UX and UI analysis comes down to five questions: where do people click, where do they get stuck, where do they leave, why they do it, and does a change actually make things better? This is a shortlist of the tools that answer them — each one checked against the vendor’s own site in September 2026, which is also how two of the tools this article used to recommend lost their place on it.
What UX and UI analysis tools actually do
UX and UI analysis breaks into five jobs, and most tools do one or two of them well rather than all five. Heatmaps show where attention, clicks and scrolling land on a page. Session replay shows a single visit as it happened, including the dead clicks and the rage clicks. Funnel and conversion analysis shows where people leave a path you defined. A/B testing proves whether a change is an improvement or just a preference. And usability testing with real participants answers the question no clickstream can: why.
So pick by job, not by feature list. One more thing belongs in the decision rather than after it: everything below records real visitor behaviour, so consent, masking of form fields and a retention limit are part of choosing the tool, not part of cleaning up later.
Heatmaps and session replay: what visitors do on the page
This is where to start if you have never measured UX before, because it needs no hypothesis. You install a tag, watch a few dozen sessions and the obvious problems show themselves — a form field everybody retypes, a button nobody sees, a mobile layout that hides the price below three screens of text.
Microsoft Clarity — the quickest way to start
Microsoft Clarity gives you heatmaps, session recordings and automated insight summaries, and Microsoft’s own page describes it as free with no traffic limits (checked September 2026). That combination makes it the default first tool: there is no budget conversation to have before you can look at your own site. What you do not get is a full analytics suite — Clarity answers “where do people struggle”, not “which channel pays for itself”.
Contentsquare, formerly Hotjar — heatmaps, replay and surveys together
If you know this category at all, you know Hotjar. It no longer exists on its own: hotjar.com now redirects to Contentsquare, which states that the two platforms have merged into a single platform and that the familiar parts — heatmaps, recordings, surveys — are still there, alongside funnels and error monitoring. There is a free plan (checked September 2026). Treat any older article that still sends you to a Hotjar plan as out of date, including the earlier version of this one.
Inspectlet — session replay with form analytics
Inspectlet records mouse movement, clicks and keystrokes and plays them back, and adds dynamic heatmaps, form analytics, feedback surveys, error logging and its own A/B testing. Form analytics is the reason to look at it specifically: if a checkout or a contact form is where your conversions die, field-by-field drop-off is a faster diagnosis than watching sessions.
“You can identify weak points and stumbling blocks quickly and improve your website accordingly.”
Matomo — heatmaps you host yourself
Matomo records clicks, mouse movement and scrolling, and can be run on your own servers instead of someone else’s cloud. It is the honest answer when data ownership, a specific hosting location or the works council are the deciding factors rather than the feature list, and it brings its own privacy controls, including masking sensitive elements out of heatmaps.
Metrics and funnels: where visitors drop out
Heatmaps tell you about a page; funnels tell you about a path. This is also where UX work meets reporting — the numbers only settle an argument if everyone is looking at the same definition of a conversion, which is a dashboard and KPI question rather than a tooling one.
Clicky — real-time analytics down to the single visit
Clicky shows activity on your site as it happens, with visitor-level logs, heatmaps at both page and visitor level, segmentation, bot and referrer-spam filtering, uptime alerts and privacy-friendly tracking. It is a small, fast tool for people who want to watch individual visits rather than build a reporting layer. The Twitter statistics the previous version of this article praised are no longer advertised; what Clicky puts in that slot today is segmentation — visits from AI chatbots are categorised separately, with direct AI bot tracking announced but not yet shipped.
PostHog — funnels, retention and replay in one place
PostHog autocaptures pageviews, clicks and form submissions and turns them into trends, funnels, retention curves and path analysis, with session replay, surveys, feature flags and experiments in the same product. It is the right kind of tool when your website is really an application: sign-up, onboarding and repeat use are event problems, not page problems.
A/B testing: proving a change is an improvement
Testing is what turns an opinion about a design into a decision. It also has the highest failure rate of anything on this list, and almost always for the same reasons: the metric was agreed after the test, two changes shipped at once, or the test was stopped the moment it looked good. Decide the metric and the runtime before you build the variant.
