Commercial risk is the risk that the revenue in the plan does not turn into cash. It has two faces that are usually managed by different people: the customer who leaves or buys less, which sales worries about, and the customer who does not pay, which credit control worries about. Both come from the same place, the customer portfolio, and both are measured badly when they are measured one customer at a time.

Concentration: how much rests on a few names

The first question is structural. In a company with 400 million of revenue, the largest customer buys 18% and the top ten buy two thirds. Each of those relationships may be excellent. The portfolio is still fragile.

Share of revenue of the ten largest customers and the cumulative share. The largest customer is 18% of revenue and the top ten 66%. Illustrative company.

Revenue share understates the risk. If the largest customer leaves, the company loses 72 million of revenue but keeps most of its fixed costs: at a 35% contribution margin, EBITDA falls by 25 million, 42% of a 60 million total. Expressed that way, a single relationship carries almost half the profit, and the board should know which customers are in that position, how long their contracts run and what it would take to replace them.

A simple portfolio measure is the Herfindahl-Hirschman index, the sum of the squared revenue shares, the same index competition authorities use to measure market concentration. Tracked over time, it tells whether growth is diversifying the business or deepening its dependence on a few accounts.

Credit: expected loss, not overdue days

The second question is whether customers will pay. Credit teams typically report overdue receivables, which describe the past. Banks measure the future with three numbers, and the same method works for a trade receivable book:

  • Probability of default (PD): the chance that the customer fails to pay within the year, from ratings, financial statements, payment behaviour and sector data.
  • Exposure at default (EAD): what the customer would owe at the time: receivables plus orders already shipped or committed.
  • Loss given default (LGD): the share of the exposure that would not be recovered, after collateral, insurance and recoveries.

Expected loss is the product of the three, and it is also the logic behind the expected credit loss provisions required by IFRS 9. Mapping the main customers on PD and exposure shows where the risk actually sits.

Main customers by probability of default and exposure, with bubbles sized by expected loss at a 55% loss given default. A mid-size distributor with weak finances carries more expected loss than the largest customer. Illustrative company.

The largest exposure is not the largest risk. The biggest customer is a strong company with a low default probability; its expected loss is under 0.1 million. A mid-size distributor with an 11 million exposure and a 6% default probability carries almost four times as much. Credit limits set as a fixed multiple of sales would treat them the same way.

Pricing and managing the risk

  1. Price the credit. Expected loss and the cost of financing the receivable are real costs of serving a customer; they belong in the customer's margin and in the price.
  2. Set limits on expected loss, not on sales. A weak customer gets a lower limit, shorter terms, a guarantee or credit insurance; a strong one can get more room.
  3. Watch the leading signals. Payment days drifting longer, partial payments, orders shrinking and management changes come months before a default.
  4. Set a concentration policy. A maximum share of revenue or profit per customer, and a plan to grow the next tier when a customer passes it.
BS

Bartor S.

Founder

All insights