> ## Content Index
> Fetch the complete content index at: https://www.datadab.com/blog/llms.txt
> Use this file to discover other available public pages before exploring further.

# The Comparison Page Library: Why 69% of Vendors Publish None, and What the Other 31% Do
- URL: https://www.datadab.com/blog/comparison-page-library/
- Published: 2026-10-12T04:30:00.000Z
- Updated: 2026-10-17T11:54:18.000Z
- Description: Most vendors publish no comparison page, and much of what looks like one is a glossary post. Build a library your buyers actually use.
- Author: Amit Ashwini
- Tags: Competitive Intelligence

*One comparison page is a post. Twenty is a library with an owner. Most teams confuse the two and stall at one.*

Of the 479 YC-backed B2B software companies in our [comparison-page study](https://www.datadab.com/research/saas-comparison-page-map), 148 published at least one page comparing themselves to a named rival.

That is 30.9%. Which leaves 331 companies publishing nothing at all.

Two thirds of the sample.

We expected the rate to be lower at the big companies and roughly flat at the small ones. The opposite turned out to be true.

At 1-10 employees, 24.0% publish a comparison page. At 11-50, 27.4%. At 51-200, 36.9%. At 200 or more, 51.2%.

A 2.1x spread. It climbs at every step. Category explains much less. Sector rates run from 0.0% in Supply Chain and Logistics to 48.3% in Analytics against 30.9% overall. Size explains more of the variance than category does.

Under fifty employees, publishing comparisons is a minority position, and a perfectly reasonable one. Above that, your buyer's shortlist assumes rival names exist. Absence stops reading as neutrality and starts reading as having nothing to say.

![Editorial cover for an article on whether a comparison asset is a single post or a maintained library.](https://www.datadab.com/blog/content/images/2026/10/comparison-page-library-01-hero-1.png)

![Distribution of 479 B2B SaaS companies by the number of vendor comparison pages they publish. Non-publishers hold 69.1% of the sample; 2.7% have reached programme depth.](https://www.datadab.com/blog/content/images/2026/10/comparison-page-library-02-archetype-ladder-2.png)

## The gap between one page and a library is where the work is

The publisher with exactly one comparison page and the publisher with twenty look similar in a search result. They are not the same thing.

Here is how the 148 publishers distribute:

| Archetype        | Companies | Share | What it looks like                                  |
| ---------------- | --------- | ----- | --------------------------------------------------- |
| Non-publishers   | 331       | 69.1% | No page we could classify                           |
| The One-Pager    | 47        | 9.8%  | Exactly one page, usually against the nearest rival |
| The Occasional   | 51        | 10.6% | Two to five pages                                   |
| The Committed    | 37        | 7.7%  | Six to twenty pages                                 |
| The Programmatic | 9         | 1.9%  | Twenty-one to a hundred pages                       |
| The Library      | 4         | 0.8%  | A hundred pages or more                             |

Twenty and a half percent of the sample sits in the light-touch tiers of one to five pages. Two point seven percent has reached anything like programme depth.

The distribution is the argument. One company alone produces 15.7% of all comparison pages in the sample. The top three produce 31.8%. The top ten produce 57.3%.

That is not a level playing field, and pretending otherwise wastes a quarter. If you are one of the 331, you are not competing for the shortlist on content depth. You are competing on the twenty minutes a buyer spends with the three pages that exist.

If you are one of the four, you are not winning because you write well. You are winning because you have a page for every tool a prospect might name.

![Of candidate URLs containing the string vs, 77.0% were concept articles such as lead versus prospect and only 19.3% were vendor-versus-vendor comparisons.](https://www.datadab.com/blog/content/images/2026/10/comparison-page-library-03-not-a-comparison.png)

## Most of what looks like comparison coverage is not

This part quietly breaks most content audits.

We classified 1,683 comparison pages from 8,717 candidate URLs across those 479 companies. Of the candidate URLs containing "vs", 77.0% were concept articles ("lead vs prospect", "inbound vs outbound", "onboarding vs activation"). Only 19.3% compared vendors.

If you counted your own coverage by looking for URLs that match a pattern, you have almost certainly counted a lot of glossary posts.

