The invoice that taught me what risk is
One invoice sat 370+ days overdue.
It was a blue chip industrial customer in Tanzania.
The name felt safe.
That is what made it dangerous.
Other settlements from the same period took 4 to 5 months.
Same customer.
Same logo.
Two different realities.
When it first happened, the instinct is to explain it away.
Port delays.
Internal approvals.
Someone on holiday.
A bank issue.
All of those can be true.
None of them change the only number that matters.
Did they pay, on time, like they said they would.
That is when the rule got sharp for me.
Risk is not the market.
Risk is not the currency.
Risk is not the competition.
Risk is a customer who cannot pay.
The rest is inconvenience.
Here is what changed in how I run accounts after that.
1) I stop classifying customers by how impressive they look.
I classify them by observed payment behaviour.
2) I do not call it “growth” until I have seen them settle properly.
If I have not seen on time payment, I do not have momentum.
I have exposure.
3) I treat the first order as a test.
Not a trophy.
Not a forecast.
Just a test of whether the clearing path is real.
It is like lending your car keys to someone you just met.
You do not start with a road trip.
You start with one short drive, then you decide.
A big brand can buy.
That does not mean they can pay.
The operating rule I use now: before you scale an account, make them earn it with payment, not promises.
A quote is a commitment, so I delay it
In a new country, my first quote is not about price. It is about whether the deal can clear.
When Syria reopened, everyone wanted to be first in 2025 and 2026. I refused to quote until the banking and clearing path was real.
I check five things before I quote: who wrote the spec, how the buyer got paid last time, whether anyone local is already approved, what happens at the port, and who I can call at 9pm.
If nobody answers at 9pm, I am visiting that market, not entering it.
Tanzania: I chose the wrong first step, it cost months
Last month in Tanzania, I approved the wrong first step and it cost us months.
Crisis management is a sequencing test. I now force an order:
1) Eliminate what cannot work, before research.
2) Force one focal point. Extra private channels multiply expensive mistakes.
3) Lock the decision in writing, same day.
4) Demand evidence, not a description.
5) Protect cash twice. Profitability and liquidity are separate calculations.
In a crisis, I do not chase answers. I run the order.
The checklists are here if you want the rest: https://checklists.hardmarkets.net
What is one account you call “good” today that has not yet proven on time payment?
Nazem