What Hotjar shows, and what it doesn’t
Hotjar shows how visitors behave on a page: where they click, how far they scroll, what they do in a real session and, if you ask, why they did it. It doesn’t tell you how many people came, from where, or which campaigns paid off; that’s your analytics tool’s job.
| Hotjar tool | The question it answers | What it can’t tell you |
|---|---|---|
| Heatmaps (click, move and scroll) | Where do people click, and how far down the page do they get? | Why they clicked, or whether they found what they wanted |
| Recordings | What does a real visit look like, step by step? | How common that behaviour is, unless you count it |
| Surveys and feedback | What do visitors say stopped them, or what they came for? | The view of people who don’t answer |
Hotjar is part of Contentsquare, and its plans and features change, so check hotjar.com before you choose one. Hotjar also offers features such as funnels and simple trend reports on some plans. Check what your plan includes before you plan an analysis around them.
The rule that makes Hotjar useful: start with one page and one question. “Why do mobile visitors leave the pricing page?” gives you something to look for. “Let’s see what users do” gives you a hundred recordings and no conclusion.
How to set up Hotjar
Setting up Hotjar is quick; the decisions about privacy and scope take longer than the install. Do it in this order:
- Create your account and add your site. Hotjar gives you a tracking code for that site.
- Decide where it runs. You don’t need Hotjar on every page. Load it on the pages you’re studying, and remove it when you’re done.
- Set privacy before you record. Review Hotjar’s privacy settings: what it hides from recordings, what it does with text typed into forms, and how to mark extra elements on your page to hide. Check these against your own privacy notice.
- Handle consent. For visitors in the UK and EU, analytics and recording tools are usually loaded only after the visitor agrees through a cookie banner. Ask your privacy adviser what your site needs, then load Hotjar through your consent tool or Google Tag Manager.
- Install the code in the site’s head or through Tag Manager, then use Hotjar’s install check to confirm data is arriving.
- Send events for the actions you care about. Hotjar can receive custom events from your site, such as “form submitted” or “pricing toggle used”, so you can filter recordings by what visitors did. See Hotjar’s developer documentation for the current events API.
- Set up the heatmaps and a survey for the page you’re studying, and wait for enough visits before you read them.
The first five things to look at
Start with these five checks. They find most of the obvious problems on a typical marketing or product page.
- Scroll depth against your main call to action. Open the scroll heatmap. If most visitors never reach the button or form, move it up, or give them a reason to keep reading.
- Clicks on things that aren’t links. Click heatmaps often show people tapping images, icons or bold text that do nothing. Either make those elements work or make it clear they aren’t clickable.
- Frustration in recordings. Filter recordings for signs of frustration, such as repeated rapid clicks or a quick return to the previous page. These often point to a broken element, a slow response or a confusing step.
- Where forms lose people. Watch recordings of visitors who start a form but don’t send it. Look for the field where they pause, go back or leave: a required field they don’t understand, an error message that’s hard to see, or a question they’re not ready to answer.
- Mobile and desktop separately. Read heatmaps by device. A layout that works on a laptop can push the key content far down on a phone, and the problem disappears if you look at all devices together.
Add one short on-page survey to the page you’re studying, such as “Is anything stopping you from signing up today?” One question with an open answer often explains what the heatmaps can only show.
How to watch recordings without losing a day
Watch recordings with a question and a filter, not from the top of the list. A few habits make it fast:
- Filter first. By page, device, country, traffic source or a custom event. “Mobile visitors on the pricing page who didn’t click the plan button” is a set of recordings worth watching.
- Skip the idle time. Turn on the option to skip inactivity, and watch at a higher speed.
- Keep a tally. In a simple sheet, write the session, what happened, and a tag such as “missed CTA”, “form error” or “confused by price toggle”. Patterns show up in the tally, not in your memory.
- Stop when you stop learning. When new sessions only repeat what you’ve already tagged, you have enough to act.
- Share clips, not conclusions. A 20-second clip of a visitor struggling persuades a team faster than a slide.
