Home / Patterns / Unsourced statistics

Unsourced statistics: why the pattern reads as AI-written and how to fix it

A percentage or similar figure stated as fact with nothing in the post saying where it came from.

A percentage or similar figure stated as fact with nothing in the post saying where it came from. The free checker at klype.ai/check flags it as "Figures with no source". It is one of the heavier findings: the checker deducts twice as much for it as for a formatting tell.

What it looks like

  • 87% of B2B buyers research a vendor on LinkedIn before they ever talk to sales.
  • Posts with a question in the first line get 3x more engagement.
  • Only 2% of founders post consistently, which is why the ones who do win.

Example text on this page is written to show the pattern. Figures and people in it are illustrative, not reported.

Why readers distrust it

People who know a field recognise its real numbers, and they recognise the invented ones faster. A single figure that cannot be traced tells those readers that the author did not check, and once that is established every other claim in the post is read as unchecked too. The figures that circulate most on LinkedIn, the "87% of buyers" family, have usually been copied from post to post until nobody can say where they started, which is why experienced readers skip past them.

Why it shows up in generated drafts

A model asked for a persuasive post produces the shape of persuasion, and in business writing that shape includes a statistic near the top. The model does not retrieve a figure; it generates a plausible one. The result looks like evidence and is not.

How the free checker spots it

The checker finds every percentage in the draft, meaning a number followed by a % sign. If it finds at least one and the draft contains none of the phrases "according to", "source:", "study by", "report by", "survey by" or "data from", it flags the first sentence that contains a percentage. It does not verify the figure, and it does not catch figures written without a % sign, such as "nine in ten", "3x" or "$2 million". The second example above would pass this rule and still be unsourced.

The checker runs in your browser; nothing you paste is sent anywhere.

How to fix it

  1. Find where the figure came from. If you cannot, cut it. A post with no statistic is stronger than a post with a doubtful one.
  2. If you can, name the source inside the sentence, with the year, so the reader does not have to ask.
  3. Prefer your own numbers. A figure from your own pipeline or your own posts needs no citation and cannot be found on anyone else's feed.

Before and after

One draft, rewritten
Before

87% of buyers research on LinkedIn before they talk to sales. If you are not posting, you are invisible to most of your market.

After

Of the 31 inbound calls we booked this quarter, 19 mentioned a specific post before I had said anything about the product. I did not expect that ratio when I started writing.

The rewrite removes the pattern without adding anything the author could not stand behind.

Test your draft

The checker reports every pattern it finds, not only this one.

Check drafts against your own voice. The free checker tests the writing. Inside Klype, Check also scores a draft against your own measured voice, matches every figure to the facts saved for your business and applies your publishing rules. Start a 7-day trial.

Common questions

If I add "according to", does the checker pass the figure?

Yes, because it looks for a sourcing phrase, not for the truth of the number. The check removes the pattern; it does not do your fact-checking. Inside Klype, Check also compares each figure with the facts you have saved for your business, which catches a number that contradicts your own records.

What about numbers from my own business?

Use them. "19 of our 31 inbound calls" is specific, verifiable by you, and impossible to copy from another post. Write it in a way that shows it is yours, such as "of our" or "in my pipeline", so readers do not take it for an industry figure.

Why do generated drafts contain so many statistics?

Because business writing in a model's training data contains them, and the model reproduces the shape. It has no record to check against, so the figure is a plausible invention. Treat every percentage in a generated draft as unverified until you have found it yourself.