DAVID GUSTIN, Chief Strategy Officer, The Interface Financial Group
June 11, 2019
“The phrase ‘it’s better to be lucky than good’ must be one of the most ridiculous homilies ever uttered. In nearly any competitive endeavor, you have to be damned good before luck can be of any use to you at all.” ― Garry Kasparov, “Deep Thinking: Where Machine Intelligence Ends and Human Creativity Begins”
Bob Solomon, former SVP of Network and Financial Solutions for Ariba, has recently been writing a few short blogs on the economics of networks. It’s good stuff, and I recommend reading the short series.
Suffice it to say that many source-to-pay (S2P) networks struggle to monetize their supplier ecosystem, and a few are looking to change things by using their network data to be more innovative with early pay finance, particularly invoice finance.
For those not familiar with invoice finance, there are three stages where it can be done:
- During Invoice submission — as a start, the submitted invoice is a set of instructions sent on a piece of paper (or virtual piece of paper) or electronically by a supplier to a customer with varying amounts of detail about services performed or items purchased (or to be purchased) along with various other details.
- When delivery has been verified, typically the funder confirms with the buyer — “Hey have you received the goods or services performed”?
- When invoice has been approved by the buyer and scheduled for payment based on payment terms.
Risks are very different under the three stages, and they matter greatly:
- Invoices submitted obviously have the highest risk, for example, a fraudulent invoice could be created or mistakes can be made, i.e., double billing, etc.
- Invoices verified still can suffer from what is known by dilution, i.e., where some deduction is made off the invoice value for many reasons. In fact, when factoring organizations lose money, it’s typically not because Sears or Kmart goes bankrupt (they may in fact have credit insurance for this event if they can obtain, or they have long smelled this coming and stop funding these invoices), it’s because of dilution.
- Approved invoices scheduled for payment have the lowest risk, but just because a buyer approves an invoice does not fully guarantee payment will be made 100% on the invoice value. Up until buyers send their payment file to a bank or third party to remit funds via ACH, check, card or wire, they reserve the right to dilute the invoice value. The technical term is “post-confirmation dilution.”
Can you manage the risks?
Source-to-pay and multi-enterprise platforms provide specific functionality (think an e-invoicing suite for suppliers with a portal, an AP automation suite, a supplier relationship management application) or much broader functionality, think e-sourcing, e-procurement, or even product development and supply chain collaboration.
These S2P networks sit on buyer driven data and could have various supplier information such as:
- Annual volume
- Buyer-seller transaction history (i.e., length of doing business)
- Invoices submitted
- Payment (tying back bank payment details)
- History of credit note adjustments, etc.
There is a belief by some S2P executives that the above data can be the driver in doing invoice finance on submitted invoices, particularly if the network has data on Purchase order, invoice and what’s been paid. Their take is we have this historical view on the buyer-seller relationship that we can use to fund before an invoice is approved. The thinking goes if you can tie this data together, together with some third-party data, you can do invoice finance based on submitted invoices. You have the raw material, all you need is the algorithms to assess risk.
Is it better to be lucky than good?
This is where the saying is “it’s better to be lucky than good” comes into play. But you don’t want to be lucky with credit. The zeitgeist is changing around the credit cycle — and most of it is not a positive narrative. Network data, combined with third-party data, still may not be enough to control for the risks of submitted or verified invoices. And network data alone will probably fall far short. When it comes to purchase orders, invoices and what’s been paid, my experience working with networks say this data is either incomplete or not present, and certainly not present over at least one credit cycle.
In Bob Solomon’s piece on networks, he took the view of how many buyers do you support per supplier. As he said, “If a supplier has only one buyer on your network, are you a network to them, or are you a nuisance?” Taking his point from a credit perspective, if you are only seeing 5% or 10% of a supplier’s business, how well do you really know them?
A best practice is to have a fully digital underwriting engine continuously and intelligently interact with the network or platform to manage the risk profile of the suppliers who request supplier finance on the platform. But in addition to network data, the use of fast data to instantaneously assign a cumulative score to a supplier based on factors such as liens, taxes, credit reports, judgments, fraud-threat and compliance helps reduce risk. Also, using drag and drop with financial and bank statements and automating the data extraction, combined with API pulls from accounting systems, helps provide the data needed to properly assess risk. Taken together, this represents a complete picture.
I think Gary Kasparov is right, by the way!