Hotjar A/B Testing: What It Can and Can't Do (And What to Use Instead)

If you searched for "Hotjar AB testing," here's the short answer: Hotjar doesn't run A/B tests. It never has. What it does is help you figure out what to test and why a test won or lost. That distinction matters, and most people confuse it.

I see this question on Reddit and indie hacker forums every week. Someone signs up for Hotjar expecting to split-test their headline against a variant, and they get confused when they can't find the feature. So let's clear it up, then talk about what to actually use.

What Hotjar Actually Does

Hotjar is a behavior analytics tool. It records what visitors do on your page. The core features are:

• Heatmaps: aggregated click, move, and scroll data shown as a color overlay on your page
• Session recordings: video-like playbacks of individual visitor sessions
• Surveys and feedback widgets: on-page polls and exit-intent questions
• Funnels: drop-off tracking between steps you define
• Rage click detection: flags users who click the same spot repeatedly out of frustration

None of these run a split test. Hotjar can't show 50% of visitors a green button and 50% a red one and tell you which converts better. That's not its job.

For a deeper look at what Hotjar covers and where it fits, I wrote a separate breakdown on what Hotjar shows for landing pages.

Why People Think Hotjar Does A/B Testing

A few reasons this confusion sticks around:

1. Hotjar markets itself as a "conversion" tool. Anything in the CRO category gets lumped together in people's heads.
2. It integrates with A/B testing platforms. Hotjar has native integrations with tools like Optimizely and VWO, which lets you watch session recordings of people who saw each variant. That makes it feel adjacent to testing.
3. Some competitors do both. Tools like VWO and Mouseflow blur the line by offering recordings and split tests under one roof. Hotjar doesn't.

So if you want to actually run a test, you need a different tool. Or a different tool plus Hotjar.

What Hotjar Is Good For in a Testing Workflow

This is where Hotjar earns its place. Think of it as the diagnostic layer that wraps around your A/B test.

Before the test: You watch session recordings to spot friction. Maybe people scroll past your hero without clicking. Maybe they hover over the pricing toggle for ages, then leave. These observations become test hypotheses. Without them you're guessing.

During the test: You filter recordings by variant. If you're running a test in another tool and pushing a variant cookie or URL parameter, you can watch sessions from each version side by side. Did variant B users actually engage with the new layout, or did they bounce harder?

After the test: You explain why a variant won. The numbers tell you what happened. The recordings tell you why. This is the part teams skip and then wonder why their next test fails.

For diagnosing rage clicks and confusion before you even start testing, see my walkthrough on using Hotjar click tracking to find rage clicks.

What to Use Instead for Actual A/B Testing

Here's a practical breakdown of options, organized by what kind of builder you're working with.

If your landing page is in Webflow or Framer

Native Webflow Optimize (if it's available on your plan) lets you run split tests inside the Webflow Designer. You build the variant directly in the canvas. No code injection, no separate dashboard. The tradeoff: it's tied to Webflow, so you can't test pages built elsewhere.

Framer's built-in A/B testing works similarly. You duplicate a page, set traffic split, and Framer tracks conversions on whatever event you define. Good for simple tests. Less flexible for multi-step funnels.

If you're choosing between these two, I compared them in Framer vs Webflow for landing pages.

If you want a dedicated A/B testing tool

VWO runs visual editor tests, has session recordings built in, and handles server-side testing for more advanced setups. Pricing is quote-based for most plans, see VWO's pricing page for current details.

Optimizely is the enterprise option. Powerful, expensive, overkill for most indie SaaS pages. Skip unless you have a CRO team.

AB Tasty sits between VWO and Optimizely in terms of complexity. Good for marketing teams that want personalization plus testing.

Convert.com is worth a look if you care about privacy and don't want a heavy script slowing your page down.

If you want free or near-free

Google Optimize is dead. Google killed it in September 2023. If you find old tutorials recommending it, ignore them.

GrowthBook is open-source, self-hosted or cloud. You can run feature flags and A/B tests with it. Steeper setup, much cheaper.

PostHog includes experiments as part of its product analytics suite. If you already use PostHog for events, the testing module is a natural add-on.

Plausible + custom split logic: if you're technical, you can roll your own variant assignment with a few lines of JavaScript and track conversions in any analytics tool. Not pretty, but it works.

The Hotjar + Testing Tool Combo I Actually Recommend

For most SaaS founders and indie hackers, the practical stack is:

1. A landing page builder that loads fast and lets you edit quickly
2. An A/B testing tool that matches your scale (free until you have traffic, paid once you do)
3. Hotjar or a Hotjar alternative for session recordings and heatmaps

The reason you want all three: testing tells you which version won, recordings tell you why, and the builder lets you iterate fast on the next test. Missing any one of them slows you down.

If Hotjar's pricing has pushed you to look around, I covered Hotjar alternatives that are free or much cheaper and what changed in Hotjar pricing for 2026.

Common Mistakes When Combining Hotjar with A/B Tests

A few traps I see repeatedly:

Running a test without enough traffic. If your page gets 200 visitors a week, you'll never hit statistical significance on a normal test. Recordings are still useful at that volume. Splits aren't. I wrote a guide to A/B testing without wasting traffic that covers what to do when sample sizes are tiny.

Testing too many things at once. A button color change and a new headline and a different hero image. You can't tell which caused the lift. Pick one variable.

Ignoring the recordings of losing variants. The losing variant teaches you more than the winning one. Watch those sessions.

Calling tests early. If your tool says "70% probability variant B wins" after two days, that's not a result. That's noise. Wait for the sample size you planned for.

So, Should You Still Use Hotjar?

Yes, but for what it actually does. Use Hotjar (or an alternative) to understand behavior. Use a real A/B testing tool to test. Don't expect one tool to do both unless you've picked a platform like VWO that genuinely offers both features.

If your goal is to ship a higher-converting page this month, here's the order I'd work in: install a recording tool, watch 20 sessions, write down three things that confused users, fix the worst one, then set up an A/B test on the second-worst one. That sequence outperforms any "best practices" checklist because it's grounded in what your visitors actually do.

---

Want to know which fixes will move the needle before you even open Hotjar? PagePulse audits your landing page and tells you which elements are likely costing you conversions, with specific rewrites you can ship today. Paste your URL at pagepulse.page and get your report in under a minute.