How to create comparison landing pages to steal your competitors' Google traffic

A competitor comparison landing page is the page you send someone to after they searched for your competitor's name and clicked your ad instead. It names the competitor your ad wasn't allowed to name, shows what's different in one table, tells the truth about price, and asks for one thing. A lot of companies that bid on competitor keywords skip the page and send the click to their homepage or a generic demo form. That's the whole opportunity.

This post is the paid-traffic sequel to our landing pages for ads playbook. It covers what actually shows up when you search a competitor's name today, what the pages behind those ads look like, and how they differ from the "versus" pages built for SEO. Where it says "we checked", we did: the numbers come from our Y Combinator ads database and from a scrape of 50 live Google results pages this week.

How the traffic actually moves

The mechanism is simple and slightly unfair. Someone types "twilio" into Google. Twilio pays for the top ad, as every brand does on its own name. But Google's trademark policy doesn't restrict trademarks used as keywords, so anyone can bid on that search.¹ When we searched it in September 2026, right below Twilio's ad sat one from Allo, a phone-system startup, for the same query.

What the challenger can't do is put "Twilio" in the ad. Since July 2023, a trademark complaint from the owner blocks the named advertisers from using the mark in headlines, descriptions and the display URL.² So the ad has to sell on its own terms ("AI-powered 2nd phone number", "Gong alternative", "Switch now") and the actual comparison has to happen after the click. Landing pages fall under Google's destination and misrepresentation policies, which care about the page matching the ad and not pretending to be the competitor, and say nothing about naming them.³

That asymmetry is the argument for a dedicated page. The searcher had a specific product in mind. Your ad interrupted them with a vague promise. The only way the click pays off is a page that immediately says "yes, you were looking at X, here's how we're different." A homepage can't say that. A generic "book a demo" page can't either, and as you'll see, that's where most competitor traffic still lands.

Two more reasons the page matters more here than in a normal campaign. First, Google grades your landing page experience relative to the other advertisers on the same search, and on a brand search that includes the brand itself, whose page is the one the searcher actually wanted.⁴ Second, the intent is unusually good. In Grow and Convert's analysis of 95 client articles, "alternative" and comparison keywords converted at 8.43% and "versus" keywords at 5.45%, against under 1% for top-of-funnel content.⁵ Those are organic numbers, and someone intercepted by an ad on a brand search is more skeptical than someone who typed "alternative" themselves, so expect the paid version to convert lower. It's still the best intent you can buy.

What 898 YC advertisers and 50 live searches show

We keep a database of every ad run by the 6,055 Y Combinator companies, scraped from the Google, Meta and LinkedIn ad libraries in July 2026. The headline numbers are in the YC ads study, and the pages those ads land on are in the landing pages playbook. 898 of those companies run Google ads. Here's how many of them run this play:

One in five is more than I expected. Abacum alone runs ads against Planful, Vena, Datarails, Cube, Pigment and Adaptive. BrioHR, a Malaysian HR tool, runs "X users are switching" ads against eleven local competitors. Buy Me a Coffee bids on Patreon, Ko-fi, Gumroad, Substack, Memberful, Teachable and Ghost. Blink, an AI app builder, runs "ChatGPT advises, Blink ships" against ChatGPT, Cursor, Copilot and Codex.

Then we tried to open the pages. Google's Ads Transparency Center exposes the display URL an advertiser chose, not the final URL, so we fetched the 192 distinct display paths that looked like comparison pages. 146 of them, 76%, return a 404, and six wouldn't load at all. The path "abacum.ai/planful/alternative" is not a page. Most likely it's a line of ad copy, typed into the display-path field so the ad reads like a comparison, and the click goes somewhere else. Of the 40 that resolved, 15 redirected to a different page, three of them to the homepage.

The 40 pages that resolved are a mixed bag, and this is the part that matters:

To check the live picture beyond YC, we scraped the US Google results for 50 well-known software brand names on one afternoon (Twilio, Clay, HubSpot, Notion, Deel, Zapier, Asana and so on) and followed every ad click to its final URL. 25 of the 50 searches showed at least one ad. 15 showed an ad from a rival. Where those rivals sent the click:

So among companies that pay real money to intercept a competitor's search, roughly a third built a page for it, a third reused an SEO page or a blog post, and a third sent the click to the homepage or a generic form. The MenaITech campaign is literally named "Competitors" in its tracking parameters, and it lands on the homepage. Two more I collected while researching this, Linq's page for "twilio" and Reevo's page for "apollo", are both the generic-form pattern: a clean demo page that never mentions the product the person was searching for.

