Why this is hard to get right
The Problem With Writing Competitor Comparison Pages
Maria is a product marketer at a mid-market HR tech company. She's been asked to build a "Us vs. Rippling" comparison page before their next sales quarter. The page needs to live on the website, rank for competitive search terms, and give sales reps something credible to share in deals.
She knows her product cold. She's sat through dozens of competitive calls. She understands where they win and where they lose. But when she opens a doc and starts writing, the copy comes out flat. Either it sounds like a sales pitch with nothing to back it up, or it reads like a feature checklist that no buyer would finish.
She tries feeding a basic prompt to an AI assistant: "Write a comparison page that shows why we're better than Rippling." The output is confident, but it's full of vague claims like "intuitive interface" and "enterprise-grade security" with no specifics. Worse, some of the feature comparisons are wrong because the AI hallucinated capabilities neither product actually has.
She revises by adding more context in a follow-up message, but the AI doesn't weave it together cleanly. The structure is still off. The tone veers between diplomatic and condescending. The CTA at the end is generic. She's spent two hours and the draft still isn't usable.
The real challenge is structural. A good comparison page has to do several hard things at once: be specific enough to be credible, fair enough to be trusted, and opinionated enough to be persuasive. It needs a format that matches how buyers actually evaluate tools. And it needs to stay on the right side of legal and compliance review, which means avoiding unverifiable superlatives and unfair characterizations.
When Maria builds a proper prompt — one that specifies the audience (IT managers at 200–2,000 person companies), names the real competitor (AcmeSoft), lists verified proof points (2-week onboarding, SOC 2 Type II, 99.95% uptime), acknowledges one genuine limitation (no on-prem), and defines the page structure (8 sections including a comparison table, objection FAQ, and hard CTA) — the output is categorically different.
The AI produces a draft that leads with buyer priorities, not marketing fluff. The comparison table shows real feature differences. The FAQ addresses objections she's actually heard on calls. The CTA is specific. The disclaimer is in place. Her legal team approves it with two minor edits instead of a complete rewrite.
That's the difference a structured prompt makes. Not just better copy — a faster path from blank page to published page that actually converts.
Common mistakes to avoid
Naming Competitors Without Providing Facts
If you name a specific competitor but don't supply verified feature data, the AI fills the gaps with assumptions. This produces comparison tables with inaccurate claims — a legal and credibility risk. Always input what you can actually prove. If you lack data on a competitor's specific limitation, leave that row blank and let your copy team verify it first.
Skipping the Audience Definition
Without a named buyer persona, the AI defaults to a generic audience and optimizes for the wrong priorities. A security-focused IT manager and a cost-focused CFO need completely different comparison frames. Define the role, company size, and primary buying criteria — that single addition reshapes what the AI emphasizes in every section.
Asking for "Better" Without Defining the Win Criteria
Prompts like "show why we're better" give the AI no anchor. It defaults to subjective language — "easier to use," "more powerful" — that buyers distrust and legal will reject. Specify your 3–5 must-win criteria (onboarding speed, compliance certifications, support SLA) so the AI structures the comparison around what actually matters in your deals.
Leaving Out Limitations
Comparison pages that omit your product's genuine weaknesses read as one-sided and erode trust. Buyers already know your limitations from review sites. Including one or two honest trade-offs in your prompt signals fairness and prompts the AI to handle objections proactively rather than ignoring them.
Not Specifying the Page Structure
Without a defined layout, the AI produces a blob of paragraphs — no table, no FAQ, no CTA. You'll spend as much time reformatting as you would writing from scratch. Name every section you need: hero, comparison table, proof section, FAQs, CTA. The AI fills structure you give it; it won't invent the right one.
Ignoring Tone and Compliance Guardrails
Comparison pages frequently get flagged for legal review when they make unverifiable claims or disparage competitors by name. Prompts that don't set a tone guardrail (e.g., "confident, fair, and factual — no insults") often produce copy that legal will reject entirely. Bake the tone requirement and a disclaimer instruction directly into the prompt to avoid rewrites.
The transformation
Write a comparison page that shows why we’re better than our competitors.
You’re a **B2B SaaS conversion copywriter**. Write website copy for a **“Us vs Competitor”** page. 1) **Audience:** IT managers at 200–2,000 employee companies evaluating vendor options. 2) **Competitor:** AcmeSoft. 3) **Tone:** confident, fair, and factual. No insults. 4) **Inputs:** Our strengths: 2-week onboarding, SOC 2 Type II, 99.95% uptime, live support in <2 minutes. Limits: we don’t offer on-prem. 5) **Format:** 8-section page: hero, who it’s for, key differences table, feature-by-feature comparison, proof, FAQs, objection handling, CTA. Include **one disclaimer** on assumptions. Keep it under **900 words**. End with a **“Book a demo”** CTA.
Why this works
Role Assignment Sharpens Output
The After Prompt opens with "You're a B2B SaaS conversion copywriter." That single instruction shifts the AI's entire frame. It stops writing like a content generalist and starts optimizing for conversion — tighter claims, stronger CTAs, buyer-centric language. Role assignment is the fastest way to elevate the register of any marketing copy prompt.
Named Audience Drives Relevance
Specifying "IT managers at 200–2,000 employee companies" in section 1 of the After Prompt tells the AI which pain points to lead with — security (SOC 2), reliability (uptime), and onboarding speed. Without that anchor, the AI guesses at buyer priorities. Named personas prevent generic copy and produce language that resonates with the actual decision-maker.
Proof Points Replace Hype
Section 4 of the After Prompt lists specific, verifiable inputs: "2-week onboarding, SOC 2 Type II, 99.95% uptime, live support in under 2 minutes." These replace adjectives like "fast" or "secure" with claims buyers can evaluate. Concrete numbers anchor the AI's language and give legal reviewers something defensible to approve.
Structured Layout Produces Usable Sections
Section 5 of the After Prompt defines an explicit 8-section format. That means the AI outputs a hero block, a comparison table, a proof section, FAQs — each as a distinct, editable unit. Defined structure prevents unformatted blobs and lets you hand individual sections to designers or engineers without reformatting.
Output Controls Keep Drafts On-Brief
The After Prompt ends with three hard constraints: under 900 words, one disclaimer on assumptions, and a "Book a demo" CTA. These aren't suggestions — they're guardrails. Word limits force prioritization. The disclaimer protects from compliance risk. The named CTA ensures the page ends with a specific action, not a vague sign-off.
The framework behind the prompt
The Strategy Behind Competitor Comparison Pages
Comparison pages sit at the highest-intent intersection of SEO and sales enablement. A buyer searching "Product A vs. Product B" has already defined their short list. They're not discovering categories — they're making a final call. That makes this content type uniquely valuable and uniquely risky.
The psychology of comparative evaluation follows what researchers call a "trade-off contrast" model: buyers don't evaluate products in isolation. They evaluate them relative to a reference point. Whoever frames that reference point controls the comparison. A well-structured comparison page lets you set the frame — defining which criteria matter, in which order, and why.
This connects to the Jobs to Be Done (JTBD) framework popularized by Clayton Christensen. Buyers don't choose software; they hire it to do a specific job. Effective comparison pages identify that job first ("IT managers need compliant, fast-to-deploy tools") and then show how each product performs against that job — not against a generic feature list.
From a conversion standpoint, comparison pages follow the AIDA model (Attention, Interest, Desire, Action) but compress the journey. Attention comes from the search query itself. Interest builds through the feature comparison. Desire peaks in the proof and FAQ sections where objections are resolved. Action happens at the CTA. Every section of a well-structured comparison page maps to one of these stages.
Legal risk is a genuine constraint. The FTC and equivalent bodies in the EU regulate comparative advertising. Claims must be truthful, substantiated, and non-deceptive. Superlatives like "best," "fastest," or "most secure" without supporting evidence create liability. This is why specifying verifiable proof points and including a disclaimer in your prompt isn't just good copywriting — it's risk management.
Finally, comparison pages perform best for SEO when they match search intent precisely. Google categorizes these searches as "commercial investigation" queries and rewards pages that genuinely help users decide — not pages that just pitch one option. Neutral framing, structured headers, and FAQ schema markup all signal to search engines that the page is a decision-support resource, not a promotional asset.
Understanding these dynamics is what separates a comparison page that converts from one that just exists on your site.
Prompt variations
You are a positioning strategist and conversion copywriter.
Write a "Us vs. Notion" comparison page for a new project management tool called Stackr, targeting solo consultants and two-person agencies.
Audience: Independent consultants billing 30+ hours per week who currently use Notion but complain about setup time and lack of client-facing views.
Our edge: Pre-built consulting templates, one-click client portals, and built-in time tracking. Setup in under 10 minutes.
Our limitation: No native integrations with enterprise tools like Salesforce or SAP.
Tone: Direct and honest. Acknowledge Notion's strengths before explaining the trade-offs.
Structure: Hero headline, who this is for, 5-row feature comparison table, 3 customer quotes, 3 FAQs, CTA to start a free trial.
Keep it under 700 words. Add one disclaimer noting comparison data reflects publicly available information as of today.
You are a sales enablement writer.
Write a one-page competitive comparison document that sales reps can attach to follow-up emails after a demo call where the prospect mentioned they're also evaluating Salesforce.
Audience: VP of Sales at B2B companies with 50–500 reps who prioritize forecast accuracy and onboarding speed.
Our product: Forecastly — a revenue intelligence platform.
Our edge: 3-day implementation, direct CRM sync in under 15 minutes, dedicated onboarding rep for first 90 days, GDPR and SOC 2 compliant.
Salesforce's known trade-offs: Typically 8–16 weeks to implement, high admin dependency, add-on cost for advanced AI features.
Tone: Peer-to-peer, confident, and factual. No disparagement.
Format: Three sections — Quick Summary, Side-by-Side Table (6 rows), Why Teams Switch. End with a "Talk to a rep" CTA.
Keep it under 400 words. Flag any claims that require verification before sending.
You are an SEO content strategist and B2B copywriter.
Write a long-form "Zendesk vs. Freshdesk" comparison page designed to rank for the search query "Zendesk vs Freshdesk for small business."
Primary audience: Customer support managers at e-commerce companies with 5–30 support agents comparing tools on a budget.
Frame the comparison neutrally. Do not favor either product in the opening. Present trade-offs honestly and help the reader decide.
Key comparison dimensions: Pricing at 20-agent scale, ticket automation depth, reporting dashboards, onboarding complexity, and mobile app quality.
Structure:
- Introduction with target reader definition
- Summary comparison table (5 dimensions)
- Deep-dive section per dimension (100 words each)
- Who should choose Zendesk
- Who should choose Freshdesk
- Bottom-line verdict
Target 1,200 words. Use H2 and H3 headings. Include a note that pricing data should be verified on vendor websites before purchase decisions.
You are a customer success copywriter.
Write comparison page copy for a situation where existing customers are being approached by a competitor (HubSpot CRM) and asking "why shouldn't we switch?"
Audience: Current users of our platform, Clientbase, who are mid-contract and have received an outbound pitch from HubSpot.
Our strongest retention arguments: Dedicated CSM included in all plans, no seat-based pricing, unlimited data history, and migration support included at no cost if they ever do switch.
HubSpot trade-offs to acknowledge fairly: Broader ecosystem, more third-party integrations, stronger brand recognition.
Tone: Warm, honest, non-defensive. Acknowledge HubSpot is a strong product. Focus on fit, not fear.
Structure: Opening empathy statement, "What you'd gain vs. what you'd give up" table, 3 FAQs from real customer conversations, retention CTA to schedule a review call with their CSM.
Keep it under 600 words. Do not use the phrase "best in class."
When to use this prompt
Product marketers building comparison pages
Create a credible “Us vs X” page for high-intent traffic without overstating claims.
Sales teams supporting competitive deals
Generate on-brand comparison copy reps can reuse in follow-ups and one-pagers.
Customer success leaders handling churn risk
Write fair competitive positioning content to address “why not switch” objections.
Founders refining positioning fast
Draft a structured comparison page for a new category narrative before a launch.
Pro tips
- 1
Name the buying stage so the copy matches intent, not curiosity.
- 2
List only claims you can prove because compliance issues kill trust.
- 3
Add 3 customer objections you hear on calls so the FAQ does real work.
- 4
Specify your “must-win” criteria because it shapes the comparison table.
The comparison table is the most-read element on any competitor page — and the most frequently botched. Here's how to prompt for one that holds up to scrutiny.
Choose rows that reflect buyer priorities, not your product roadmap. If your buyer is a security-focused IT manager, lead with compliance, uptime, and support SLA — not design or integrations. In your prompt, include: "Structure the comparison table around the top 5 buying criteria for this persona."
Limit rows to 5–8. More than 8 rows dilutes attention and invites claims you can't verify. Every row you add is a potential compliance flag.
Use three-state values. Instead of just check marks and X's, prompt for: "Yes / No / Partial" values with a one-line explanation in each cell. This reads as fairer and more informative than a binary comparison.
Name the sources. Add a table footnote instruction: "Add a footnote citing where each competitor data point comes from (vendor website, G2, Gartner, etc.)." This forces the AI to flag inferences and gives your team a clear verification checklist before publishing.
