The 552-page tell: what happens when a comparison page is arithmetic

The largest single content cluster in the product-demo market is 552 pages under one URL pattern, comparing screen-recording tools against each other. We read all of them.

It is not 552 articles. It is 187 tool pairs, multiplied by three angles.

The multiplication is in the URLs

The pattern is legible without reading a word:

• camtasia-vs-bandicam-enterprise-readiness
• camtasia-vs-bandicam-feature-comparison-guide-2026
• camtasia-vs-bandicam-pricing-value-comparison-guide-2026

Three pages, one pair of products. Across the cluster, 159 of the 187 pairs have exactly three pages. The angle suffixes are drawn from a fixed set: enterprise readiness appears 183 times, pricing 174, features 100 in various spellings.

That is a decision made once, in a spreadsheet, and executed 552 times. A person choosing what to write about does not produce 159 groups of exactly three.

What three siblings actually contain

We read the Camtasia and Bandicam set. All three carry the same publication date. All three open with a bolded statistic:

• “83% of enterprise IT teams” requiring SSO and automated provisioning
• “64% of content creators” citing post-production as their bottleneck
• “40% of the software’s advanced features go unused”

None of the three names a study, a year, or a source. Across the whole cluster we measured roughly two such figures per thousand words.

All three also define the same two products from scratch. “What is Camtasia?” appears on every page in the pair’s set, reworded each time. A reader arriving from search at the pricing page and then the features page reads the same two definitions twice, differently phrased, as if they had never met.

The byline

One author name appears on 98% of all 552 pages, described as a head of learning and development with over fifteen years of experience in enterprise software implementation.

We have no way to know whether that person exists, and we are not going to assert that they do not. What we can say is measurable: one name is credited with 552 comparison pages covering 187 product pairs, published in bulk, none of which cites a source.

Why this is the interesting case

A cluster like this is easy to dismiss and hard to detect.

Only 12% of any given page’s text appears anywhere else in the set. Every page was worded fresh, so nothing registers as copying. Run the standard duplicate check and these pages score 95% unique. They would pass almost any automated quality gate, including, until we changed it, our own.

That is the part worth taking seriously. The multiplication is not in the words. It is in the structure, the dates, the byline, and the absence of anything a reader could verify.

We have no traffic data for this cluster and we are not claiming it fails. It may do very well. Our position is narrower and, we think, harder to argue with: we could not build it and still describe honestly how it was made.