If you’ve set up GA4 (we’ve written a full guide to what’s new there), you already know how many people visited a page, how long they stayed, and whether they converted. What GA4 can’t show you is why they didn’t. It’s a numbers story — it tells you a page underperformed, not what actually happened when a real person landed on it and quietly gave up.
That’s the gap user behavior analytics (UBA) tools fill: heatmaps and session recordings that let you watch, in effect, over a visitor’s shoulder. Here’s what they actually show you, which tools are worth using in 2026, and the one honest limitation every one of them shares.
What these tools actually show you
Three views, each answering a different question:
- Click and heatmaps — where on the page people actually click, tap, and hover, visualised as a heat-coloured overlay. Instantly reveals when people are clicking something that isn’t a button, or ignoring the button that matters most.
- Scroll maps — how far down the page people actually get before leaving. If your key offer sits below the point where 80% of visitors have already scrolled away, that’s now visible instead of invisible.
- Session recordings — literal replays of real visitor sessions: their mouse movement, their hesitation, the moment they get stuck. Most tools now flag “rage clicks” (rapid, frustrated repeated clicking) and “dead clicks” (clicking something that isn’t actually interactive) automatically, so you don’t have to watch hours of footage to find the friction.
The tools worth using
Microsoft Clarity — the default starting point, and it’s genuinely free. Unlimited session recordings and heatmaps, with soft limits (100,000 pageviews and sessions per project per day) that almost no SME site will ever hit. It plugs natively into GA4, so behavioural data sits alongside the numbers you already track, and Microsoft’s Copilot integration now generates plain-language summaries of what happened in a session or heatmap automatically — genuinely useful when you don’t have time to watch recordings manually. It also covers native mobile apps, not just websites, which most competitors don’t. The honest catch: data retention is a relatively short 30 days, and as with any free tool from a major platform, it’s worth being aware of the data-ownership terms rather than assuming they match a paid, privacy-first product.
- Best for: almost every SME as a starting point — there’s little reason not to install it.
- Cost: free.
Hotjar — for teams that want direct visitor feedback, not just observation. Beyond heatmaps and recordings, Hotjar bundles surveys, feedback widgets, and user interview tools — a fuller research suite for actually asking visitors what they think, not just watching what they do. Paid plans offer far longer data retention (up to 365 days) than Clarity’s 30. One 2026-specific thing worth knowing: Hotjar was acquired by Contentsquare, and its plans and billing are gradually transitioning onto Contentsquare’s structure — worth checking current terms before committing, since this kind of transition can shift pricing and features.
- Best for: teams who want qualitative feedback (surveys, interviews) layered on top of behavioural data.
- Cost: limited free tier; paid plans from roughly US$32/month.
Lucky Orange — for online stores specifically. Bundles heatmaps and session recording with live chat in one package, and has particularly strong native Shopify integration — tying behaviour data directly to cart abandonment and purchase patterns.
- Best for: Singapore SME e-commerce stores, especially on Shopify.
- Cost: from roughly US$32/month.
Worth a brief mention: Crazy Egg for teams running a tight conversion-optimisation loop, PostHog for technically capable teams wanting a self-hosted, open-source option with full data control, and enterprise platforms like Contentsquare or Glassbox — genuinely powerful, and genuinely overkill for a typical SME’s traffic and budget.
The honest limitation, every single tool shares
These tools are excellent at showing you what happened — where people clicked, where they hesitated, where they gave up. None of them tell you why. A heatmap can show a form being abandoned at the third field; it can’t tell you whether that’s confusing wording, a fear of committing, or simply the wrong field to ask for at all. That interpretation is still a human job — which is exactly why pairing behavioural data with something qualitative (a quick survey, actual customer conversations) closes a gap no heatmap alone will.
What this actually looks like in practice
The pattern shows up often enough to be worth describing generally: a business installs a free heatmap tool, and within weeks discovers something that had been invisible for months — say, a services page quietly losing a large share of mobile visitors to a scroll or layout issue nobody noticed on desktop. Fix the specific thing the recording revealed, and it’s common to see a meaningful, measurable lift in form submissions or enquiries shortly after — the kind of return that costs nothing but the ten minutes it takes to install a free script and actually watch a few sessions.
The 30-second decision
- Want free, unlimited, and you already use GA4 → Microsoft Clarity
- Want surveys and direct visitor feedback alongside behaviour data → Hotjar
- Running an online store, especially Shopify → Lucky Orange
- Technical team, want full data control → PostHog
Where Oasis Web Asia comes in
We set these tools up as part of a proper analytics stack alongside GA4 — and more importantly, we help interpret what they actually show. A rage-click on a button, a form abandoned at field three, a scroll-map showing your offer nobody reaches — these are meaningless without someone turning the observation into an actual fix. It’s the same evidence-first web development Singapore SMEs trust us with, aimed at the moments your own visitors are quietly telling you something’s wrong.
If you’ve never actually watched a real visitor use your website, that’s exactly the conversation we like to have.
Start a conversation → — get a free consultation with our Singapore-based team.