Two owners describe the same manager to me and use the same three words. "They overreact."
One manager sees the business accurately and responds four times harder than the problem warrants. The other is confident about a market they have almost no evidence about. Both get called overreacting, but the first needs a smaller response and the second needs a better test. Give either one the wrong correction and it gets worse.
You need six words for this and no theory: set point, sensor, error, gain, damping, delay. If you want where they come from, the piece on what an emotion actually is builds the whole thing off a thermostat.
Carver and Scheier made the case forty years ago that feedback control describes self-regulation in a person (Carver & Scheier, 1982), and a company steering by numbers is the same architecture with slower sensors (and much worse reporting).
First your head produced a read: enrollment is falling. Then, separately, it produced a size: and I'm rebuilding the schedule this week. The read may be perfect while the size is ridiculous. It runs the other way too, where a perfectly proportionate response is aimed at a picture that's simply wrong.
Most business advice mashes those together, which is why so much of it lands as noise. "Calm down" is aimed at job two. "Look at the data" is aimed at job one. Aim either one at the wrong job and nothing moves… except the relationship, which gets slightly worse!
An owner sees three cancellations in a week. Three cancellations in a week is real and it usually deserves attention. What follows is a full curriculum revision, a price change, two staff meetings and a new marketing agency inside two weeks.
Tell that owner to relax and they will often decide, correctly, that you aren't listening, because the thing they saw was actually there. You just spent your credibility arguing the one point they're right about (and you don't get that back cheap).
Agree with the read out loud and negotiate the size. Yes, three cancellations is real. What's the smallest thing we could do that would tell us something in thirty days?
Reducing gain does NOT mean the response stops. It means matching the size of the response to the size of the problem.
Owners hear "smaller response" as "do nothing," and doing nothing is not what you're asking for.
An owner believes a segment won't buy. That belief produces less effort with that segment, less effort produces poor results, poor results confirm the belief. Reducing the effort is what stopped the evidence from ever showing up. Nobody was irrational at any point, and now the belief has numbers behind it (numbers the belief itself manufactured).
OK you ask, why do pep talks fail here and tests work? Because the update machinery runs on evidence and nothing else. Behrens and colleagues showed people adjust how fast they learn to how volatile the world is, which means the system is watching the environment and setting its own learning rate to match (Behrens et al., 2007). Give it evidence and it moves. Encouragement gives it nothing to chew on.
A clinician does this with a behavioral experiment. Craske and colleagues argue that expectancy violation, not repetition, is what updates a belief (Craske et al., 2014). They have to predict something, watch it not happen, and see the gap themselves.
Write the number down first. "If we call 40 of them this month, I predict fewer than 2 will book." Then run it. Then read the result against the paper, not against how the month felt.
Test one. Make them say it as two sentences. "What do you think is happening?" and then "What are you going to do, and how big is that?" If they can't split the two, that by itself tells you plenty. Most people say the response and assume the read came with it.
Test two. Hand them one piece of contradicting evidence and watch. An accurate read absorbs a contradiction and stays roughly put, because it was built on something real. A confident bad read doesn't move at all and often gets LOUDER. Volume isn't the signature though – immovability is.
Test three. Ask what would change their mind. Somebody with a good estimate can usually name it in a sentence. Somebody defending a prior tells you the question isn't fair, or names something that could never happen, like a competitor calling to apologize.
What if you run all three and get both? That happens a lot, and I would usually fix the accuracy first, because if the picture is wrong, sizing the response perfectly just gets you to the wrong place on time.
I've watched the second one run for a full year in a school that had decided adults wouldn't sign up. They ran no adult intros, adult numbers stayed flat, and the flat numbers got treated as proof. Nobody ever tested it, so nobody could see that the effort and the result were the same number. Less effort was producing the flat adult numbers that proved adults wouldn't sign up.
If you want the longer version of why we defend a position we've already committed to, the eight biases piece walks the shortcuts that make it feel completely reasonable from the inside.
Separate what you think is happening from how hard you're going to act on it and say them as two sentences. Decide your review interval in advance, because under stress you will set it too short. And write down what would change your mind before you have the answer, which is the only reliable way to tell whether you're updating on evidence or defending a prior.
Pick one decision on your desk this week and run test one on yourself. Two sentences, out loud, written down. If the response is clearly larger than the read justifies, you've found your fix and it costs you nothing to make it smaller. If they're already sized to each other – you're in good position!
I'd rather you find that than a consultant does. Reach out any time.
Behrens, T. E. J., Woolrich, M. W., Walton, M. E., & Rushworth, M. F. S. (2007). Learning the value of information in an uncertain world. Nature Neuroscience, 10(9), 1214–1221. https://doi.org/10.1038/nn1954
Carver, C. S., & Scheier, M. F. (1982). Control theory: A useful conceptual framework for personality–social, clinical, and health psychology. Psychological Bulletin, 92(1), 111–135. https://doi.org/10.1037/0033-2909.92.1.111
Craske, M. G., Treanor, M., Conway, C. C., Zbozinek, T., & Vervliet, B. (2014). Maximizing exposure therapy: An inhibitory learning approach. Behaviour Research and Therapy, 58, 10–23. https://doi.org/10.1016/j.brat.2014.04.006
A companion business article to this series.
Addenda
Companion business articles
The full documents, free and ungated