I was the API
For a few seasons, I was the API.
In software, an API is how one system asks another for something. In a lot of small businesses, the API is a person. At Adventure Idaho, for a while, it was me.
“Ask Justin.”
“Justin knows how that works.”
“Check with Justin before you schedule it.”
“I think it’s on the spreadsheet, but ask Justin.”
The booking system knew who had paid. The schedule knew who was working. The group text knew who had a flat tire. The whiteboard knew which van was going where. None of them knew about each other. I was the part that connected them.
It works, which is the problem
Being the API isn’t a failure of effort. It’s usually the opposite. The person in that seat is fast, knows the exceptions, and answers at ten at night. At small scale it’s the most efficient system there is.
That’s why it hides so well. Nothing breaks. Every question gets answered. The cost doesn’t show up as an error. It shows up as everything else.
Decisions wait until one person is free. Planning happens at night, because the day goes to answering questions. For us, weekly planning took twenty to forty hours before it was rebuilt. Training a new manager takes a season, because the rules aren’t anywhere they can read. And the business can only grow as fast as one head can carry it.
Why the software didn’t fix it
We had software, and it was good at what it did. It recorded transactions: who booked, who paid, what was sold.
What it didn’t hold were the rules. Which guide can run which section. Which van can pull which trailer. What happens when a trip runs long and the next one needs the same driver. Those lived in my head, and in the heads of the few people who had been around long enough.
The software stored the data. I stored what the data meant.
That gap never had a product built for it, because until recently writing a business down cost more than it saved.
What changed
The goal was never to replace the person. It was to get the repeatable parts out of the head, so the person is only needed for the parts that genuinely need one.
For us that meant a few things, in roughly this order. Writing the standards and checklists down, so the work went the same way whoever showed up. Building a model of the operation that held the real rules, so conflicts showed up the day before instead of the morning of. Capturing changes where the crew already talked, instead of asking them to fill in a form. And a short list every morning of what actually needed a person that day.
The crew can see the plan now without calling me. Weekly planning went from twenty to forty hours to two to four. Most of the questions that used to come to me don’t anymore, because the answer lives somewhere they can reach.
What stays human
I’m still the API for some things, and I should be. Judgment calls. Exceptions nobody has seen before. Anything that depends on knowing a person rather than a rule.
That’s the difference. It used to be every question. Now it’s the ones that deserve a person.
A test worth running
If you run an operation, count for one week the questions that reach you that someone else could answer if the information lived somewhere they could get to.
That number is your API traffic. Each one is a piece of the business that only works while you’re in the room.
It’s also the first thing worth looking for in any business someone is thinking about buying, because it’s the part of the value that leaves when that person does.