HubSpot A/B testing: what's built in and how to test any HubSpot page

Where HubSpot's native tests fall short

The session rule is the biggest limit for B2B pages. Because HubSpot randomizes visitors again on each new session, a prospect who reads a landing page on Monday and returns on Thursday to fill in the form may see the other version on the second visit. That conversion then reflects a mix of both variants, which blurs the difference you're trying to measure, especially when buying decisions take more than one visit.

The other limits are about scope and reporting. Classic A/B tests compare exactly two versions; going beyond two means an adaptive test, which shifts traffic as it learns and requires a form on every variation. HubSpot's knowledge base lists performance metrics for test results but doesn't describe a significance calculation, so check a result with our significance calculator before choosing a winner. And all of it sits behind Professional tiers, so a team on a lower tier has no native page testing.

How to A/B test HubSpot pages with Otter

Otter has no HubSpot integration and no HubSpot-specific install guide. It runs on HubSpot CMS pages the same way it runs on any site: paste the standard Otter snippet into the site header HTML in HubSpot's website settings so it loads in the head of every HubSpot page, publish, and verify the installation from Otter. Otter only needs to add HTML to the page head, so confirm that your HubSpot plan and permissions allow editing the site header HTML before you start.

Once installed, you have two ways to build a test. The visual editor opens a live HubSpot page and changes headlines, copy, links, image URLs, styles and inserted content, with desktop, tablet and mobile previews. To test something the editor shouldn't touch, such as the fields in a HubSpot form, build a second HubSpot page with a different form and run a split URL test between the two. Split URL tests can carry UTM parameters through the redirect, so the variant page still receives the campaign's tags.

Assignment is sticky. Otter stores a random visitor ID in a first-party cookie for 365 days, so a returning prospect sees the same variant on every visit in that browser, and a form filled in days later still counts toward the right variant. Otter doesn't follow visitors from one device to another.

Goals for HubSpot landing page tests

Otter has no form-submit goal type and doesn't read HubSpot's form data, so pick one of these goals. If the form redirects to a thank-you page, a pageview goal on that URL counts completed submissions. A click goal on the submit button counts attempts, including ones that fail validation. If you already push form submissions to the GTM dataLayer or record them in GA4, Otter can use that dataLayer event or import the GA4 event as the goal.

Two things won't happen. Otter doesn't write test data to HubSpot contact records, and it doesn't use HubSpot lists or properties for targeting, because there's no integration. What Otter can target is the visit: UTM source, medium and campaign, traffic source, device, country, new or returning visitors, cookies and URL parameters. Otter can also send exposures and conversions to GA4 if you want to analyze them alongside your other data.

HubSpot A/B testing vs Otter: which fits

Use HubSpot's native tools if you're on Content Hub Professional or Enterprise, a two-version page test or an adaptive test with forms covers what you need, and you want results next to your contact data. Email tests stay in HubSpot either way, because Otter tests web pages, not emails.

Use Otter if your team is on a HubSpot tier without page testing, if visitors often return before converting and need to keep one variant, if you want more than two versions with a fixed split, or if you test pages on HubSpot and on another platform, such as a Webflow marketing site plus HubSpot landing pages, under one account. Otter's statistics are frequentist or Bayesian at a threshold you choose, with automatic correction for tests with several variants, and Starter costs $39/month for one site with unlimited visitors and tests. We make Otter, so compare it with HubSpot on a real page first. See Otter pricing and our A/B testing tools comparison.

What to test on HubSpot landing pages

HubSpot pages usually exist to capture a lead, so the best tests sit around the form. Test the headline against the ad or email that sent the visit, the offer (an ebook, a demo, a free consultation), the copy and proof next to the form, and where the form sits on the page. Test the number of form fields with two pages and two forms, and decide in advance whether success means more submissions or more qualified leads, since a shorter form can bring in more leads and weaker ones at the same time.

Thank-you pages are worth testing too: a meeting link, a case study or a product video can move a new lead further along. For more ideas, see the landing page A/B testing guide, and size the test before you start. A form page converting at 10% needs roughly 6,700 visitors per variant to detect a 15% relative lift at 95% significance and 80% power.

Pitfalls and SEO for HubSpot page tests

Don't run a HubSpot native test and an Otter test on the same page at the same time: HubSpot would split visitors between its two versions while Otter splits them again, and neither result would be clean. Avoid editing a page in HubSpot while an Otter test runs on it, because changing the control mid-test changes what you're measuring. Otter also hides each page until its test settings load, with a 3-second failsafe, so expect that short wait on every HubSpot page that carries the snippet.

For search, follow Google's website testing guidance. HubSpot's native A/B tests keep one URL, and so do Otter's visual tests, which change content with JavaScript, an approach Google lists as valid. For an Otter split URL test between two HubSpot pages, add a rel="canonical" tag on the variant page pointing at the original, keep the redirect temporary, don't show crawlers different content, and remove the losing page when the test ends. Otter doesn't assign known bots to a variant, so crawlers see the original.