It also means competitor research built the same way is unreliable. A tool that reports "this site has 60 comparison pages" may have found sixty pieces of conceptual marketing.

We had to read the titles. That is a property of the format rather than a flaw in the tools. A URL cannot tell you whether it argues about two products or two definitions.

Forty-nine candidate URLs came back as fetch failures. We exclude those from the classification rather than counting them as non-comparison pages. A page we could not load is not evidence of anything.

![Of 1,683 classified comparison pages, 90.3% name a single target vendor, 9.4% name two, and fewer than one percent name three or more.](https://www.datadab.com/blog/content/images/2026/10/comparison-page-library-04-one-rival.png)

## Most pages name exactly one rival

Across the 1,683 pages we classified, 90.3% named a single target vendor. Nine point four percent named two. Three pages named three. One named four.

Worth pausing on, because it runs against the instinct to build a comprehensive grid.

A single-vendor page is easier to write accurately, easier to keep current, and closer to how buyers search. Somebody comparing you to one competitor is running a single-vendor check. Somebody comparing you to four is usually reading a listicle a third party wrote.

The multi-vendor grid is a real format. It is also the format most likely to be quietly wrong in six months.

![A wireframe of the sections that make a single-vendor comparison page useful, from pricing unit through to the disqualifying conditions that tell a reader to choose the rival.](https://www.datadab.com/blog/content/images/2026/10/comparison-page-library-05-single-vendor-anatomy.png)

## What actually belongs on a single-vendor page

If you are writing one page against one rival, thinness is not the failure mode. Vagueness dressed as fairness is.

A page saying both products are strong, and that the right choice depends on your needs, has told the reader nothing they could not have guessed. It reads as a comparison and works as a sales page, which is worse than either.

The pages that work state the things a buyer is actually choosing between. What each one charges and in what unit. What each is structurally good at. Where the rival is genuinely the better answer, said plainly enough that a reader can act on it.

Say that last part even when it costs you. A page that names its own disqualifying conditions is the only kind a buyer can act on with confidence, and the only kind worth being cited for.

Two structural rules do most of the work. Keep the pricing units visible without a click, since a buyer comparing units is the reason the page exists. Put the page on a permanent path, because a page that expires cannot be updated, and one that cannot be updated cannot be trusted.

![Two vendors point comparison pages at the same rival while an unmaintained pair is left empty. Of 496 vendor pairs, 216 are contested by two or more companies.](https://www.datadab.com/blog/content/images/2026/10/comparison-page-library-06-contested-pairs.png)

## Nearly half of vendor pairs are contested

Of 496 vendor pairs mentioned across the sample, 216 are attacked by two or more different companies. That is 43.5%.

The top five targets absorb 18.7% of mentions. The top ten take 29.5%. The top twenty-five, 50.9%.

Comparison effort spreads across a long tail rather than piling onto one incumbent. Which means that for a large share of pairs, your competitor is publishing about you whether or not you have published anything at all.

A third party will do the comparison if you will not. That is uncomfortable to sit with. It is also the strongest argument for controlling the framing yourself.

![Of 1,683 classified comparison pages, 65.5% put the publisher's own product in the frame and 34.5% compare two vendors that are both third parties to the publisher.](https://www.datadab.com/blog/content/images/2026/10/comparison-page-library-07-editorial-share.png)

## A third of comparison pages are editorial, and that is the good stuff

Sixty-five point five percent of the pages we classified put the publisher's own product in the frame. Thirty-four point five percent compared two vendors, both third parties to the publisher.

Those 580 pages are not flattering to the publisher. They are also the only pages in the set a buyer can read without assuming a sales motive.

Comparison-shaped questions are the most citation-prone category in our data. Editorial third-party comparisons are the format most likely to be referenced, because they carry a claim about the market rather than a claim about you.

You do not own those pages. You can make sure your facts are extractable by whoever writes them.

![How concentrated attention is across target vendors: the top 5 targets absorb 18.7% of mentions, the top 10 29.5%, the top 25 50.9% and the top 50 70.6%, leaving a long tail of uncontested pairs.](https://www.datadab.com/blog/content/images/2026/10/comparison-page-library-08-target-concentration.png)

## Measure the library

A single comparison page has one measurable job: appearing when somebody compares you to a named competitor. Narrow, and checkable.

