The Count That Ended The Argument
Years ago, in a bunk during basic military training, I watched a man lose an argument he never got to make.
He had reported sick. That was his right. The standard response was Panadol: take it, rest, report back when you are fit. He took the sick leave. He did not take the tablets.
The 2IC walked to his bunk, picked up the strip, and counted. Same number. Untouched. Nobody raised a voice. No accusation, no cross-examination, no debate about whether he truly felt unwell. The count settled it.
I have carried that moment for decades, because it taught me something leaders take years to learn. The 2IC did not win with rank. He won because he produced evidence in the one form that could not be argued with. A sealed strip of Panadol is not an opinion. It is a fact with a serial number.
Most leaders never find that form. They walk into a room with a spreadsheet, a deck, a forecast, and they watch it slide off the face of the person opposite them. Then they conclude the other person is being difficult.
The other person is not difficult. They are running a filter nobody has shown you.
Same Numbers, Two Rooms, Opposite Results
Here is a pattern I have watched repeat in boardrooms, in founder one-to-ones, and in family businesses.
Two people sit in the same review. You walk them through the same numbers: the same pipeline, the same churn, the same twelve-month trend. The first leans forward, asks two sharp questions, and commits budget before the meeting ends. The second nods, says "interesting," thanks you, and changes nothing.
Same data. Same room. Opposite response.
When that happens, the instinct is to repeat yourself. Louder. With one more slide, one more model, one more round of proof that already failed to move them. It never works, because volume is not a channel. If the evidence is not arriving in the form the listener needs, more of it lands in exactly the same place as none of it.
Now the part that stings. The person who did not move is not lazy, and they are not stupid. They are the second subject in a study you have probably never read, and you are about to make them more certain of the opposite of what you want.
Picture a management team I once sat with. The chief operating officer wanted to stop a loss-making service line. She brought three years of margin data, a break-even model, and a recommendation. The chief executive listened, asked two questions, and agreed within four minutes. The finance director, in the same meeting, heard the same numbers and asked for "another look at the assumptions." The CFO was not protecting the service line. She had never seen a break-even model she trusted on the first reading. She needed to run the figures herself, with her own hands on the cell references, before the number would live in her body. The COO read that as resistance. It was a channel mismatch.
Two weeks later, the CFO rebuilt the model overnight, reached the same conclusion, and became its loudest defender. The evidence had not changed. Its container had.
The Filter Nobody Installed On Purpose
There is a structure underneath this, and it has a name in the modelling tradition.
Neuro-Linguistic Programming calls it a meta-program — a pattern-level setting that governs how a person sorts experience before they ever reason about it. The one that matters here is the convincer pattern: the channel a person has to receive evidence through, and the amount of that evidence they need before they will act.
Gayle Laborde documented it in Influencing with Integrity (1983), working from the modelling of Bandler and Grinder (1975) and the meta-program framework formalised by Dilts, Grinder, Bandler, and DeLozier in 1980.
Strip the jargon and it is this. Two people can look at the same document and one will call it proof while the other calls it a claim. That is not a difference in intelligence or honesty. It is a difference in what counts as admissible evidence inside their head.
Let me be honest about the status of this idea. NLP meta-programs are a modelling framework, not a validated clinical instrument. Tomasz Witkowski's 2010 review in the Polish Psychological Bulletin examined thirty-five years of NLP research and found the field's evidence base far weaker than its marketing. The convincer pattern is useful the way a working hypothesis is useful. It changes what you test and what you notice. Treat it as a lens, not a law.
What the lens reveals is that your evidence problem is almost never a quantity problem. It is a format problem.
Why The Same Evidence Pushes People Apart
If meta-programs are the view from the practitioner's chair, cognitive psychology has photographed the same phenomenon from the laboratory.
Lord, Ross, and Lepper (1979), in the Journal of Personality and Social Psychology, gave supporters and opponents of capital punishment the same two pieces of research evidence — one supporting deterrence, one undermining it. Both groups became more certain of their original position. The evidence did not settle the question. It polarised the room. Same pages, opposite effects, because each side was reading for confirmation rather than verdict.
Kunda (1990) named the mechanism in Psychological Bulletin: motivated reasoning. People do not reason like scientists hunting for the most accurate answer. They reason like lawyers building a case for the conclusion they already prefer, and they do it without noticing.
