The System Is the Memory
The invoice arrived late on a Friday night. It was for a year of website hosting, billed up front, on a domain that appeared nowhere in my records.
The email said almost nothing. It gave an amount and the name of a plan. The domain and the twelve-month term were buried in an attached PDF I would never have opened at that hour. The part of my operating system that reads the mail opened it anyway. It searched every record I keep for that domain and found nothing: no project, no note, no line in the list of what I own. By the time I woke up, the invoice was waiting for me with a plain question attached. Is this yours, and do you still want it?
It was mine. Once I saw the name, I remembered. It belonged to a project from years ago, long since finished. The contract was still alive because, as far as I could tell, other domains I do want were tied to it.
That was the whole of my memory on the subject. As far as I could tell.
The charge went through before I had a better answer, and I let it stand. Canceling a contract I could not explain might break something I cared about. So a project that ended years ago bought itself another year of life. The only defense anyone could offer for it was a guess.
That is the real cost here, and the money is the small part of it. Years ago I made a careful decision. I tied those domains to that contract for reasons that were obvious at the time. The decision survived perfectly. An auto-renewal is the purest trigger there is, and it carried my decision forward every year exactly as instructed. The reasons did not survive at all.
Intent is the deciding: what matters, why, and what should happen about it. The contract kept the last part flawlessly. It never held the why.
In 1885 a German psychologist named Hermann Ebbinghaus published the first careful measurement of forgetting. He was his own test subject. He memorized lists of nonsense syllables until he could recite them without error. Then he measured how much work it took to learn the same lists again later. After one day, roughly two-thirds of his original effort was gone. After that the loss slowed down. Whatever survived the first few days tended to hold for weeks.
Most people cite Ebbinghaus to sell a study method. Review on a schedule and the curve bends back up. That is true, and it is the least useful thing in his work for anyone running an operation.
The useful thing is why he chose nonsense. Meaning helps memory. A real word is propped up by everything else you know. A syllable like DAX connects to nothing, so nothing can hold it in place. He wanted to watch forgetting with the meaning taken out.
An operator does not need a laboratory to run that experiment. The details that run a working life arrive with the meaning already stripped away. A contract number. A renewal date. The fact that one account quietly holds up three others. Each one made complete sense on the day it was decided. A year later, each one is a nonsense syllable. Nobody rehearses a hosting contract to keep it fresh, and nobody should have to.
The part of the curve that unsettles me is the long flat tail. Ebbinghaus measured effort, so what follows is my reading rather than his finding. What survives the steep drop is rarely the detail. It is a residue: the sense of having known. "Other domains are tied to it, I believe." That sentence felt like memory. It had the confidence of memory. It had almost none of the content.
The organizational tax is remembering, re-deciding, re-finding, and worrying about the thing you know you are forgetting. That week I paid every part of it on a decision I had already made once.
The system had always held the what: the date, the amount, the contract. My head held the why, and my head let it go the way heads do.
At work, the discipline meant to prevent this has a name. It is called vendor management. In regulated industries it is a formal program that examiners review.
Most of the programs I have seen are diligent about the paperwork. They track the contract, the renewal date, the security reports, and the annual review. That is the what, filed beautifully. The why usually lives in the head of whoever signed. Sometimes that person left years ago.
Done well, vendor management is the discipline of keeping the why next to the what for everything you depend on and do not control. A vendor file earns its place by answering four questions. What does this vendor do for us? What breaks if it goes away? Who owns the relationship? When is the last day we can still say no?
A vendor list without those answers is a list of nonsense syllables.
The last question matters most once the others are lost. Its answer is the time you need to recover them. In most business contracts, the last day to say no arrives weeks or months before the renewal does. Mine had passed before the invoice ever arrived.
So my own record now holds one more sentence. It says why this contract exists and what I believe depends on it. The belief is marked as a belief. Weeks before the next renewal, the system will ask me to confirm it while there is still time to act on the answer.
When I make a decision now that will outlive my attention, I ask where its reason is going to live. If the answer is "with me," I already know how that ends.
Somewhere in your operation, something renews on its own. A subscription. A contract. A meeting that has met every Tuesday for three years. Pick one and try to say out loud why it exists. If you can, write the reason down next to the thing. If you cannot, find out now, while someone still remembers.
§