Guide · Practitioner
Affiliate and Comparison Sites in AI Answers
Affiliate and comparison sites shape AI answers because engines often cite them for "best" and "X vs Y" prompts. Work with them on facts, not favours: give accurate data and testing access, and expect affiliate relationships to be disclosed ([eCFR](https://www.ecfr.gov/current/title-16/chapter-I/subchapter-B/part-255/section-255.5)) and paid links marked rel="sponsored" ([Google](https://developers.google.com/search/docs/crawling-indexing/qualify-outbound-links)). Google's spam policies also cover "thin affiliation": affiliate pages with copied product descriptions and no added value ([Google](https://developers.google.com/search/docs/essentials/spam-policies)).
Facts are from Google Search Central and the US eCFR, accessed 30 Oct 2026. This isn't legal advice. Points marked "Our view" are interpretation; examples are illustrative.
On this page
- Why do comparison sites matter for AI answers?
- What rules apply?
- How do you tell a useful comparison site from a thin one?
- How do you work with publishers honestly?
- What should your affiliate programme terms say?
- What should you never do?
- Should you publish your own comparison pages?
- How do you keep facts current across many sites?
- What do the best comparison sites need from you?
- How do you measure results?
- Exercise
- Checklist
- From MentionX data: which questions lead to comparison sites
- Worked example: Ledgerline cleans up its comparison listings
- How do you find the comparison sites that matter?
- What makes a comparison site trustworthy?
- Should you join affiliate programmes?
- How do you correct an inaccurate listing?
- How do you respond to a negative review?
- How do you keep many listings current?
- How should your own comparison page be written?
- What should an editor pack include?
- Which numbers show the work is paying off?
- Which sites should you skip?
- Common questions
- Related
- Sources
Checked on 31 October 2026. "Our view" marks MentionX opinion. Ledgerline is a fictional company used for illustration.
Why do comparison sites matter for AI answers?
Our view: a buyer asking "best payroll software for small businesses" gets an answer drawn from pages that compare options. Affiliate review sites publish many of those pages. If they're out of date about your pricing or features, the AI answer may be too. Find the ones that matter with the prompt-logging method in finding Reddit threads AI engines cite, logging comparison-site URLs instead.
What rules apply?
| Rule | Source |
|---|---|
Advertisements and paid placements should use rel="sponsored"; our view: treat affiliate links the same | |
| Thin affiliation, copying merchant descriptions without original content, is a spam policy | |
| Material connections, such as commissions, must be disclosed clearly | eCFR |
| Reviews should show evidence of first-hand testing and explain comparisons |
Our view: the publisher carries most of these duties, but as the advertiser you shouldn't ask anyone to break them.
How do you tell a useful comparison site from a thin one?
Our view:
| Signal | Useful site | Thin site |
|---|---|---|
| Testing | Screenshots, hands-on detail | Rewritten vendor copy |
| Disclosure | Clear affiliate statement | None or hidden |
| Dates | Prices and reviews dated | Undated |
| Balance | Pros and cons for each option | Only praise |
| Methodology | Explains how products are rated | None |
Google's review guidance describes similar signals: evidence of use, quantitative measurements where relevant, and explaining what sets products apart (Google).
How do you work with publishers honestly?
Our view:
- Give testing access with no conditions on the verdict.
- Send a fact sheet with dated prices, plans, limits and best-fit users.
- Offer standard affiliate terms if you run a programme; don't negotiate rankings.
- Correct errors with a public source, one fact at a time.
- Accept fair criticism; fix the product issue it reveals.
What should your affiliate programme terms say?
Our view:
- Affiliates must disclose the relationship clearly and near the link.
- Links must be marked
rel="sponsored"or nofollow (Google). - No fake reviews, false claims or invented test results.
- No bidding on your brand name in ads unless you allow it.
- Prices and claims must be kept current or dated.
- You can end the partnership for breaches.
What should you never do?
| Tactic | Problem |
|---|---|
| Paying extra for the "#1" spot without disclosure | Misleads readers |
| Supplying pre-written "independent" reviews | Deceptive; may breach FTC rules on fake reviews (eCFR) |
| Creating your own fake comparison sites | Deceptive; spam risk |
| Asking publishers to remove disclosures | Breaches disclosure duties |
| Threatening to end commissions over a fair verdict | Pressure that corrupts reviews |
Should you publish your own comparison pages?
Our view: yes, if they're honest. Say it's your product, compare on criteria buyers care about, date prices, and include fair pros and cons for competitors. Pages like this give engines a clear, first-party source for your facts. See getting into best-of listicles ethically.
How do you keep facts current across many sites?
Our view: keep a public pricing page and changelog, and email active affiliate partners whenever prices or plans change. A short quarterly note with the current fact sheet reduces stale comparisons. Track the ten most-cited comparison pages and check them after every pricing change.
What do the best comparison sites need from you?
Our view: editors running serious comparison sites are usually short of time, not goodwill. They need a working demo account, a named contact who answers within a day, current pricing in a format they can copy accurately, and honest notes on who your product isn't for. Meeting those needs makes it more likely your product is tested properly, and makes it less likely the review relies on old marketing copy.
How do you measure results?
Our view:
| Metric | How |
|---|---|
| Inclusion in cited comparison pages | Monthly prompt log |
| Accuracy of facts on those pages | Quarterly audit |
| Affiliate revenue and referral quality | Programme reports and analytics |
| AI answer mentions | Monthly prompt reruns |
AI answers vary by run, so look at trends over months.
Exercise
List the five comparison pages most cited for your top buyer prompt. For each, note whether affiliate relationships are disclosed, whether your prices are current and one fact you'd ask them to correct.
