Should You Remove Navigation From Landing Pages? What the Tests Show
Removing landing page navigation doubled signups in one famous test. We sort the A/B tests, NN/g and Baymard research, and a test plan for your paid pages.
On this page
Should you remove navigation from landing pages? The conversion rate studies people quote say yes, and the most famous one doubled signups on a kitchenware retailer's registry page. Read the original write-ups and the evidence gets thinner: a few vendor case reports that leave out sample size, test length and confidence level, next to usability research that argues for giving visitors a way back. Our view is that dropping the header menu is a sensible default for a single-offer paid page and a weak one for any page where visitors still need to decide whether to trust you.
Where the claim comes from
Most of the research evidence people cite for removing navigation from a landing page traces back to two write-ups. The first is a case study published by Wingify, the company behind the VWO testing tool. Yuppiechef, a South African retailer of premium kitchen tools, removed the main navigation bar from its Wedding Registry landing page and left the rest of the page identical. In the post's words, "The conversion rate jumped from 3 to 6%." The page now shows an October 27, 2026 date, but its structured data gives a first publish date of December 15, 2011, so the most repeated number in this debate comes from a test about fifteen years old.
The second is a test HubSpot ran on five high-traffic landing pages of its own, removing the top navigation, footer navigation and social share links. The post adds two more examples, from Career Point College and SparkPage, and concludes that "although the percentages are fairly small, removing the top nav doesn't hurt conversion rates, so it's a change worth making." It shows an updated date of August 26, 2025 and does not say when the test ran.
What the published A/B tests show
Every A/B test result on removing landing page navigation that we found is positive, which likely reflects which tests get written up. The table sets out what each one tested and what each report leaves out.
| Study | Page tested | Traffic | What changed | Reported result | Missing from the report |
|---|---|---|---|---|---|
| Yuppiechef, via Wingify (2011) | Wedding Registry landing page, registry signups | AdWords search and display, Facebook, other advertisers, directories, organic, direct | Main navigation bar removed | Conversion rate from 3% to 6% | Visitors, test length, confidence |
| HubSpot, top of funnel | Three high-traffic free content pages | Not stated | Top nav, footer nav and social share links removed | "0-4% bump" in conversion rate | Visitors, test length, significance |
| HubSpot, middle of funnel | Two high-traffic pages offering software exposure and a sales connection | Not stated | Same as above | "16% and 28% lift" | Visitors, test length, significance |
| Career Point College, cited by HubSpot | Not stated | Not stated | Top nav removed and form layout changed | Conversion rate up 336% | Test design; the two changes are not separated |
| SparkPage, cited by HubSpot | Not stated | Not stated | Top nav removed | From 9.2% to 17.6% over one month | Test design, visitors |
HubSpot's free content pages moved 0–4%, while the two pages closer to a sales conversation gained the most, and the post offers no explanation for the gap. None of the reports says how many visitors saw each version, so a real doubling cannot be told apart from a lucky fortnight, and Career Point's 336% bundles a form redesign with the navigation change.
What Nielsen Norman Group's research says
Nielsen Norman Group's usability research on navigation and landing pages does not measure conversion rates, and we found no NN/g study that removes navigation from a landing page and reports the result. What it offers is a model of trust. In a 2016 article, Aurora Harley of Nielsen Norman Group cites four credibility factors from a 1999 list by Jakob Nielsen, "design quality, up-front disclosure, comprehensive and current content, and connection to the rest of the web," and writes that the landing-page content and the main navigation both need to be well organized. She also recommends prominently displaying contact information, with the utility navigation as a good place for it.
An older article by Jakob Nielsen at Nielsen Norman Group, from January 2000, gives both camps something. He observed that users "look straight at the content and ignore the navigation areas when they scan a new page," so a menu on a focused page may get little attention. He also notes that users often reach pages through search or other routes that bypass the higher-level navigation pages, and argues that pages without structural links become orphans.
