How to Install Hotjar and Get Usable Heatmap Data on Day One

October 31, 2026

Hotjar combines session recordings, heatmaps and surveys under one script. The free Basic plan covers 35 sessions per day, and paid plans start around $32/month — enough for a small site to get real qualitative data without a procurement conversation.

What is Hotjar actually good for?

Click and scroll heatmaps on a specific page, plus recordings of the sessions where something went wrong. Its clearest advantage over free Microsoft Clarity is survey and feedback tooling in the same product: you can show a poll to the 20% of visitors who never scrolled past the fold.

If you only want recordings and heatmaps, Clarity is free with no daily cap and Hotjar’s case is thinner. Buy the surveys.

How do you install the tracking script?

  1. Create the site in Hotjar and copy the tracking code from Sites → Tracking code.
  2. Paste it just before the closing head tag, on every page you want measured.
  3. Verify installation in Hotjar’s own check — it crawls your URL and reports whether the script fires.
  4. Set up user identification if you run an app with logins.
  5. Publish the first heatmap on a page with meaningful traffic.

Step 5 has a minimum bar: a heatmap on a page with 40 visitors a month tells you nothing. Hotjar needs roughly a few hundred pageviews before click distribution stabilises.

Does the 35-session cap ruin heatmaps?

Not for heatmaps. A heatmap aggregates clicks over all recorded sessions, and 35 sessions a day adds up to about a thousand a month — enough for one or two high-traffic pages.

It does limit recordings. You will capture a sliver of traffic, and finding a specific user’s broken session becomes luck rather than lookup. That is the practical wall on Basic, and it is why most teams upgrade once recording becomes part of a support workflow. Compare FullStory if the workflow depends on looking up a named user’s session.

How do you set up a heatmap worth looking at?

Pick a page with one clear question attached. “Do people click the pricing link in the hero?” is a heatmap question. “Are people engaging?” is not.

Configuration that gives a clean read:

  • Desktop and mobile separately. Aggregating them produces a smear.
  • Click heatmap first, scroll second. Clicks point at intent, scroll points at layout.
  • Exclude internal IPs and your own traffic, or your testing inflates the top-left corner.
  • Segment by source if you run paid traffic. Ad visitors behave differently from organic.

Once you have the heatmap, open three recordings from the same viewport. The heatmap says where, the recording says why, and the pairing is what makes a finding actionable.

What should you do with the data afterwards?

Write down one sentence per finding and one proposed change. “People click the non-link feature card 40% more than the pricing button — make the feature card clickable.” That is a ticket someone can act on.

Changes without a follow-up measurement are why teams conclude heatmaps do not work. Re-run the same heatmap four weeks later, same segments.

For funnel-level numbers alongside the qualitative layer, pair with Amplitude or PostHog — neither replaces a heatmap, and a heatmap does not replace a funnel.

Where does Hotjar fall short?

Sampling surprises. On paid plans, recordings are sampled by default, and people compare a sampled week with a full previous week and conclude traffic dropped.

Recorder performance on heavy pages is another one — the script adds weight, and mobile Lighthouse scores take a hit. Load it after interaction if you care about Core Web Vitals.

Key Takeaways

  • Free Basic caps at 35 sessions per day; paid plans start near $32/month.
  • Heatmaps need a few hundred pageviews to be meaningful; below that, wait.
  • Split desktop and mobile, exclude internal traffic, and start with click heatmaps.
  • Pair every heatmap with three recordings from one viewport — where plus why.
  • For surveys on specific visitor segments, Hotjar is the differentiator over free Clarity.
Tags: