A First Step to Predictive Analytics: Quantifying the Risk of Paying an Invoice on Receipt

Is it possible to quantify the risk of paying an invoice on receipt – before the usual approvals process? Remitia, an upstart in the trade financing and payables market, thinks so. In a blog post, originally published on PaymentEye, Remitia suggests it’s possible to use big data to accurately price the risk associated with paying an invoice on receipt.

The key is peering “deeply into accounts payable data” to understand the type of approval risk an invoice submitted from an existing or new supplier brings, the post notes. Further, “this is a data source that is rarely mined internally or by third parties,” and similar and additional datasets can provide “insights into payment patterns and A/P operational patterns and their impact on working capital.”

The pricing of payment risk for unapproved invoices is actually not new. It’s something factors have done since the advent of the lending model, but with incomplete information based on access to supplier systems and data. But the notion of taking historical payment and invoice files to quantify and group different risks to arrive at the probability that a submitted invoice will ultimately be approved and paid without any modifications takes the concept into the big data era from a buy-side systems perspective.

Whether or not Remitia succeeds with its pricing algorithms is actually of secondary importance to the future of trade financing, at least in my view. What is arguably more important is that someone is finally tying payment analytics to accounts payable and procurement data sets to come up with predictive models – something I suspect we’ll see far more of in the near future.

Imagine, for example, being able to estimate the probability of a purchase order being fulfilled exactly as specified even before issuance, or how small, system-recommended changes to a purchase order could result in a better outcome. Or think about being able to dynamically sub out different payment mechanisms without a user even being aware of it, such as masking a p-card-type payment model with a proprietary one that captures an even larger rebate but goes through an invoice consolidator or prime partner to circumvent the card companies.

The future of predictive analytics around procurement, accounts payable and payments is bright indeed. Perhaps it will start with the simple binary quantification – pay or don’t pay – of paying an invoice on receipt from a supplier based on historical data. But of course, even before this becomes a common practice, the datasets will need to prove that such an outcome is even possible to achieve

Disclosure: Jason Busch is one of the founders and investors in Remitia.

Share on Procurious

Discuss this:

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.