Turning observations into a fix list
Hotjar insights only matter when they turn into changes. Turn each pattern into a row in a fix list and rank it before anyone starts work. The rows below are an illustration of the format, not findings from a real site.
| Observation | Evidence | Likely cause | Fix | Effort | How we’ll know |
|---|---|---|---|---|---|
| Most mobile visitors don’t reach the form | Scroll heatmap, mobile | Hero and image take up the first screens | Shorter hero, form link near the top | Small | Form starts on mobile |
| Visitors tap the plan cards, not the buttons | Click heatmap | Cards look clickable | Make the whole card a link | Small | Plan clicks |
| People stop at the “company size” field | 14 of 20 form recordings | Unclear options | Simpler options, or make it optional | Small | Form completion rate |
Rank by impact and effort: fix the small, high-impact items first. Then measure each fix in your analytics, and run the heatmap again on the changed page. This is the same loop behind a conversion-focused website: observe, change one thing, measure.
Pairing Hotjar with your analytics: which question each one answers
Use your analytics tool, such as Google Analytics 4, to find where the problem is, and Hotjar to see why it happens. They answer different questions, so you need both.
- Analytics answers “how many” and “where”: how many visitors reached the page, where they came from, which pages lead to sign-ups, where in a funnel they drop out.
- Hotjar answers “what did they do” and “why”: what visitors did on the page before they left, what they clicked and what they told you.
A good working pattern: analytics shows that mobile visitors from paid ads leave the landing page faster than others. Hotjar recordings of those visitors show the offer is hidden below a large image. You fix the layout, and analytics tells you whether the change worked. For the analytics side, see website analytics for data-driven decisions.
Does Hotjar slow down your site or affect SEO?
Hotjar doesn’t affect your search rankings directly, and its script loads asynchronously, so the effect on page speed is usually small. Any SEO effect comes through speed, so measure it rather than assume it.
- Measure before and after. Check your Core Web Vitals and a lab test, such as Lighthouse, with and without Hotjar on the page.
- Run it only where you need it. Loading Hotjar on a few pages for a few weeks costs far less than loading it everywhere, forever.
- Load it after consent where you need consent. That also means it never runs for visitors who decline.
- Remove it when you’re done. An old tracking script nobody reads is pure cost.
Hotjar’s recordings and heatmaps aren’t seen by search engines, and they don’t change your content. Where Hotjar helps SEO is indirect: it shows you why visitors leave a page that ranks, so you can fix the page they land on.
Privacy, masking and consent
Recording real visits is useful because it’s real, which is also why it needs care. Before you turn recordings on:
- Know what’s hidden by default. Check what Hotjar hides from recordings and what it does with text typed into form fields, in its current help documentation.
- Hide anything else that’s personal. Account pages, order details, names and addresses shown on screen should be hidden from recordings, or the page excluded altogether.
- Exclude pages that don’t need studying, such as account settings or checkout confirmation pages.
- Say it in your privacy notice. Tell visitors you use behaviour analytics and what it collects.
- Respect consent choices. Where your site uses a consent banner, Hotjar should load only for visitors who accept.
If your site serves visitors in India, the UK, the EU or the US, privacy rules differ by place. Treat this section as a checklist to discuss with your privacy adviser, not legal advice.
The limits of Hotjar
Hotjar is a tool for questions, not a source of answers on its own. Keep its limits in mind:
- Small samples mislead. A heatmap from forty visits can look convincing and mean nothing.
- Behaviour isn’t intent. A recording shows what someone did, not what they meant. Surveys and user interviews fill that gap.
- Data piles up. Without a question, you’ll collect thousands of recordings and watch none of them.
- It can’t test a fix. Hotjar shows a problem; an A/B test or a before-and-after measurement shows whether your fix solved it.
How we’d use Hotjar in a UX audit
In a UX audit, we’d use behaviour recordings to confirm what the heuristics review suggests, then turn the evidence into a ranked fix list. The method is the one above: pick the pages that matter most to the business, agree one question per page, collect heatmaps and recordings for long enough to see patterns, add a short survey, tally what we see, and rank fixes by impact and effort.
The useful part comes after the audit. A fix list is only worth something if someone changes the page and measures the result, which is where our UX audit services meet the work of fixing the page after the click, so that paid traffic lands on a page that works.
A short Hotjar checklist
- One page, one question.
- Privacy settings and consent handled before recording.
- Hotjar loaded only on the pages you’re studying.
- Events sent for the actions you care about.
- Heatmaps read by device.
- Recordings filtered and tallied, not browsed.
- One short survey on the page.
- A ranked fix list, with a metric for each fix.
- Before-and-after measurement in your analytics.
- Hotjar removed or paused when the question is answered.