If your competitor campaign lands on a page that doesn't contain the competitor's name, you're paying a premium CPC for a homepage visit.

Seven pages, torn down

Here are the pages worth learning from, good and bad, scored on the things that decide whether a competitor click converts. Every one was live on September 14, 2026.

Alta vs Clay: the fairest table of the set

Alta sells an AI sales team; Clay is a data enrichment workspace. The page opens with "Clay is a workspace for GTM engineers. Alta is the AI GTM team for closers", which is the two-sentence fairness test passed in the first line. The table is grouped by job ("get to first meeting", "get the data", "run the outreach") and Clay wins two rows outright: 150+ data provider integrations and custom enrichment waterfalls, where Alta admits to a "curated set". That concession is what makes the other twelve rows believable.

Two things I'd change. The hero uses Clay's logo next to Alta's, which is exactly the thing trademark lawyers tell you not to do (more on that below). And there's no pricing, just a "predictable pricing" row against "credit-based". If the reason people leave Clay is bills they can't predict, a number would close harder than a checkmark.

Rippling vs Deel: dated data and a price

This is the page Rippling's ad sent us to when we searched "deel". The URL says what it is (/lp/paid-compare-deel), the menu is gone, the email field is above the fold, and the G2 comparison is stamped "data as of 06/10/2025", which is the right habit even if fifteen months is too long between refreshes. Lower down it lists EOR pricing from $499 a month, global payroll from $35 and contractors from $30. The weakness is the table: Rippling checks every box and Deel checks none, so the reader has no reason to trust it. The capture date helps; a refresh would help more.

Akiflow vs Motion: the textbook paid page

One of only two pages in the YC set on noindex, and it shows in the discipline. "Unlike fully automatic scheduling tools like Motion, Akiflow gives you AI support with full control" is the entire argument, and it's the right argument for someone who searched "motion alternative" because they're annoyed at Motion rescheduling their day. No table, which I'd normally flag, but the page is selling a philosophy rather than a feature count. Akiflow runs sibling pages for Reclaim, Sunsama and Todoist, each behind its own ads.

Buy Me a Coffee vs Patreon: price in the headline

"At only a 5% fee vs Patreon's 12%" is in the first paragraph, and the table below converts it into money: "for every $1,000 in support, you keep ~$950" against "$800 to $850". That's how to show price on a comparison page, as a consequence rather than a list. The "Compare using AI" button is the first one I've seen on a landing page. It hands the visitor to ChatGPT with the comparison prompt, which is either brave or a sign that they've measured what the AI answer says about them. The table itself has no row Patreon wins, and a couple of rows ("you're a ticket number") are opinion dressed as fact, which is the kind of thing that draws complaints.

Rillet vs Campfire and Camtasia vs VEED: SEO pages doing paid work

Rillet's page is well built: proof numbers in the hero, a logo strip, a table that starts with G2 scores (5.0 against Campfire's 4.8) and gives Campfire exactly one tie, on AI agents, and nothing it wins. But the full navigation, a "Rillet Recon" event banner and a "Log in" link are all exits, and the page is indexed, so the paid traffic and the organic traffic share it. Camtasia's VEED page has the same structure with a worse table: row after row of green checks for Camtasia and red Xs for VEED. Nobody believes a table like that, and a table shaped like the one the UK regulator banned in February 2026 (below) is a harder sell to legal.

Linq for "twilio": the page that forgot the competitor

This is the most common failure and it's not a design failure. The page is fine: no menu, form above the fold, proof chips, logo wall. It's just the wrong page. Someone searched "twilio", clicked an ad that couldn't say Twilio, and landed on "Build AI agents on iMessage" with a five-field form. Nothing tells them why this is the answer to the thing they typed. Reevo's page for Apollo traffic has the same problem with a full menu on top. Both would convert better with one added line under the headline: "Looking at Twilio? Here's what's different."

The anatomy of a page that converts competitor clicks

Putting the teardowns and the research together, here's what I'd build. Most of it is the standard landing page playbook, tightened for a visitor who arrived with someone else's brand in their head.

1. Logo and one button, no menu. HubSpot's A/B test found removing navigation lifted conversions 0 to 4% on top-of-funnel pages but 16 to 28% on mid-funnel pages.⁶ A competitor search is about as mid-funnel as it gets. 80% of the YC pages still keep the menu.
2. Name the competitor in the H1, then say who each product is for. "The control-first Motion alternative" and "Clay is a workspace for GTM engineers, Alta is the AI GTM team for closers" both do this in under twenty words. The ad couldn't say the name. The page must, immediately, or the visitor assumes they clicked the wrong thing.
3. Match the page to the intent behind the query. "X alternative" means they're leaving; lead with the frustration and the switching offer. "X vs Y" means they're evaluating; lead with the table. "X pricing" means they want a number. Build one page per intent rather than one page per competitor, because the headline has to mirror the query.
4. One table, six to ten rows, and let them win one. Conversion Factory's advice is to "acknowledge the dimensions where the competitor is better or even", with a test I like: read the page aloud to someone who works there and see if they nod or feel attacked.⁷ Alta's two conceded rows are why its other twelve land.
5. Show the price, with a date. HockeyStack's analysis of 31 million visitors across 80 B2B SaaS companies found transparent-pricing pages get fewer form fills (2.8% against 4.6%) but convert them to pipeline at 17.5% against 10.3%, so the lead you get is worth more.⁸ Stamp the competitor's price with the capture date, as Rippling does with its G2 data, and refresh it quarterly or remove it.
6. Proof from switchers, above the fold. A quote from someone who moved from the competitor beats a logo wall, because it answers the visitor's actual question: did anyone else do this and survive? If you use customer quotes, the FTC's reviews rule, in force since October 21, 2024, applies to them: real, attributable, not edited to change the meaning.⁹
7. A switching offer. Free migration, contract buyout credit, a trial with no card. The reason "alternative" queries convert above "versus" queries is that the searcher has already decided to move; the offer removes the last excuse.
8. One call to action, and it's the right one. Demo for enterprise, trial for self-serve, and pick based on your actual sales motion rather than habit. Linq's page has one button. Rillet's has a menu, a banner, a login link and a button.
9. Noindex, tracked, fast. The paid page carries noindex so it can't cannibalize your SEO comparison page, passes the click ID through to your CRM so you can judge the campaign on closed revenue rather than form fills, and loads in under two seconds on a phone. Give it 60 to 90 days of data before you judge it.

How SEO comparison pages differ from the paid ones

lemlist's /versus hub and Botdog's /comparison hub are the other kind of comparison page, and they're a useful contrast. Both are built to rank, not to catch a click from an ad, and almost every design choice flips.

lemlist's hub page is titled "Best lemlist Alternatives", a query about leaving lemlist that lemlist would rather answer itself. The sitemap behind it lists 30 "lemlist vs X" pages and four roundups. The spoke pages are around 2,000 words, name prices on both sides (lemlist from $69 a month, Instantly from $47 plus a separate credits subscription), and concede rows: Instantly scores 4.8 on G2 against lemlist's 4.6, and wins on price for pure email volume. That's the same honesty rule as the paid pages, applied at ten times the length.

Botdog's hub does something clever, three-way comparisons ("lemlist vs Waalaxy vs Botdog") that capture the "X vs Y" query without Botdog having to be one of the two names people already know. It also has a bug worth learning from, and since our team runs Botdog I can say so: every page in the hub carries a robots: index, follow tag, and every page's canonical points to the homepage. The canonical quietly cancels the index instruction. The title tags also still say "2024". Both are being fixed. If you build a hub, check the canonicals before you check the rankings.

Here's the split, dimension by dimension:

Three numbers explain why the SEO version has to be a program rather than a page. Ahrefs' 2025 study of a million URLs found only 1.74% of new pages reach the top ten within a year, and the average page ranking first is five years old.¹⁰ Minuttia's study of 1,000 "alternatives" keywords found G2 in 8.78% of results and Reddit in 5.19%, with ClickUp the only software vendor in the top twenty domains.¹¹ And Foundation's analysis of HubSpot's 71 comparison pages put them at 11,000 organic visitors a month, worth about $75,400 in equivalent ad spend, with "HubSpot vs Salesforce" getting roughly twice the searches of "Salesforce vs HubSpot".¹² Those hubs work because they're 30 to 70 pages that have been live for years, not because any one page is brilliant.

One more change for 2026: Google stopped showing FAQ rich results on May 7, so the FAQ schema that most SEO comparison pages carry no longer expands your listing.¹³ The markup still gets read by AI crawlers, which is why we keep it on this post, but it's no longer a reason to bolt an FAQ onto a comparison page.

