GuidePractitioner11 min readPart of GEO Practitioner

Building 'Best X for Y' Pages Without Spam

How to build 'best X for Y' list pages that earn trust from Google and AI engines: evidence for every pick, clear criteria, honest disclosure and no spam.

Dipuck Jones
Written by Founder of MentionX · 13+ years in B2B SaaS growth and SEO
ReviewedUpdated
Short answer

A "best X for Y" page earns trust when every pick has a stated reason and first-hand evidence, the criteria fit the use case, and conflicts are disclosed. Google asks that recommendations of "best overall" or "best for a purpose" include why, with first-hand supporting evidence, and that ranked lists have enough useful content to stand alone (Google). Thin, mass-produced lists risk Google's spam policies (Google).

Key takeaways
  • A best-X-for-Y page earns trust when each pick has a stated reason, first-hand evidence and criteria that fit the use case.
  • Google asks review content to show evidence of experience, explain what sets products apart and cover pros and cons.
  • In MentionX data, best-of prompts named the tracked brand in only 26.6% of answers, against 91.9% for comparison prompts.
  • When the brand did appear in a list, it was sixth or lower 56% of the time, so where you appear matters as much as whether.
  • Include your own product only where it genuinely fits, disclose it and score it on the same criteria as everyone else.

Checked on 31 October 2026. "Our view" marks MentionX opinion. Ledgerline is a fictional company used for illustration.

01Why do these pages matter for AI engines?

Buyers ask engines "best X for Y" constantly. Our view: engines that cite sources pick list pages that clearly state picks and reasons. Google says AI Mode helps with complex comparisons (Google), and its reviews system covers ranked lists, aiming to reward in-depth research over thin summaries (Google).

02What does Google ask for?

Requirement (Google)How to meet it
Why each pick is best, with first-hand evidenceA "why we picked it" paragraph with test results or photos
Enough useful content for the list to stand aloneSpecs, pros, cons and fit for each pick
Focus on key decision factorsCriteria stated up front
Quantitative measurementsNumbers from your testing, with method
Which is best for which circumstancesSub-picks ("best for small teams")
Links to multiple sellersLet readers choose where to buy
Quality and originality, not lengthCut filler

03How do you set criteria?

  1. Define "Y" precisely ("for remote teams of 5–20", not "for business").
  2. List 4–6 factors that matter for that use case.
  3. Weight them and publish the weighting.
  4. Score every candidate the same way.
CriterionWeightWhy
Async video support30%Core remote need
Price per seat25%Small-team budgets
Integrations25%Fit with existing tools
Admin controls20%Security

Example criteria for a remote-team video tool, to show the format.

04What evidence should each pick have?

  • Hands-on testing period and what was tested
  • Screenshots or photos you took
  • Measured numbers (load time, battery, setup time) with method
  • Price and plan with "as of" date and source
  • At least one honest drawback

05How do you handle your own product?

Our view: disclose clearly at the top that you make one of the products, score it on the same criteria, and be willing to rank it lower where it's weaker. If it can't win any sub-category honestly, don't publish the list.

Google says reviews often use affiliate links and points to its affiliate programme position (Google). Disclose affiliate relationships near the top of the page in plain language. Advertising rules in your market may also require disclosure.

07What counts as spam?

Google's spam policies cover scaled content abuse: many pages generated primarily to manipulate rankings with little value (Google). Our view: warning signs include hundreds of "best X for [city/profession]" pages with the same picks, lists of products you've never used, and copy rewritten from manufacturer pages.

08How do you keep lists current?

TaskFrequency
Prices and availabilityMonthly
New modelsQuarterly
Re-test picksYearly or on major release
Update date with change noteEach update

09How do you measure impact?

Run your "best X for Y" prompts across engines monthly; record whether your page is cited and whether engines repeat your picks and reasons correctly.

10How many picks should a list have?

Our view: as many as you've genuinely tested and can justify, usually five to ten. Google asks that ranked lists contain enough useful content to stand on their own (Google), so each pick needs its own reasoning, not a one-line blurb. If you can only justify three, publish three.

11How should you structure each pick?

ElementExample
Pick label"Best for noisy rooms"
One-sentence reason"Tightest pickup pattern we tested"
EvidenceRecorded sample, measured noise floor
Specs and dated priceWith source link
Drawbacks"Needs an audio interface"
Who should skip it"Beginners wanting plug-and-play"

12What about methodology pages?

Our view: publish a separate "How we test" page and link it from every list. Describe equipment, test duration and scoring. It shows evidence of experience across all your lists, which Google's guidance asks for (Google), and gives engines one clear source for your method.

13How do sub-picks help AI answers?

Our view: buyers rarely ask for "best overall"; they ask "best for my situation". Labelled sub-picks ("best for small teams", "best on a budget") give engines a clear statement to match to a specific prompt, and Google's guidance asks reviewers to say which option is best for which uses (Google).

14From MentionX data: how brands fare in best-of answersOriginal data

Method. Read-only aggregate queries over MentionX scans from 22 August to 29 October 2026 (UTC): 2,217 engine runs (2,154 completed) across 18 anonymised projects and six engines.

Best-of prompts are where buyers ask engines to choose for them, and where brands find it hardest to appear. Across all six engines:

IntentBrand mentionedShare of voice
Brand-specific100%64.3
Comparison91.9%44.6
Recommendation29.2%26.8
Best-of26.6%22.7
Problem5.7%7.7

The tracked brand was named in 26.6% of best-of answers, with a share of voice of 22.7. When it appeared in a list (102 answers), it was first in 18%, second or third in 14%, fourth or fifth in 13%, and sixth or lower in 56%.