Test the table with a skeptical reader. Before publishing, ask someone unfamiliar with your product to read the table and identify any claim that feels unsubstantiated. Those are your legal review targets.
Comparison pages serve two audiences: the buyer reading them and the search engine indexing them. With a few prompt additions, you can optimize for both simultaneously.
Target a specific keyword phrase. Add to your prompt: "Optimize the page for the search phrase '[your product] vs [competitor] for [use case or industry].' Use this phrase naturally in the H1, first paragraph, and at least two subheadings."
Prompt for semantic coverage. Include: "Address related questions buyers search for, including pricing differences, integration options, and which product is easier to set up. Use H2 headers for each topic."
Control internal linking opportunities. Add: "End each major section with a one-sentence transition that could support an internal link to a related page (e.g., our pricing page, security overview, or case study)."
Request schema-ready FAQ format. Structured FAQs at the bottom of comparison pages frequently appear in Google's People Also Ask results. Prompt for: "Write FAQs in a direct question-and-answer format, each answer under 60 words, suitable for FAQ schema markup."
These additions don't require a separate SEO prompt. They layer search optimization into the same structured output your copy team already reviews.
The competitor comparison format isn't limited to website pages. The same prompt structure — with adjustments to audience, length, and format — produces content across your entire competitive sales motion.
Top-of-funnel (SEO page): 1,000–1,500 words, neutral framing, H2 structure, FAQ section. Goal: rank for "[product] vs [competitor]" searches and capture undecided buyers.
Mid-funnel (sales deck slide): 200–300 words of speaker notes plus a 5-row comparison table. Goal: give reps a credible, rehearsed answer when a prospect names a competitor in a discovery call.
Bottom-of-funnel (follow-up email): 150–250 words, first-person from the rep, referencing the specific objection raised on the call. Goal: address the "we're also looking at X" hesitation before the next meeting.
Retention (customer success): 400–600 words, warm tone, acknowledging the competitor's strengths before addressing fit. Goal: reduce churn risk when an existing customer is being actively courted.
For each format, adjust the audience, tone, word limit, and CTA in your prompt. The structural logic stays the same. You're not building four different pages — you're running one comparison framework at four different buying stages.
When not to use this prompt
Don't use this prompt pattern in these situations:
-
When you lack verified proof points. If your inputs section would be empty or filled with adjectives rather than metrics, the AI will hallucinate specifics. Build your evidence base first — then write the page.
-
When the competitor is much larger and well-known. Directly naming Salesforce, HubSpot, or Microsoft in a comparison page invites intense scrutiny and can backfire if buyers see your claims as overreach. Consider a "built for companies like yours" positioning angle instead, without naming the giant directly.
-
When you're pre-product-market-fit. Comparison pages lock in positioning. If your differentiation is still evolving, publishing a structured comparison locks you into claims you may need to retract. Wait until your wins are consistent and repeatable.
-
When a third-party review would carry more weight. Buyers know you're biased. For high-skepticism categories, a G2 review summary, analyst report, or customer case study will outperform self-authored comparison copy. Use this prompt for self-published pages, but don't expect it to replace independent validation.
-
For internal competitive intelligence. This prompt produces external-facing marketing copy. For internal battle cards or sales coaching materials, use a separate prompt structure focused on objection handling scripts rather than web copy.
Troubleshooting
The comparison table includes claims I can't verify about the competitor
Add this instruction to your prompt: "Only populate competitor rows with information from publicly available sources. If a data point is uncertain, insert '[Verify before publishing]' in that cell instead of guessing." This shifts the AI from confident hallucination to transparent flagging, giving your team a built-in review checklist.
The output sounds one-sided and promotional, not credible
Strengthen your tone guardrail and add a fairness instruction. Replace your tone section with: "Acknowledge one genuine strength of the competitor before presenting trade-offs. Avoid superlatives without supporting data. Write as if a skeptical analyst is reviewing every claim." Also, make sure your inputs section includes at least one honest limitation of your own product.
The page structure doesn't match what I asked for
Number your sections explicitly and add an output format instruction. At the end of your prompt, include: "Output each section as a separate H2 block with the section title bolded. Do not combine sections. Follow the order listed exactly." Vague structure requests like 'include a comparison table' give the AI too much discretion — numbering forces the layout you need.