Checklist
- Cited comparison sites identified
- Fact sheet with dated prices shared
- Affiliate terms require disclosure and rel="sponsored"
- No paid rankings or supplied reviews
- Partners told about price changes
- Monthly tracking in place
From MentionX data: which questions lead to comparison sites
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.
Comparison sites matter most on buying questions. Our sample shows brands are named far more often on some question types:
| Intent | Brand mentioned | Share of voice |
|---|---|---|
| Brand-specific | 100% | 64.3 |
| Comparison | 91.9% | 44.6 |
| Recommendation | 29.2% | 26.8 |
| Best-of | 26.6% | 22.7 |
| Problem | 5.7% | 7.7 |
And comparison and review pages are a meaningful share of cited sources:
| Cited URL type | Share of 9,750 citations |
|---|---|
| Blog or article | 19.6% |
| Homepage | 10.0% |
| Review page | 2.7% |
| Docs | 2.6% |
| Pricing | 1.9% |
| Wikipedia | 1.6% |
Our view: comparison and recommendation questions are where affiliate and review pages tend to appear in answers. If you sell in a crowded category, being listed accurately on the reputable sites matters more than almost any single owned page.
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.
Worked example: Ledgerline cleans up its comparison listings
Ledgerline is a fictional accounts-payable tool. Figures are illustrative.
Ledgerline audits the comparison sites cited on its main buying prompts. Three list it with an old price, one omits it and one is a thin site that copies others. The team sends correct facts to the three editors, with a link to its pricing page and a short product summary. It contacts the editor of the site that omitted it, offers a trial account for review and makes no payment.
It ignores the thin site. Two of the three editors update their pages within a month. The reviewer tests the product and adds it, noting a missing feature honestly. The team adds that feature to its own product roadmap discussion and doesn't argue the point.
How do you find the comparison sites that matter?
Run your buying prompts in the engines your buyers use and log every cited URL. Group by domain. The sites that appear across several prompts and engines are the ones to prioritise. Add sites your sales team hears buyers mention. Ignore sites that never appear in answers, unless they're important to buyers for other reasons.
What makes a comparison site trustworthy?
Look for named authors, a stated testing method, dates on reviews, disclosure of affiliate relationships and balanced coverage that includes weaknesses. Sites that list every product as excellent, have no authors and copy descriptions word for word are thin. Engines may still cite them sometimes, but they're poor partners.
Should you join affiliate programmes?
Affiliate programmes are a normal business arrangement, and many reputable review sites earn that way. If you run one, make sure partners disclose it clearly, as consumer protection rules generally require. Don't pay for rankings or positive verdicts. Pay for referrals, and let the editor decide what to say.
How do you correct an inaccurate listing?
Email the editor with the specific error, the correct fact and a link to where it's stated on your site. Keep it short and polite. Don't demand changes to opinions or ratings; ask only for factual corrections. If the site has a corrections form, use it. Log the request and check back after a few weeks.
How do you respond to a negative review?
If it's factually wrong, ask for a correction with evidence. If it's a fair criticism, accept it, and consider whether your product or messaging should change. Arguing with fair reviews rarely helps and can make things worse.
How do you keep many listings current?
Keep a spreadsheet of every comparison site that lists you, with the facts shown, the date checked and the contact. When pricing or plans change, email every contact the same week with the update. Our view: a quarterly sweep of the top sites prevents most outdated facts from reaching AI answers.
How should your own comparison page be written?
Yes, if you do it fairly. A page comparing your product with alternatives, with accurate facts about each and clear notes on where others are stronger, helps buyers and can be cited. Name competitors only where you can state facts you can source, date the page and update it when products change. A one-sided page that claims you win on everything is unlikely to be trusted by readers or engines.
What should an editor pack include?
Make their job easy. Provide a short product summary, current pricing, a clear list of features and limits, screenshots they can use, and a contact who replies quickly. Offer trial access for hands-on testing. Don't send long marketing copy; editors want facts they can check. When something changes, tell them before they find out from a reader.
Which numbers show the work is paying off?
Track how often the comparison sites you work with are cited on your buying prompts, and whether those cited pages now name you accurately. Track your brand mentions in answers to the same prompts. Log every outreach and correction with a date, so you can see which changes were followed by gains in answers over the next month or two.
Which sites should you skip?
Skip sites that ask for payment to be listed without disclosing it, that publish fake user reviews, or that copy content from other sites. Working with them can damage trust, and they're rarely cited for long. If one of these sites lists you with wrong facts, a short correction request is still fine; just don't build a relationship or pay for placement there. Focus your limited outreach time on the sites that engines actually cite for your buying prompts. Review that list every quarter, as cited sites change.
Common questions
Do affiliate links hurt a site's value in AI answers? Not by themselves. Engines still cite well-made review sites that use affiliate links.
Can I pay to be ranked first? Paying for rankings without disclosure is deceptive. Pay for referrals, not verdicts.
What if a site won't update an old price? Make the correct price clear on your own site and on other sources, and keep asking politely.
Related
- G2, Capterra and review sites
- Listicle outreach template
Sources
- Google, Qualify your outbound links; Spam policies; Write high-quality reviews
- US eCFR, 16 CFR 255.5; 16 CFR 465.2
All accessed 30 Oct 2026.
Questions
Do affiliate sites influence AI answers?
Our view: often, for product and software comparisons, because engines cite comparison pages. Check your own prompts to see which sites are cited.
Do affiliate links need special markup?
Google asks that advertisements and paid placements be marked rel=sponsored. Our view: treat affiliate links as paid and mark them the same way, and disclose the relationship to readers.
Is a high commission a good way to rank higher?
Our view: tying placement to commission without disclosing it misleads readers. Compete on product facts, testing access and fair terms instead.