Nickerson (1998) catalogued the same bias across every domain he could find — politics, medicine, investment, science itself — and concluded that confirmation bias is the most pervasive error in human judgement.
Mercier and Sperber (2011) pushed the point further in Behavioral and Brain Sciences. Their argumentative theory of reasoning says the human mind did not evolve to find truth. It evolved to win arguments and defend positions. That is why a founder can present flawless data and watch it fail. The listener is not weighing your data. They are protecting a position.
Finally, Petty and Cacioppo's Elaboration Likelihood Model (1986) split persuasion into two routes: a central route, where evidence is genuinely examined, and a peripheral route, where decisions rest on cues like credibility, delivery, and familiarity. Most pitches are judged on the peripheral route long before the central route is ever consulted.
Put those together and the picture is unflattering but useful. Evidence does not speak for itself. Evidence is spoken for, or against, by whatever filter receives it.
Four Channels And Four Clocks
The convincer pattern has two dials. Laborde called them the channel and the mode.
The channel — how the person has to receive the evidence:
- See — they trust what they can look at. Charts, a prototype, a demonstration, a finished sample.
- Hear — they trust what they can hear. A conversation, a verbal walkthrough, a call from someone they rate.
- Read — they trust what is written. A report, a memo, a case study with pages and dates.
- Do — they trust what they have run themselves. A pilot, a trial, a day in the seat, their own hands on the work.
The mode — how much of it they need before they move:
- Automatic — once is enough. Show it once and they are in.
- Number of times — they need to see it a fixed number of times, often three.
- Period of time — they need to see it hold for weeks or months.
- Consistent — they need it to be the same every single time, without exception.
Now combine them, because people carry one of each.
A read-and-automatic person will act on a single clean memo. Send them a pilot and you will bore them. A do-and-consistent person will ignore ten memos and commit the moment they run the pilot and it holds twice in a row. Show them a beautiful chart and you have shown them nothing.
This is where most organisational frustration lives. A chief executive who is see-and-automatic concludes his team is slow because a single convincing chart did not move them. The same executive fails to notice that his own confidence arrives in seconds, while the people around him are built to be convinced by repetition, duration, or personal trial. He reads their caution as a character flaw. It is a setting.
Reading The Filter Before You Present
You do not need a questionnaire. You need three observations, gathered before the meeting where the decision has to happen.
Their language. Listen for the verb that carries trust. "I see what you mean" points to visual. "That rings true" points to auditory. "Let me look at it in writing" points to reading. "Let me try it" points to doing. People leak their channel in the words they reach for without thinking.
Their questions. The first question a person asks under pressure names the evidence they require. A reader asks for the assumptions and the methodology. A doer asks for the sample size and who has used it. A seer asks to see it work. A hearer asks what other people said.
Their history. Ask how they made their last three confident decisions and what finally settled each one. Most people can name the moment. That moment tells you the channel and the clock in one answer.
Notice what you are not doing. You are not changing the facts. You are choosing the container they arrive in.
The Protocol: Find The Filter, Then Match The Evidence
Run this the next time you need a decision and the evidence is not landing.
- Stop producing more evidence. Start calibrating. Ask how they have been convinced in the past. "What was the last decision you felt sure about, and what finally settled it for you?" Their answer names their channel.
- Name the clock out loud. Ask how many times, or over what period, they need to see something hold. A person who says "show me a quarter of it working" has told you their mode, and no clever pitch will shortcut it.
- Re-deliver the same facts in their format. The identical numbers become a printed one-pager for the reader, a live demonstration for the seer, a scripted conversation for the hearer, a two-week pilot for the doer.
- Match the clock, not your patience. If they need it three times, plan the three exposures on purpose. If they need a quarter, build the review into the quarter. Giving up at exposure two wastes everything you built.
- Track the shift in the body, not the words. Agreement is cheap and nodding is free. Watch for the breath, the posture change, the pause before the yes. That is the filter opening.
Two mistakes ruin this more often than any other. The first is delivering in your own preferred channel, because it feels like proof to you. The second is mistaking a clock for an objection. A person asking for more time is often telling you which mode you failed to match.