Microsoft Clarity vs Hotjar: the free session recording tools compared

by Harry Watts · Last reviewed 21 October 2026

For a small UK site, Microsoft Clarity is the free tool that records every visit. Hotjar's free plan is now Contentsquare Free, which replays one visit in twenty. Whichever you choose, in the UK it shouldn't record anyone until they've accepted your cookies, and it should mask what people type into your forms.

This guide is for a business paying for clicks that don't become enquiries, who wants to watch what those visitors actually do. It compares what each free plan records as of October 2026, lists the free plans that have gone, and covers what "free" still costs you in consent and in the terms.

Why watch recordings at all

Your conversion report tells you a paid visit didn't enquire. It can't tell you why. A session recording (also called session replay) plays back what that visitor did on your site: where they scrolled, what they clicked, and where they gave up. It is how you find the leaks on your own pages, the ones a wasted spend checklist can only point at.

I can't point to a single recording, because the pattern has repeated itself countless times. Recordings have shown me bugs that only appear at one screen size, repeated clicks on an element visitors expect to be a link, and visitors who click the call to action, reach a long contact form and leave immediately. For checking issues like these, I find them invaluable.

What each free plan actually records

Most comparison pages lead with features. For a small site the number that decides it is how many visits the free plan actually records, because a tool that replays a sample can leave you with almost nothing to watch.

Contentsquare's free plan headlines 200,000 sessions a month, but that is for its heatmaps and analytics. Only 5% of those sessions become replays (Hotjar's help centre). A site with 2,000 visits a month gets about 100 replays, and the paid share of a small site's traffic might give you a handful. Clarity says it has no sampling and no paid tier, and records up to 100,000 sessions per project per day (Clarity FAQ).

I've used Hotjar for years. I inherited it at a company I worked for, and it never gave me a reason to switch: recordings, heatmaps and exit surveys, easy to use, and at a price finance never questioned. So I set it up at the companies I joined after that, and for many of the SMEs I work with.

The free plan recorded 35 sessions a day. That's not enough for deep, statistically significant analysis of course, but plenty of small businesses don't receive 35 website visits a day, let alone buy 35 Google Ads clicks. So it was more than enough to spot obvious usability problems and the patterns holding a conversion rate back.

When I added it to this site, it came packaged as Contentsquare. Search Console was showing more visits from Google than I had recordings, and that's when I realised I no longer had full coverage. The free plan now records one visit in twenty. That's what made me look again, and I switched this site to Microsoft Clarity, which I'm very happy with so far. I still use Hotjar and Contentsquare on plenty of the accounts I work on. It just isn't necessarily my go-to any more.

I have no affiliate or referral relationship with Microsoft Clarity, Contentsquare or any other tool on this page.

Free plans that have gone

A lot of what ranks for "free session recording tool" still lists plans you can't sign up for:

• Hotjar Free. Existing accounts keep their 35 sessions a day, but Hotjar's site now sends new sign-ups to Contentsquare.
• Smartlook. Cisco is shutting it down. New sales ended on 31 May 2026, and Smartlook says the platform closes on 30 September 2027.
• LogRocket and Lucky Orange. Neither lists a free plan any more, only a trial.
• OpenReplay. Free only if you host it on your own servers. The cloud version is a seven-day trial.

Free in money, not in obligations

In the UK, a session recording needs the visitor's consent before it starts. The ICO's guidance says you must get consent for "logs or recordings of individual visitors to your website and the actions they took", and that the new statistics exception "does not allow you to monitor or track individual visitors" (the ICO's exceptions). The 2026 change that lets some analytics run without consent covers counting visits, not recording them. How the banner and Google's consent signals fit together is in Consent Mode v2 for UK advertisers.

Clarity adds three things to check before you install it:

• Microsoft is a controller. Clarity's terms make you and Microsoft "independent Controllers", and let Microsoft use the personal data it collects to improve its products and "create user profiles for purposes that include advertising" (Clarity terms of use). Your privacy notice has to say so, and because Microsoft isn't your processor, the statistics exception could never apply to it anyway.
• Some sites are out of bounds. The same terms rule out content such as "health care, financial services or government-related information". A clinic, a dentist or a financial adviser should read them before installing it.
• Its own consent mode isn't enough. Since 31 October 2025 Clarity has needed a consent signal for visitors from the UK, the EEA and Switzerland (Clarity consent API). Without one it still loads and still collects page views and basic interactions, just without cookies. The safe setup is not to load the tag at all until a visitor has accepted analytics cookies.

When a client tells me they don't want a tool filming their customers, I start with consent. Set up properly, a recording tool records nobody until they've agreed to it on the cookie banner, and anyone who declines isn't recorded at all. What it does record isn't a film of the person. It's their clicks, scrolls and taps on the page, and anything they type into a form is masked before it leaves their browser. Then I explain what they get in return. Even at a small scale, a handful of recordings of people who clicked an ad and didn't get in touch will often show exactly why, which no report can.

Masking differs by tool. Clarity masks input boxes and dropdowns in every mode, and its default also masks numbers and email addresses (Clarity masking). Contentsquare masks all text and numbers until you change it, PostHog masks inputs but not page text, and Fullstory's free plan leaves out configurable form privacy.

Your first five recordings

Don't start by watching recordings at random. Start with the visits you paid for that didn't enquire.

• Filter to paid visits. In Clarity, the Channel filter has a Paid search option, and the Source, Medium and Campaign filters read your UTM tags (Clarity filters). Add the entry page you send your ads to, and you are watching the visits you paid for.
• Watch where they stop. The page they leave from, and how far down it they got, is the first clue. A visitor who reaches your form and leaves without starting it is telling you something different from one who leaves the landing page in five seconds.
• Look for clicks that do nothing. Clarity flags rage clicks (repeated clicks in one small area within a short time) and dead clicks (a click that gets no response), and you can filter to recordings that have them. Both can mean a visitor expected a link or a button that isn't there.
• Check the screen size. A layout bug that only shows on one phone width won't appear when you check the site on yours.
• Check the form itself. Count the fields a visitor sees after they click your call to action. If the recordings show people arriving at the form and leaving, the form is the first thing to shorten.

If the recordings show visitors enquiring and the account still reports nothing, the problem is the measurement, not the page: start with the conversion tracking mistakes, and if your enquiries come by phone, count the calls.

Quick answers

If you'd rather have it set up

Consent is one of the checks in conversion tracking audit and repair, where I look at every route an enquiry takes into the account and every tag in the container. If you'd like to talk it through first, a free intro call is thirty minutes.