My opinionated take: a paid comparison page should be built by whoever writes your ads, and the SEO version by whoever writes your content, and they should share the table and nothing else. The moment one page serves both, it grows a menu and 2,000 words, and the click you paid for lands in the middle of an essay.

The legal lines

"Can I do this?" is the question I get most, so here's the state of it in September 2026. I'm not a lawyer; these are the primary sources.

• Bidding is allowed. Google's trademark policy explicitly doesn't restrict trademarks as keywords. The mark in ad text or display URL is restricted for direct competitors once the owner complains, with exemptions for resellers and informational sites.¹
• US courts keep ruling that buying the keyword, on its own, isn't trademark infringement. The latest case is Lerner & Rowe v. Brown Engstrand, decided October 22, 2024. A law firm bought a rival firm's name as a Google keyword. The rival sued and showed 236 phone calls from people who had mixed the two firms up. The Ninth Circuit said 236 calls, set against the number of times the ad was shown, was too little confusion to matter, and the bidder won.¹⁴ Two earlier appeals courts reached the same result, in 1-800 Contacts v. Lens.com (2013)¹⁸ and Network Automation v. Advanced Systems Concepts (2011).¹⁹ The line the courts keep drawing: buying the keyword isn't confusion. An ad or a page that implies you're the competitor is.
• Naming the competitor is fine, their logo isn't. Nominative fair use in the US and the FTC's comparative advertising policy both encourage naming a competitor, on three conditions: you use only the name, only as much as needed, and nothing implies they endorse you.¹⁵ The logo fails the second test. Add a "not affiliated" line; Botdog's LinkedIn disclaimer is the model.
• In the EU and UK, the comparison has to be fair, not just true. Directive 2006/114/EC allows comparative advertising only if it objectively compares "material, relevant, verifiable and representative" features, doesn't denigrate the competitor and doesn't trade on their reputation.¹⁶ The UK's CAP Code copies that language.
• A table where the competitor loses every row has already been banned once. On February 25, 2026 the UK's Advertising Standards Authority ruled against Practice Pal, a school-sports software vendor, over a comparison table on its website. The table showed a cross next to features its competitor actually offered, and left out the areas where the competitor was equal or better. The regulator called that misleading and banned the table in that form.¹⁷ If your table has no row the competitor wins, ask whether that's because it's true or because you left those rows out.
• Testimonials are regulated. The FTC's Consumer Reviews and Testimonials Rule has been in force since October 21, 2024, with civil penalties for fake or misleading reviews.⁹ Quotes from switchers are the strongest proof on these pages, which is exactly why they need to be real.

The checklist

Before a competitor comparison page goes live behind an ad:

1. The competitor's name is in the H1 or the first line under it.
2. Two sentences say who each product is for, and the competitor's sentence is one they'd agree with.
3. The menu is gone. Logo, one button, legal links in the footer.
4. One table, six to ten rows, at least one row the competitor wins or ties.
5. Every competitor price or feature claim links to where you got it, with the capture date.
6. No competitor logo, screenshots or brand colors. Plain text only, plus a "not affiliated" line.
7. At least one quote from a customer who switched, with a real name.
8. A switching offer the visitor can act on today.
9. One call to action that matches your sales motion.
10. noindex on the paid page. A separate, indexed version at /compare/competitor if you want the organic traffic too.
11. Click IDs pass to the CRM. The campaign gets judged on qualified pipeline after 60 to 90 days, not on form fills after a week.
12. Negative keywords on the campaign: login, sign in, support, help, docs, status, careers, jobs, free, coupon, discount, promo code, and "pricing" unless the campaign has its own pricing ad group. Exact match expands to close variants, so read the search terms report weekly and keep adding.

That last one isn't about the page, but it's where most of the budget leaks. A lot of people who search "hubspot" want to log in to HubSpot.

Build one with your agent

Everything above is written up as a runbook at adside.ai/blog/competitor-comparison-landing-pages.md: the inputs to collect, the research steps, the page and campaign specs as JSON, the copy prompt verbatim, and the acceptance checklist. It's built for Claude Code, Cursor or any agent that can read a URL. Paste this to start:

Two things the agent will push back on, and should: it won't put a row in the table it can't verify on the competitor's own site, and it won't write the page until you've named one row the competitor wins. Both are features.

Frequently asked questions