The FAQ section doesn't address real buyer objections
Seed the FAQ with your actual objections. Add to your prompt: "Base the FAQ section on these 3 objections heard on sales calls: (1) 'Your onboarding takes too long,' (2) 'We already use Salesforce integrations,' (3) 'Your pricing is higher per seat.'" Generic FAQ prompts produce generic questions. Real objections produce answers that actually handle deals.
The word count is way over the limit I specified
Add a hard constraint at the end of your prompt: "Stop writing at 900 words. If you reach the limit before completing all sections, shorten the feature comparison descriptions rather than cutting the CTA or FAQ. Prioritize the comparison table and CTA over body copy." Giving the AI a prioritization rule prevents it from cutting the wrong sections when space runs out.
How to measure success
How to Evaluate Your Comparison Page Output
Before sending the draft to design or legal, run it through these checks:
Credibility signals:
- Every claim in the comparison table is either sourced or flagged for verification
- The competitor's genuine strengths are acknowledged at least once
- No superlatives appear without a supporting metric
Structural completeness:
- All 8 sections (or your specified sections) are present and distinct
- The comparison table has 5–8 rows covering buyer-priority criteria
- The FAQ addresses real objections, not hypothetical ones
Conversion effectiveness:
- The hero headline speaks to the buyer's job, not your feature list
- The CTA is specific ("Book a demo") not generic ("Learn more")
- Word count falls within your specified range
Legal and compliance:
- A disclaimer is present noting data sources and potential for change
- No phrases like "best in class," "industry-leading," or "only solution" appear without evidence
- No competitor claims exceed what public sources support
If the draft clears all these checks, it's ready for legal review. If it fails two or more, revisit your prompt inputs — the issue is usually missing proof points or an undefined audience.
Now try it on something of your own
Reading about the framework is one thing. Watching it sharpen your own prompt is another — takes 90 seconds, no signup.
Turn your competitive positioning into a credible, conversion-ready comparison page — without the legal headaches.
Try one of these
Frequently asked questions
Be explicit in your prompt about what you know vs. what you're inferring. Add an instruction like: "Only include claims sourced from publicly available information. Flag any row in the comparison table that requires verification before publishing." This produces a draft with built-in review flags rather than confident-sounding hallucinations. Your team can verify flagged items before legal review.
Yes, but keep the structure tighter. In your prompt, name both competitors and limit the comparison table to 5–6 rows covering only your strongest differentiators. More than 6 rows across three columns becomes hard to read and harder for AI to populate accurately. Add an instruction: "If data for one competitor is unavailable for a given row, mark it as 'Unverified' rather than guessing."
Three prompt-level controls reduce legal risk significantly:
- List only claims you can document in your inputs section
- Include a tone guardrail: "Confident, fair, and factual. No superlatives without evidence. No disparagement."
- Add a disclaimer instruction: "Include one disclaimer noting that comparison data reflects publicly available information and may change." These three steps won't eliminate legal review, but they dramatically reduce the number of edits required.
Replace the tone instruction in section 3 of the prompt with your exact standard. For aggressive: "Confident and direct. We are the clear choice — lead with our wins and address trade-offs briefly." For neutral: "Balanced and editorial. Present both products fairly. Let the evidence speak." The AI follows explicit tone instructions closely — vague words like "professional" give it too much latitude.
Add a constraint to your prompt: "Avoid adjectives without data to support them. Do not use the phrases 'best in class,' 'industry-leading,' or 'world-class.' Every claim must be followed by a specific metric or example." Promotional drift usually happens when the AI fills gaps with marketing language. Closing those gaps with specifics is the fastest fix.
One page per competitor performs better for both SEO and conversion. High-intent buyers searching "our product vs. competitor name" want a focused answer, not a multi-competitor overview. Separate pages also let you tailor the comparison table and FAQ to that specific competitive dynamic. Use the multi-competitor format only for internal sales decks or analyst briefings.
Replace feature-based inputs with outcome and process-based inputs. For example:
- Instead of: "SOC 2 Type II, 99.95% uptime"
- Use: "Average client ROI of 3.2x in 6 months, dedicated project lead on every engagement, fixed-fee pricing with no scope creep clauses" Also replace the feature table with a process or outcome comparison table. The structure works for any business model — the proof points just look different.
For SEO-first pages targeting competitive search queries, aim for 1,000–1,500 words with H2/H3 structure — enough to cover key dimensions thoroughly. For conversion-first pages linked from paid ads or outbound emails, 600–900 words with a clear CTA performs better. Specify the goal and word limit explicitly in your prompt — the AI will calibrate density and depth accordingly.