct debits for utilities.

Innovation doesn’t happen in isolation, it takes a village to optimize payments systems. When developing new features we work with some of our key merchants to better understand their issues and what we can do to address them. As co-chair and co-founder of the Women in Payments and Fraud chapter of the Merchant Risk Council, I work to support, promote, educate, and inspire other women to pursue careers in payments and fraud. One of the big advantages of this role is that it has introduced me to an incredible community of payments experts who I can rely on for everything from career advice to sound boarding new product features in development.

While we were building a new feature to address the challenge of failed subscription renewals, I reached out to three of my fellow members of this community:

  • Karla Hensley | Senior Payments and Fraud Manager | Pinterest
  • Christine Maaß | Director Infra, Platform, and Payments | TeamViewer AG (Former Head of Payments Babbel)
  • Ruth Vero | Head of Customer Operations | GoHenry

Their feedback helped define what needed to be considered to build a successful retry logic that could solve insufficient fund declines on subscriptions renewals.

Building retry logic: Three focus areas

1. Let data guide you

The more data you have about your transactions, the more you can learn from the patterns they form. For example, understanding why your renewals fail is the first step to creating a process to ensure that the frequency of their occurrence is reduced. A blanket approach to retry logic, i.e. the same day every month, is the easiest option but has limited effectiveness. Using your data to formulate a more targeted approach will take more time but will also increase recovery rates.

Using your own data is useful, but benefitting from ours is better. We analyze millions of transactions every day, meaning our merchants can benefit from our learnings from transaction patterns of thousands of companies.

At Pinterest we know how important it is to factor in payday into our retry logic. Using data that helps us take a country by country approach to determine when to bill our users is really important in reducing customer churn.


2. Allocating resources

Approaching, analyzing, and acting upon the patterns you see in your data can be a daunting, labor intensive task. Take dunning for example, one of the ways to solve for involuntary churn. Dunning is the process of creating a methodical means of communicating with customers to ensure payment. This could be in the form of a text message or email to remind them to have funds in their account ahead of your attempt to authorize a payment. If 10% of transactions fail, a sophisticated dunning approach may save up to half of these failures.

However, dunning is an unsophisticated method that can involve a lot of labor-hours to implement. For many growing businesses, it comes down to making a strategic decision about how they want to allocate their precious resources. Outsourcing the complexity involved with retry logic can save you time and money and allow your team to concentrate on other business objectives.

When it came to the collection of our monthly subscription fees, our biggest challenge was the resource to remind customers when their subscription was due. We’re a young company and have tried hard to keep our operation lean.


3. Create the right business rules

Taking a targeted approach to retry logic, doesn’t need to be rocket science but there are some considerations to take into account:

A.Factoring in decline reasonsand then experimenting with rules such as:

  • Only retrying on soft declines (when a payment is declined but retrying it in the future might recover it e.g. insufficient funds, temporary technical failures - as opposed to hard declines which can never be recovered with retries e.g. card account closed).
  • Differentiating the number of retry attempts by decline reason (e.g. for decline reasons where the recovery rate might be 0% on the 3rd retry attempt onwards, limit your logic to 2 retries).
  • Timing between retry attempts by decline. For example, allow more time between retries where an insufficient funds decline has been received.

B.Splitting logic by market- Factors such as the time of day and day of the month can have a significant impact on transaction success rates. In the US for example, many customers are paid every two weeks, typically on a Friday, with an uplift of successful payments at the beginning and middle of each month. In the UK, people tend to be paid monthly, with one spike towards the end of the month.

If you treat every decline differently, you’ll need fewer retries which means lower costs as well as compliance with scheme rules. Visa, for example, currently mandates four retries over a 16 day period, without sacrificing lost orders. By leveraging our data and knowledge you’ll never again have to worry about decisions such as the best day to take payment, no matter what country your customers are in.

Retries need to be invisible and if your retry logic isn’t extremely accurate it can reduce the chances of authorization and also lead to increased fees from the card networks.

I know how much effort you ideally need to invest to make retry logic really effective.


Leave it to Adyen

We developed Auto Rescue, a new feature of Revenue Accelerate, based on feedback from our customers, to take care of all of this heavy lifting for you. Auto Rescue automatically retries declined shopper-not-present transactions such as subscription renewals and also solves for all of the focus areas mentioned above.

The results, monitored during the pilot period and since launch have been impressive:

  • Karla, Pinterest: Early results point to a 4% improvement in recovery rate on retried transactions. My expectation as we grow is that the recovery rates will continue to improve up to as high as 10%. What does that mean to me? I might be able to offset my entire cost of Gateway/Processor fees with the implementation of this feature.
  • Ruth, GoHenry: Working with a trusted partner, Adyen Auto Rescue provides us with a solution that relieves our challenges. We now have full visibility of the recovery process; it is using machine learning to ensure our revenue collection continuously improves and at a price that makes sense.
  • Christine, Babbel: The implementation effort to use the Auto Rescue feature on our end was low, which made it easy to prioritize within our payment roadmap. Our A/B test results during the pilot phase had been convincing and the uplift statistically significant.

If you’d like to join Babbel, GoHenry, and Pinterest in reducing your churn and optimizing revenue, contact us.

To read about more ways you can boost your subscription business, download the Adyen Guide to Subscriptions.


Fresh insights, straight to your inbox

By submitting your information you confirm that you have read Adyen's Privacy Policy and agree to the use of your data in all Adyen communications.