Across all engines, blog posts and articles made up 19.6% of cited URLs and review pages 2.7%.

Our view: engines answering best-of questions lean on pages that already make a reasoned choice. A list with clear criteria and labelled sub-picks gives them statements to use; a list of logos and adjectives doesn't.

Limits. A convenience sample of mostly B2B software projects, run through engine APIs, which can behave differently from the consumer apps. Perplexity, Copilot and DeepSeek figures come from a single day (26 October 2026), so treat them as a snapshot, not a trend.

15Worked example: Ledgerline's "best AP automation for mid-sized companies" page

Ledgerline is a fictional accounts-payable tool. Figures are illustrative.

Ledgerline wants a best-of page, but knows a vendor ranking itself first invites suspicion.

The team defines the use case precisely: finance teams processing 1,000 to 10,000 invoices a month, using a mid-market ERP. They publish five weighted criteria: ERP integration depth, approval workflows, invoice capture accuracy, multi-entity support and total cost.

They trial six products, including Competitor A to D and Ledgerline, with the same set of 50 sample invoices, and record what they found for each criterion. Ledgerline wins on multi-entity support but not on capture accuracy, where Competitor B scores higher, and the page says so.

Each pick gets a labelled sub-pick, such as "best for multi-entity groups", with a two-sentence reason. The page discloses that Ledgerline publishes it and links a "How we tested" page.

16Why does criteria-first structure help?

Buyers rarely ask for "best overall". They ask for the best option given their situation, and engines try to match that. A page that states its criteria and weighting, then scores each product against them, gives engines explicit reasoning to draw on. It also gives readers a way to disagree with your weighting and still use the page, which is a mark of a trustworthy list.

17How do you handle competitors fairly?

Use the same criteria and evidence for every product. State facts that can be checked, with dates, and correct them when they change. Give a competitor credit where it wins. Our view: a list that admits your product isn't best for everything is more believable to readers, and its facts are more likely to agree with independent sources that engines compare it against.

18How do you measure a best-of page?

Track three things: whether engines cite the page for its target prompts, whether your brand is named in those answers, and where in the list it appears. Run each prompt several times, because answers vary; across engines, we saw changes for 11 of 28 prompt-engine pairs we retested. Add Search Console data for the page, and referrals from AI assistants in analytics.

19What evidence convinces readers and engines?

First-hand testing, described specifically: what you tested, for how long, with what data, and what you found. Screenshots and measurements help readers; for engines, the findings need to be in text. Google's reviews guidance asks for evidence of experience and for explaining what sets each product apart. A sentence like "Competitor B extracted line items correctly from 47 of 50 sample invoices in our test" is evidence. "Competitor B has great accuracy" is an opinion with nothing behind it.

20How do you choose which products to include?

Include the products buyers in that use case actually consider. Check which names engines already mention for your target prompts, which ones appear on review sites for the category, and which ones your sales team hears about. Leaving out a well-known option makes the list look incomplete or biased; engines answering the same question will mention it anyway.

21How should you structure each pick?

Use the same pattern for every product, so readers can scan and engines can extract. Start with the "best for" label and a one-sentence reason. Follow with two or three findings from your testing, then the main limitation, then price with the date you checked it. Keep it in that order every time. Consistent structure makes the comparison fair, and it makes each section usable on its own when an engine quotes only part of the page.

22How long should a best-of page be?

Long enough to give each pick a real reason and its evidence, usually 150 to 300 words per product plus the criteria and method. Five well-supported picks beat fifteen thin ones. If you need more products, split the page by use case rather than stretching one list.

23Who should own best-of pages?

Content owns the structure and writing. Someone with hands-on product knowledge, often product marketing or a solutions engineer, owns the testing and the facts. Legal reviews claims about competitors, especially where pages are seen in the EU, where comparative advertising rules apply.

24Common questions

Should a vendor publish a best-of list? It can, if the testing is real, the criteria are published and the vendor's interest is disclosed.

How often should lists be updated? Whenever a pick's price, features or availability changes, and at least quarterly, with the update date shown.

Do affiliate links hurt a best-of page? Not on their own, but disclose them and make sure the picks would be the same without them.

Should picks be numbered? Number them only if the order reflects your scoring. Otherwise use labelled sub-picks, which suit use-case questions better than a single ranking.

Should a best-of page link to the products? Yes. Link each pick to its official page so readers and engines can check the facts you state.

25Checklist

  • Specific use case
  • Published criteria and weights
  • Evidence and a drawback per pick
  • Own-product and affiliate disclosure
  • Update schedule

Questions people ask

3 answers

Are 'best X for Y' pages spam?

Not inherently. Google's guidance asks that ranked lists explain why each pick is best with first-hand evidence and contain enough useful content to stand alone. Thin lists produced at scale risk violating spam policies.

Can I include my own product in a best-of list?

Yes, if you disclose it and judge it by the same criteria. Readers and engines distrust lists where the author's product always wins.

Do affiliate links hurt list pages?

Google says reviews often use affiliate links and points to its affiliate programme guidance. Disclose them and focus on original value.

Sources

  1. Google, Write high quality reviews; Reviews system; Spam policies; AI features and your websitedevelopers.google.com
Written and reviewed by Dipuck JonesHow we research and testReport an error