The ask that gets paid
The letters that worked were not firmer. They were shorter and they asked for one thing.
At one of the practices I managed, I inherited a receivables ledger with a long tail on it. Balances from eight months back, twelve months back, one or two that predated anybody still working there. The standing procedure was a monthly statement with a covering line: please find attached a summary of your account, we would appreciate settlement at your earliest convenience.
That letter had gone out, in some cases, eleven times.
What eventually moved money was not a harder version of it. It was a different message, and the difference took me an embarrassingly long time to name.
What a general reminder actually asks for
A statement of account asks the recipient to do research.
Look at it from their desk. They receive a page listing four invoice numbers and a total. To act on it, they have to find the invoices, work out whether the work described on each one was actually received, check whether somebody in their organisation already approved it, decide whether the amount is right, decide whether the timing is right, and then decide to pay. That is five or six decisions, several of which require talking to a colleague, and none of which they were doing this morning.
The rational response to a request that large is to put it aside, and putting it aside is what “we will look into it” means.
The message that got actioned did the research first. One invoice, not four. The number, the date, the amount, what it was raised against, and then a single sentence identifying the exact place their record and mine came apart. Your system shows this cleared in March. Mine shows a partial. Here is the difference. Can you confirm which one is right.
That is not a request for money. It is a request for a yes or a no, and it can be answered without leaving the chair.
The part I had backwards
For about two years I thought this was a question of tone. I wrote firmer letters. I put dates in them. I got better at the sentence that implies a consequence without stating one.
Tone changes the emotional temperature of an inbox. It does not change the amount of work sitting in the message, and the amount of work is the thing deciding whether the message gets handled today or moved into the pile of things that require a proper sit down.
I also had the causation the wrong way round on effort. I assumed the specific message worked because it demonstrated seriousness, that going to the trouble of reconstructing the account signalled I was not going away. Maybe some of that is true. But the signalling explanation predicts that a long, detailed, thorough letter should work best, and long detailed letters were the worst performers in the whole ledger. The good message was almost always the shortest one, because brevity was a side effect of having already removed every decision except one.
Where it stops working
None of this manufactures money.
A share of that tail was not an information problem at all. The payer knew exactly what they owed, agreed with the number, and did not have it. Against that, a beautifully specific ask buys you a faster refusal and a more honest conversation, which is worth something and is not a recovery.
There is also a version of this that curdles. If every message you send is engineered down to a single frictionless yes, you can end up doing all the thinking for a counterparty who was perfectly capable of doing their own, and you have quietly absorbed their accounts payable function into yours for free. I have done that. It gets you paid and it costs more than it looks.
What I notice now is mostly on the receiving side. When something has sat in my own inbox for three weeks, it is almost never because I do not want to do it. It is because the message asked me to make a decision I could not make from inside the message, and I have not yet found the twenty minutes to go and find out what the decision is.