Evidence is not a kind of number. Evidence is a relationship between a number and a claim — so the same chart is data on one slide and evidence on the next, and nothing about the chart changes.
The same chart, in two meetings. In one of them somebody says the sentence it is meant to support.
A chart of customer visits by month. Accurate, current, well drawn.
The same chart, with a line above it: “Visits fell 18% after the booking change in March — we should reverse it.”
One is a recorded fact. One is that fact standing in relation to something somebody is trying to establish.
Evidence is a relationship, not a substance. No amount of cleaning, charting or checking turns a number into evidence — the change happens when somebody names the sentence the number bears on, and it can happen in a meeting without anyone touching the data.
Which is why an unattached figure is worse than no figure. A chart with no claim invites everybody to supply their own, and a room full of privately different sentences looks exactly like a room having a discussion.
Tap a figure, then tap whether a claim is attached to it. Then see the claim each one needs.
Pick a situation. In every one, the numbers were real and correctly produced.
The chart of customer visits is accurate, well drawn and current. Nobody said what it was meant to show, so eight directors each supplied a different sentence.
Five questions. Nothing is scored.
Five terms, not two. Tap one.
Look for a claim or question. If the numbers or facts are presented without saying what they prove or challenge, you are looking at data. If the information is used to support or refute a specific claim, it is evidence.
You risk making decisions without understanding what the information actually supports. Data without a claim can mislead, leading to wasted time, missed targets, or decisions that do not solve the real problem.
No. Evidence always starts as data. You cannot support or challenge a claim without something recorded or observed. However, not all data becomes evidence—only when it is tied to a claim.
A team presents a dataset showing user clicks, claiming their AI model works better. Without comparing those clicks to a baseline or stating what counts as 'better', the data does not become evidence.
A chart is only evidence if it is tied to a claim or argument. Without a clear statement of what the chart is supposed to prove or challenge, it remains data—however polished the graphic.
Evidence focuses the discussion on claims that matter, helping you act with confidence. Raw data can overwhelm or mislead if nobody interprets it. Evidence gives you a reason to act, not only numbers to look at.
Without a clear claim, people interpret the same data in different ways. Decisions get delayed, arguments break out, and nobody is sure what the numbers are meant to show. The meeting ends without clear direction.
Nothing you can do to a number makes it evidence. Somebody has to name the sentence it bears on.
Copyright © Pawan Nayar · LLOS.ai · 2026 — Data vs Evidence: a recorded fact, versus a fact answering a question.Original pedagogy, voice, and design — all rights reserved.