On a Scale of One to Ten

I fractured my spine a few years ago. What followed was a long sequence of appointments, and nearly every one of them opened the same way. What is your pain on a scale of one to ten.

I gave a number every time. Six, mostly. Sometimes seven when the day had been bad. I had no real idea what I meant by it, and neither did the person writing it down.

The thing I kept turning over while I waited in those rooms is that my six was built out of my own life. It was calibrated against the worst thing I had ever felt, which set the ceiling, and against what I had already learned to live around, which set the floor. Someone else walks into the same room carrying a different ceiling and a different floor. Their eight might be my four. My six might be, for them, the worst hour they have ever had.

None of that survives the number. The scale looks standardized because everyone uses the same digits. It is not standardized in any way that matters. It is a private measurement reported in a public unit, and the unit is doing no work.

There are two separate things folded into a pain rating, and they come apart the moment you press on them. The first is how much a person can tolerate. The second is what they have available to make the thing tolerable. Someone with medication that works and someone with nothing but time are reporting from completely different positions, and the two of them may well say the same number. One is describing what is left over after relief. The other is describing what they are absorbing in full.

I started hearing the question everywhere once I was out of those rooms.

Rate the severity of this problem, one to five. How painful is this workflow on a scale of one to ten. How disappointed would you be if you could no longer use this. Customer discovery runs on self-reported intensity, and teams take those responses, average them, sort the list, and build from the top of it.

The averaging is the part that should worry people more than it does. You are taking numbers that were never on a shared scale and computing a mean, which produces something with the shape of information and very little of the content. The output looks like a priority order. It is closer to a record of who happened to be having a hard week.

Worse, the bias runs in the direction that costs you the most.

The customer who has adapted will under-report. They built a workaround three years ago, it is invisible to them now, it is simply how the work gets done. Ask them to rate the problem and you will get a four, because they solved it, in the sense that a person with a good prescription has solved it. That workaround might be consuming forty hours a month and two headcount.

The customer who has not adapted will over-report. There is no workaround, the friction is fresh and salient, and you will get a nine. The underlying problem might be worth a few hundred dollars a year to them.

So the severity rating is not just noisy. It is frequently inverted against the thing you actually want to know, which is whether this problem has already extracted money from them.

Which points at the better question, and it is not a better scale. It is behavior.

What are you doing about this right now. What did you try before that. Who inside the company owns the workaround. What did you spend on it last year, in tools or hours or someone's job. Has anyone else looked at this and given up.

A workaround is calibrated in a way a rating never is, because building one costs roughly the same regardless of temperament. Somebody who wrote the spreadsheet, hired the contractor, or bought a competing product has already told you the severity, in a unit that converts. They just told you with an action rather than a digit, and the action cannot be inflated by a bad week or deflated by a high tolerance.

I would like to end this with the fix, but the honest version is that I understand why the scale survives.

The behavioral question takes twenty minutes and a person who knows how to follow an answer. The rating takes five seconds and fits in a survey you can send to four hundred people. Most teams do not have twenty minutes times forty customers, and most clinics do not have it either. So the instrument stays, and everyone using it knows roughly what it is worth, and the numbers get written down and averaged anyway.

I still get asked in exam rooms. I still say six. And I still catch myself wondering what someone else would have called the same thing.