AB Tasty — experimentation and personalisation
AB Tasty covers web experimentation, feature experimentation and staged rollouts, plus personalisation and e-merchandising, with a visual editor for people who do not want to write the variant themselves. Its behaviour tools — session recordings, heatmaps and scrollmaps, funnel analysis — sit in a separate module called Insights rather than in the testing product, so check which of the two a quote covers.
VWO — testing plus behaviour analytics in one licence
VWO runs A/B, split and multivariate tests on the web, in mobile apps and server-side, and it also carries the behaviour side: session recordings, heatmaps, funnel analysis and form analysis, along with feature rollouts and voice-of-customer surveys. That overlap is the argument for it — one tool where the observation and the experiment sit next to each other.
Unbounce — variants of a landing page without a developer
Unbounce is a landing page builder rather than an analysis tool: you assemble a page, create variants and A/B test them, and its Smart Traffic feature routes visitors to the variant that fits them best. Useful for campaign pages that need to exist by Thursday; the wrong tool for improving the site you already have.
Usability testing: the part clickstream data cannot answer
Analytics tells you how many visitors abandon a step. It never tells you that they abandoned it because the word above the field means something different in their industry. A few moderated sessions usually explain more than a month of heatmaps.
Maze — tests with real participants, moderated or not
Maze runs prototype tests, live website tests, mobile tests, moderated interviews, surveys, card sorting and tree testing, and it can recruit participants from its own panel if you have no one to ask. The prototype half matters most: it lets you find the problem before the page is built, which is the cheapest moment to fix anything.
What became of the tools of 2022
A tool list is only as good as its last check. Four names that belong in the history of this article rather than in its recommendations:
- Keen — keen.io is now a managed event streaming platform on Apache Kafka for developers. It is a perfectly good product and it is not a UX dashboard.
- Smartlook — has joined Cisco: end of sale on 31 May 2026, no further renewals after 31 August 2026, platform decommissioned on 30 September 2027. Its capabilities continue inside Splunk Observability Cloud, which is a different purchase for a different buyer.
- Hotjar as a standalone product — merged into Contentsquare, as above. The name survives as a page on someone else’s site.
- Google Optimize — Google’s own free testing tool has not been available since 30 September 2023. Worth knowing even though this article never recommended it: it is the reason a lot of testing setups moved to AB Tasty, VWO or PostHog, and the reason free A/B testing is now a feature of other products rather than a product.
One more thing is gone from this version: the prices. The earlier text quoted monthly figures for four of its six tools, and four years later every one of those figures had to be re-checked — one of them described the plans of a product that can no longer be bought at all. Pricing pages are one click away from every link above, and they are right; an article is not.
How to choose without buying four tools
- Write the question down first. “Why do people abandon the quote request?” picks a tool; “we should look at UX” does not.
- Install one behaviour tool — heatmaps plus session replay — before anything else, and watch twenty real sessions before you draw a conclusion.
- Add a testing tool when you have a hypothesis worth the traffic. Below a few hundred conversions a month, qualitative testing beats A/B testing.
- Sort out consent, form-field masking and a retention period at install time, not after the first recording of somebody typing their address.
- Do not forget speed: page experience signals are part of how the page feels, and no heatmap will show you a slow one.
- Agree where the numbers land. A conversion rate that cannot be reconciled with what the CRM counts as a lead will lose every argument it is brought into — which is a data readiness question, and worth settling before the first test rather than after the third.
Conclusion: match the tool to the question
The shortlist is short on purpose. Start with Clarity or Contentsquare to see what happens on your pages, add Inspectlet or Matomo if form analytics or self-hosting is the deciding factor, use Clicky or PostHog for the path rather than the page, bring in AB Tasty or VWO once there is something to prove, and put Maze in front of the build when the design is still a prototype.
And put a date on the list. Of the six tools this article recommended in 2022, two are no longer in the business it recommended them for, and the best-known name in the category, which the list did not even include, has since lost both its own product and its own site. Whatever you choose today, check the vendor’s own page before you quote it to anyone.
