Parse, which sells AI-visibility tracking, studied 7.4 million links cited by ChatGPT Search and Google AI Mode between May and August 2026 and found that URLs with a comparison term supplied only 4.7% of them. (Parse) That slice is small, yet comparison pages are the ones buyers read right before they choose, so a sloppy one wastes your best shot. A "[you] vs [competitor]" page, a "[competitor] alternatives" page and a "best [category] tools" page each catch a buyer at a different point, and each needs a different build.
They're also some of the most abused pages in SaaS. Plenty of vs pages are thin, one-sided or generated at scale, and buyers know it.
Our view: an honest comparison page that admits where a competitor wins will outperform a one-sided page, with buyers and with AI answers, because it's the only kind either one will trust. If you aren't willing to say who should pick the other tool, don't publish the page. This guide covers the three page types, which query each one answers, how to structure each, the honesty rules that keep them credible, and how to keep them from becoming a liability.
Three page types, three different jobs
Start by matching the page to the question. They aren't interchangeable.
Page type | Query it answers | Buyer's state of mind | What they want from the page |
|---|---|---|---|
Vs page | "[you] vs [competitor]", "[A] vs [B]" | Two names on a shortlist | A fair, specific comparison and a recommendation by situation |
Alternatives page | "[competitor] alternatives", "tools like [competitor]" | Unhappy with, or wary of, an incumbent | Options, why people switch, what they'd give up |
Category page | "best [category] software", "[category] tools for [segment]" | Building a shortlist | A clear list with criteria and who each option suits |
The data backs up the split. Ahrefs counted source types behind ChatGPT's answers to 750 top-of-funnel prompts: 43.8% were "best X" blog lists. (Ahrefs) Set that beside the much larger Parse sample in the opening, where comparison-labeled URLs were 4.7% of citations.
Read together, the two numbers say list-style pages dominate "best X" questions, and comparison pages are a small slice of AI citations overall. They're cited when the question is comparative, and that's when the buyer is closest to choosing.
Marketer Neil Patel put the commercial case on X: "when someone is deciding what to buy, the comparison page is the one that sends them to you." (Neil Patel on X)
How to build a vs page
A vs page compares you with one named competitor. It's the most direct page you'll write, and the one buyers trust least by default. The structure we'd use:
``` H1: [You] vs [Competitor]: [honest one-line summary of the difference]
The short answer (2–4 sentences) Who should pick [You], who should pick [Competitor].
At a glance (table) 5–8 rows buyers actually decide on: pricing model, key capability, integrations, setup effort, support, contract terms. Date the table.
Where [You] is the better fit Specific situations, each with a reason you can show.
Where [Competitor] is the better fit Real situations. If you can't name any, you're not ready to publish.
Switching from [Competitor] What carries over, what doesn't, how long it takes.
Sources and last-updated date Links to both vendors' pricing and docs.
Section 4 is the one most teams skip. Don't. Google's guidance for reviews asks you to "discuss the benefits and drawbacks" and explain "which might be best for certain uses or circumstances." ([Google Search Central](https://developers.google.com/search/docs/specialty/ecommerce/write-high-quality-reviews)) It's written for product reviews, but a vs page is a review by a biased reviewer. The same principles apply, more so.
## How to build an alternatives page
An alternatives page lists several options to one incumbent. The buyer is usually unhappy with that incumbent, or nervous about committing to it. So the page should start with why people look elsewhere, not with you.
``` H1: [N] [Competitor] Alternatives for [segment or use case]
1. Why teams look for [Competitor] alternatives The real, common reasons (price changes, missing feature, fit). Cite reviews or public sources where you can.
1. Summary table Each alternative: best for, pricing model, key difference.
1. One section per alternative (including you, flagged) Same sub-headings every time: Best for / Strengths / Trade-offs / Pricing.
1. How to choose 2–3 questions that point the reader to the right option.Include real competitors besides yourself. An alternatives page that lists only you is just a landing page with a misleading title.
How to build a category page
A category page answers "best [category] tools". You have two options: a product-led category page about what the category is and how your product fits, or a roundup that lists several vendors.
If you write a roundup that includes yourself, be upfront. Ahrefs' data shows "best X" lists where the publisher ranks itself first still show up in ChatGPT's sources, though Ahrefs itself left the choice to readers and said it won't always put itself in the top spot. (Ahrefs) On X, SEO consultant Lily Ray has poked fun at the habit, describing "objective, deeply researched" articles where the publisher "happened to come out on top." (Lily Ray on X)
Self-ranking lists can work. They only stay credible if you flag yourself clearly, use the same criteria for every entry, and say where others beat you. Google's review guidance also asks that ranked lists include "enough useful content" to stand on their own, and that when you call something best, you explain why "with first-hand supporting evidence." (Google Search Central)
The roundups on this blog follow the same approach. The sales automation SEO agencies list, for example, judges every agency, including Miraa, on the same published tests.
