The collapse of the UK construction and outsourcing giant Carillion at the beginning of the year led to around 30,000 small and medium-sized (SME) suppliers being left with millions of pounds worth of unpaid invoices. Carillion had payment terms of 120 days, and this industry-wide issue of slow payments is the focus of UK startup Previse.

As reported by Forbes, Carillion took advantage of a 2012 government policy called the “Supply Chain Finance Scheme". As outlined by David Cameron at the time: "A bank is notified by a large company that an invoice has been approved for payment; the bank is then able to offer a 100 per cent immediate advance to the supplier at lower interest rates, knowing the invoice will ultimately be paid by the large company."

previse1600
© Previse

Everyone's a winner, right? Unfortunately, "within six months of the start of the scheme, Carillion had extended its standard supplier payment terms to 120 days. The “early payment facility” became the only way of obtaining earlier payment," as Forbes reported.

Instead of obtaining early payment of an invoice from Carillion by drawing on their borrowing facility with the banks, suppliers were actually taking out a loan directly themselves under Carillion’s facility, and the banks’ charges were interest on the loans.

Speaking to Techworld, cofounder and CEO of Previse Paul Christensen said: "Carillion highlights the unintended consequences of supply chain finance, which is that is is used to allow a buyer to extend its payments terms, while offering supply chain finance (SCF) so that its suppliers can still get paid in the same time. However SCF is only available to the largest 1% of suppliers."

The solution proposed by Previse is data-driven SCF, allowing a bank to forward the value of an invoice, minus a small charge, on to the supplier as soon as the invoice has been received by the buyer. The bank then recoups the money from the buyer.

This is an issue his cofounder, and chief product officer, David Brown has been grappling with for the best part of two decades now, originally at his startups Oxygen Finance and Remitia, and now with Previse.

"The lightbulb moment was 'it is all in the data'," Christensen said. "So the data is everything in their ERP system. So all payables and invoice information. Then we run a prediction with a high level of accuracy that the big corporate will pay this invoice."

So Previse applies machine learning to analyse this data and provide a score of a corporate buyer’s likelihood to pay the invoice. It's pretty simple regression analysis: using historic data to predict the likelihood of an invoice being fulfilled in the future. Then a threshold can be set and the lender takes on far less risk.

"Previse provides very detailed and precise analytics to enable a funder to underwrite this risk at a very low price. You have to have a funder - because the buyer is not willing to able to pay on day one, so the funder does. It's just like a credit card model," Christensen - a former solicitor and investment banker - explained to Techworld.

This not only helps to avoid the sort of issues seen post-Carillion collapse, but also helps avoid corporates having to take out expensive short term credit - with typical interest rates of over 20% APR - lowering costs for buyers and giving SMEs the confidence to reinvest in their business.

Furthermore: "With today’s technology able to apply machine learning to big data, it is perfectly possible to have all suppliers paid instantly, on the day an invoice is received, without the buyer having to change its terms or bear any risk.

"This has the additional benefit that the risk of buyer bankruptcy can be moved to places that are in a better position to manage, diversify and hedge it, than with SME suppliers. The tools now exist; they need to be used," Christensen wrote in a blog post.

The solution is due to go live this summer after Previse has worked with several clients to ingest historic finance data and applying its algorithm. "We apply our adaptive learning and modelling at the entity level, so specific to each client," Christensen said.

Previously if they wanted faster payment a supplier could use invoice 'factoring' and sacrifice a cut of the invoice as a result. Previse still takes a cut, but by using detailed analytics it can shrink the risk and therefore the cost to the supplier.

The target clients for Previse are big corporates, which Christensen admits "makes sales tricky as most of the benefit is for the SMEs". So Previse focuses on the money. "They get half the discount of the SME, who can either wait a month or get 99% cash on delivery now, so half of that difference goes to corporates."

He also believes that corporates have soft issues when it comes to supply chain finance in that suppliers are unhappy with the terms, leading to strained relationships, and politicians can also often be critical. "So we can fix all of that and pay you a boat load of money," he said.

However he admits that "inertia and bureaucracy are the barriers. What we have done has not been done before so we need to get through that".

Previse raised nearly £3 million in seed funding from angels and VCs from Hambro Perks and Founders Factory, and added chairman of British Land, John Gildersleeve and chairman at Sainsbury's, David Tyler, to its advisory board in February.