Birds of a Feather Flock Together

We have always used the people around someone to form a view about them. Who they spend time with and who they associate with tells us something about who they are.

Experienced underwriters do the same. When assessing an MSME, they look beyond the company to its customers, suppliers, investors and promoters. The financials tell you what happened. The network tells you who the business actually is.

Credit companies have tested this idea too. Lenddo used social-network data in its credit scoring around the idea that “who you know matters”. It even ran a “Trusted Network”, where a borrower’s repayment behaviour could affect the scores of people they nominated.

That was consumer lending, where social scoring raised fair-lending concerns. Business lending is different. A company’s customers and suppliers are commercial relationships with invoices behind them.

Yet most credit models still assess borrowers in isolation. And where counterparty analysis is used, it is usually focused on customers: who owes the money, will they pay, and how concentrated is the book? That is only half the picture. A business is held up at both ends. In my experience, problems often start on the sourcing side long before they show up in receivables.

This raised a question for me: can the network around a company tell us something about the company itself?

At Avexo Credit, we decided to test it. We look at both sides of the supply chain, scoring every meaningful counterparty into a single supply chain rating for the borrower, while flagging concentration, adverse media, sanctions and related-party connections.

The interesting part was what the network revealed.

We built the supplier side to answer a going-concern question: can this business keep sourcing what it sells?

It turned out to be just as useful for spotting fraud. Take round-tripping. A company can sell to and buy from the same related party to manufacture turnover. Looking at either side alone may not reveal it. Connect both sides, and the pattern can become obvious.

Doing this manually across every application at consistent depth simply isn’t feasible. That is where automation earns its place: making the same depth of analysis possible on every file, not just the ones someone has time to dig into.

None of this is technically hard. It is tedious. And that is a solvable problem.

A company doesn’t operate in isolation. Neither should the way we assess it.

This reflects how we think about counterparty analysis at Avexo Credit. It is not credit or investment advice regarding any specific company or transaction.