Failed payments can cost a business in many ways, from both
loss of income to time in chasing the payment. As many as 15 per
cent of transactions fail across all payment mechanisms. To help
combat this, GoCardless launched a new payments intelligence tool
called Success+, which helps credit control improve their payment
The tool uses machine learning from nine years of payment
behaviours gathered by GoCardless insights to analyse trends from
recurring payments data. It not only helps to identify the ideal
time to retry a failed payment but then automatically schedules
payment retries on an optimal day for each payer.
Until now, it has been difficult for merchants to know when is
the best time to retry a failed payment. This has meant businesses
have relied on retries that take place on random dates, often with
limited success. Using previous insights,
the Success+ tool allows businesses to find the optimum day for
payments via merchant data and takes into account the industry of
the business and payer data – such as a history of failed
payments. This will also let a business know if the tool thinks
that a payment is very likely to fail, so it can decide if it’s
worth retrying the payment anyway or handling the situation
Out of the 2,000 GoCardless customers that trialed Success+, 15
per cent are seeing a decrease in payment failures and 63 per cent
have seen an improvement in cashflow. With the payment tool taking
care of recovering the failed payments, businesses are saving time
and can focus on the day-to-day running of their business and
maintaining their customer relations.
So, are payments the next industry that can rely on machine
learning for insights? Duncan Barrigan, Chief
Product Officer at GoCardless thinks so. He said, “Failed
payments is a widespread business issue. They are costly – both
in time and money, can impact customer relationships and increase
the risk of bad debt and customer churn. We created Success+ to
“An interesting insight we spotted when looking at the data is
that many specific factors need to be taken into account to get to
the best results. For example, the industry of the business taking
the payment and the country of residence of the customer making the
payment can make a meaningful difference.
“We found a significant proportion of customers are willing to
remake a payment that has failed and the failure is down to
something that has gone wrong temporarily, so it’s important to
give them the chance to address it. Finally, we had some
fascinating insights in our wider research into how customers get
paid such as amazingly, 11% of companies in the UK operate a lunar
12-months ago, the smart home-insurance provider
Neos became one of the first businesses to test
GoCardless’ payment intelligence product, Success+.
The company has a current failure rate of between 4 and 5%, and
most of these are due to customers having insufficient funds in
their bank account. But the company is impressed with the tool.
Head of finance, Monsur Alam at Neos said: “Before using
Success+, we had one person spending 60% of their time on credit
control. By automatically retrying the payments, that’s now down
to 20% of their time, as it’s far more likely the payment will be
collected before we need to intervene.
“By automatically retying failed payments, it gives our
customers the time to make sure there is money in their account.
Because of this, almost 90% of retried payments are now
successfully collected, where before this was 30-40%.”
The Success+ tool is currently focussed on using macro trends,
merchant, and payer data to figure out the optimal retry schedules
for individual payments and customers; although given it is
operated by machine learning, this analysis doesn’t directly
produce high-level insights into trends over time.
This is something that the company hopes to improve in the
future. Barrigan added, “As Success+ continues, GoCardless will
continue to look for ways to help businesses learn more about
customer behaviour, payment performance and how to reduce failed
GoCardless Historical Data Helps Recover Failed Payments Through
Machine Learning appeared first on The Fintech Times.