What is comparison pages in InstinctGTM?

Part 3 · Operate

Comparison pages

The pages that convert best, and the gate that stops them getting your domain penalised.

01What it is

A programme for the pages a buyer reads a week before deciding: versus pages for somebody comparing two named options, and alternatives pages for somebody who has already decided to leave. Every claim carries the competitor's own URL and the date you read it, and nothing publishes without passing a gate.

02Why it exists

Somebody searching a competitor's name with "vs" or "alternatives" is one to two weeks from a decision. If your page does not exist they read the competitor's version of the story instead, which is the only version available.

These pages also convert better than anything else on a marketing site. Ordinary B2B landing pages run at 2 to 5 per cent; competitor comparison pages land at 7.5 per cent and above, and pages framed as "why this might not be right for you" do better still. Honesty earns what the category average cannot, which is why saying where the competitor wins is a requirement here rather than a nicety.

The reason this is built as a gate rather than a generator is the March 2026 core update. It targeted scaled content abuse specifically, and sites publishing volumes of near-identical pages by swapping a name into an identical structure lost 60 to 90 per cent of their rankings. Legitimate programmatic SEO was caught wherever the content was thin. Anyone can generate comparison pages; almost everyone generates ones that get deindexed.

03How to use it

01Add only competitors real buyers weigh against you and about whom you have something specific to say. If the honest answer is that you do roughly the same thing, there is no page.
02Record claims with the competitor's own page as the source. Their pricing, docs or changelog — never a third-party summary.
03Mark at least one claim as favouring them. A table where every row favours you is an advertisement, and a reader who notices stops believing the rest.
04Write where they are genuinely better and who should choose them. Both are required and the page will not publish without them.
05Run the gate. A refused page names what would have happened rather than showing an error.
06Record what search did with each page by hand, and prune the ones it declined to index.

04What good looks like

You give it: A competitor with four sourced claims, one of which favours them, plus a stated strategic difference and a named situation where somebody should choose them instead.
You get: The page publishes, carrying every claim with its source link and check date, and links into a pillar.
Why: It has all five traits that separated pages surviving the core update from pages deindexed by it: unique data, a hand-written opening and ending, a task served on the page, a pillar link, and a query no other page already answers.

Derived from the rules in the code

05What weak looks like, and what it costs

You give it: A second versus page targeting a query an existing page already answers.
You get: Refused, naming the page that already answers it.
Why: Two pages answering one query is the template-swap pattern, and it gets both of them deindexed rather than ranking either. That threshold — a distinct query per URL — is the one the whole programme rests on.