The honesty rules
These rules apply to all three page types. They're what separate a page buyers trust from one they bounce off.
Only claim what you can show. Every capability claim should link to docs, a screenshot or a public page.
Never invent a competitor's weakness. If you're not sure a competitor lacks a feature, check their docs or leave it out. Miraa's fintech page puts the rule plainly: a page can take a position, but "it cannot invent a competitor's defect." (Miraa)
Use their current pricing, with a date. Competitors change pricing often. An outdated number is the fastest way to lose credibility.
Name situations where they win. Buyers trust a page more when it admits trade-offs.
Put conditions next to claims. "Unlimited seats on the Enterprise plan" is accurate. "Unlimited seats" with a footnote isn't. AI summaries tend to keep the claim and drop the footnote.
Show who wrote it and when. A byline and a last-updated date help readers judge the page.
Linking them together
Comparison pages work best as a small cluster, not a pile of orphaned URLs.
Category page at the top, linking to your product page and to the vs pages for the competitors it mentions.
Vs pages linking to the alternatives page for that competitor, to your pricing page, and to relevant integration pages.
Alternatives page linking to each vs page you've published.
Product, pricing and integration pages linking back to the relevant comparisons ("Comparing us with [Competitor]? See the side-by-side").
Ahrefs also looked at which of a brand's own pages ChatGPT used when it recommended that brand. In the software category, general landing pages showed up in 37.2% of those prompts and the brand's own "best" lists in 34%. (Ahrefs) A strong cluster gives a model more of your own pages to draw from.
Technical basics matter too. Comparison pages built in interactive widgets sometimes ship their key content in client-side JavaScript that crawlers may not render. Miraa's technical SEO page calls out "orphaned comparisons" and "a comparison URL that returns a soft error" as technical problems, not just design ones.
Maintenance: the part nobody budgets for
Comparison pages go stale faster than any other content. Competitors ship features, change pricing and rebrand.
Ahrefs found 79.1% of the 1,100 "best" lists ChatGPT used as sources had been updated in 2025, and 26% in the two months before the study. (Ahrefs) Freshness isn't proven to cause citations, but stale lists are clearly in the minority among the ones that get used.
A practical schedule:
Monthly: check competitor pricing pages for your top 3–5 vs pages.
Quarterly: review every comparison page for feature changes, and update the "last updated" date only when you actually change something.
On a competitor launch or rebrand: update the affected pages within a couple of weeks.
Don't scale these pages with a template
It's tempting to generate a vs page for every competitor pair in your category. Don't.
According to Google's AI optimization guide, producing separate content for each variation of a query "primarily to manipulate rankings or generative AI responses" falls under its scaled content abuse policy. (Google Search Central) A template with a different competitor name dropped into each copy is close to a textbook case. Miraa's fintech page says the same about programmatic pages: a template stamped with a thousand names is "the mass-content pattern" to avoid. (Miraa)
Build pages for the competitors your buyers actually compare you with. For most B2B SaaS companies, that's a handful, not hundreds.
Five mistakes we see on SaaS comparison pages
The 40-row feature grid. Long checkmark tables make every product look the same. Buyers decide on five to eight things. Pick those, and explain each.
Comparing against last year's competitor. If a competitor relaunched their pricing or product, a page that still describes the old version reads as careless, or worse, as deliberate.
Hiding the page. Vs pages that sit outside the navigation and get no internal links are hard for buyers and crawlers to find. Link them from pricing and product pages.
Writing for the keyword, not the buyer. A page that repeats "[Competitor] alternative" in every heading but never says why someone would switch won't convert, even if it ranks.
No owner. Comparison pages need one person responsible for keeping them accurate. Without an owner, they go stale within a quarter.
Measuring comparison pages
Comparison pages should be judged on conversions, not traffic. Track:
Demo requests and trials started from each page.
Assisted conversions (visitors who read a comparison before converting elsewhere).
Whether AI answers to "[you] vs [competitor]" prompts cite your page, and whether they describe the difference accurately.
For the full measurement setup, see the B2B SEO pipeline attribution guide. For how answer engines choose sources, see What is AEO?.
Final thoughts
Comparison, alternatives and category pages are where SEO is closest to sales. So honesty pays more here than on any other page type, because the reader will test every claim on a trial or a sales call.
Start with the vs pages for your three most-compared competitors. Make them fair enough that a competitor's customer would find them useful, put a refresh date on each, and keep it. Only then add the alternatives page and the category page around them. Don't generate fifty templated vs pages; that's the fastest way to turn this page type into a liability.
If you sell sales software, the sales tech SEO playbook applies this to a crowded category. For how these pages fit into a full program, see our B2B SaaS SEO guide. For help building comparison pages that hold up, see Miraa's B2B SaaS SEO approach.