What Baymard Institute found in checkout testing
Baymard Institute's navigation research that bears on landing pages comes from checkout usability, where it recommends taking the main menu away. A guideline from Baymard Institute on "Enclosed Checkout" design states the issue: "Users can be distracted during the checkout process by dominant noncheckout elements such as the main site navigation." The full guideline sits in Baymard's paid Premium database, and the public half of the advice matters as much. A cart abandonment guide from Baymard Institute, updated November 2, 2026, recommends an "enclosed checkout design that removes main site navigation while providing a clear way to return to the shop if needed." The checkout research behind them draws on usability testing: Baymard Institute cites "25 rounds of qualitative usability testing with 4,400+ test participant/site sessions," a method that shows why shoppers stall and leaves the size of any conversion lift unmeasured. We found no Baymard study of header navigation on paid landing pages, so the checkout guideline is the closest evidence to borrow.
Which pages need navigation to earn trust
Google's own advice for ad landing pages sits in the middle. Google asks advertisers to "Choose a landing page that closely matches your ad and keywords" and to "Make it quick and easy for customers to perform the action you want them to take," and its mobile tips include "Use simple navigation" and "Make it easy to contact you." Read alongside the NN/g trust factors, that advice sorts pages into three groups.
- Remove the menu on single-offer pages built for one campaign, where the visitor already knows the brand: content downloads, event registrations, offers to retargeted visitors. Yuppiechef's registry page fits here.
- Trim the header on high-consideration offers such as demo requests, quotes and expensive purchases, and wherever cold traffic meets a brand it has never heard of. Keep the logo, a contact route and a few trust links (about, reviews, pricing), which is Baymard's clear way back applied outside checkout.
- Keep full navigation on pages that also rank in organic search or work as research hubs, such as comparison and category pages, because those visitors arrive deep in the site and use the menu to find their bearings.
A test plan for one paid landing page
With only vendor case reports on record, the evidence that should decide your page is a test on your own traffic.
- Pick one page with one offer and steady paid traffic, and freeze ad copy, targeting and bids for the length of the test.
- Change only the navigation: remove the header menu and footer link lists, and keep the logo, contact route and privacy link. Editing the form at the same time recreates the Career Point problem.
- Split traffic 50/50 by visitor, so returning visitors keep seeing the same version.
- Use conversions per visitor across the whole session as the main metric, counting conversions completed on other pages, and follow the result through to qualified leads or orders. Page-level form fills overstate the win, since some visitors who used the menu would have converted elsewhere, the same reasoning that favors measuring incrementality over attributed conversions.
- Set the sample size and end date before launch, run whole weeks, and resist stopping on an early spike.
Worked example (hypothetical)
Every figure in this example is an assumption. A demo-request page gets 3,000 paid visitors a week and converts at 4.0%. Detecting a lift to 4.8% (20% relative) at 95% confidence and 80% power, using a two-sided test of two proportions with an even 50/50 split, takes about 10,300 visitors per version, roughly 20,600 in total, or about seven weeks. Detecting a lift to 4.4% (10% relative) takes about 39,500 visitors per version and about 26 weeks. If HubSpot's 0–4% figures are relative lifts, gains that small would take most paid pages far longer than a quarter to confirm.
Judge the winner in money terms. A higher conversion rate lowers customer acquisition cost at the same spend, and our ROAS calculator shows what the lift is worth against your break-even ROAS. A page that converts better also gives you room to scale paid ads without killing ROAS, and teams that want the test built and run for them can talk to our paid media group. A tie on a properly sized test is a useful answer too, since it means the menu was never where visitors were leaking and the next test belongs on the offer, the form or the page speed.
Sources
- Wingify: Yuppiechef navigation menu case study
- HubSpot: landing page navigation test (updated Aug 26, 2025)
- Baymard Institute: Guideline #606, "Enclosed Checkout" design
- Baymard Institute: How to reduce cart abandonment (updated Nov 2, 2026)
- Baymard Institute: checkout usability research
- Nielsen Norman Group: Trustworthiness in Web Design (May 8, 2016)
- Nielsen Norman Group: Is Navigation Useful? (Jan 8, 2000)
- Google: Optimize your ads and landing pages