A library has a different job. Not page count. Coverage of the pairs buyers actually ask about.

The study gives you the shape of that problem. Of 496 vendor pairs, 216 are contested by two or more companies, and the top ten targets take 29.5% of mentions. So there is a long tail of pairs where no competitor is fighting, and a short head where all of them are.

The uncontested pairs are the cheapest coverage you will ever buy. No competitor is maintaining them, which means they are almost certainly out of date.

Before commissioning anything, pull your sales team's lost-deal notes and count which competitor names appear most. That list beats any framework for choosing what to write. It takes an afternoon, not a quarter.

![Of 1,683 classified vendor comparison pages, 74.3% sit on permanent structured paths and 25.7% sit on blog, learn or resources paths.](https://www.datadab.com/blog/content/images/2026/10/comparison-page-library-09-permanent-paths.png)

## Build it on permanent paths

Of the comparison pages in the sample, 74.3% sat on permanent structured URLs (\`/compare/\`, \`/alternatives/\`, \`/vs/\`). Only 25.7% lived on blog, learn or resources paths.

Seventy-four point three percent is the finding I would put in front of anyone proposing a comparison blog post. In practice, most comparison pages in this market are maintained assets sitting in navigation. They are not editorial pieces published once and forgotten.

A blog post has a publication date and an audience. A comparison page has an owner and a review cycle. Confuse the two and you produce something that looks like coverage and behaves like an archive.

If you take one structural thing from this piece, take that one. It costs a routing decision, not a writing budget.

![Four steps to start a comparison library: pick the pair you lose most often, publish it on a permanent path, state your pricing and your disqualifying conditions, then decide whether you are funding a post or a library.](https://www.datadab.com/blog/content/images/2026/10/comparison-page-library-10-start-here.png)

## How confident we are

The study is a cross-sectional crawl, and it measures form rather than quality.

We fetched 8,717 candidate pages and classified 1,683 of them as vendor comparisons. Forty-nine URLs came back as fetch failures and are excluded rather than treated as non-comparisons. A page we could not load is not evidence about that page.

Classification was done by reading titles and page structure. That means we can tell you a page exists on a permanent path, and nothing about whether it is accurate, current or fair. A maintained page is not the same as a good one.

The archetype counts are the most robust part of this work, because they only require counting pages. The between-group comparisons, such as the size ladder, carry wider uncertainty, and the smallest tiers are single-digit company counts. The shape of the distribution is the finding, not the exact share of any one row.

All of it describes 479 YC-backed B2B software companies in October 2026\. It does not generalise to bootstrapped vendors, to agencies, or to software categories the sample does not contain.

![](https://www.datadab.com/blog/content/images/2026/10/comparison-page-library-10-start-here-2.png)

## Wrap-up

None of this says every company should run a twenty-page comparison programme. It says the choice is real, it is usually made by default, and the default is invisibility.

Under fifty employees, a focused set beats a full library. A competitor's comparison page is not something you have to out-produce. You have to be findable at all. At fifty-one and up, 36.9% of your peers publish comparisons and the share is rising, which means a buyer comparing you to a rival will likely find three versions of that comparison. None of them yours.

Start with the pair you lose most often. One page, on a permanent path, with your pricing stated plainly and a line naming who should not buy.

Then decide whether you are building a library or a post, and be honest about what you can afford.

One caveat, since this post is largely about other people's pages. We classified titles and page structure from a cross-sectional crawl in October. We did not assess whether the pages we found were accurate, current or fair. A page sitting on a permanent path tells you somebody maintains it. It does not tell you they maintain it well.

Which is an argument for checking the ones in your category, not an argument that the category is trustworthy.

Want to see how your own pages hold up?

Put the page a buyer reads before contacting sales under pressure, then [scan it for the gaps that decide the shortlist](https://www.datadab.com/tools/decision-friction-scanner).

Classification came from reading titles across 8,717 candidate pages fetched from 479 companies in October 2026\. We assessed whether a page was a vendor comparison, not whether it is accurate or fair. A page on a permanent path tells you it is maintained, not that it is maintained well.

[Scan your comparison page](https://www.datadab.com/tools/decision-friction-scanner)