Your analytics can tell you a thousand people visited last month and fourteen enquired. It cannot tell you that two hundred of them tapped a button that does nothing on a phone, waited half a second, and left. That gap between the number and the reason behind it is where most owner-operated businesses quietly lose the buyers they already paid to attract. A session recording closes it. It replays one real visitor's screen: the cursor, the scrolling, the rage-tapping, the moment they gave up. And it routinely surfaces conversion bugs that no dashboard, however detailed, can show you.

Direct answer: A session recording is a screen replay of one real visitor using your website. Combined with heatmaps, it shows why people fail to convert: the broken button, the form nobody finishes, the page nobody scrolls. Analytics only show that they left. For a business too small to run A/B tests, watching a handful of real sessions is the fastest, cheapest way to find what's costing you enquiries.

Your dashboard counts. It never watches.

Open your analytics and you will find out how many people came, how long they stayed, and how many converted. Useful numbers. They describe the symptom — traffic in, almost nothing out — with precision and zero explanation.

This is the trap most founders fall into. They have been sold the idea that more data is the answer, so they add another dashboard, another tracking tool, another weekly report full of green arrows. None of it answers the only question that matters: why did the person who needed exactly what you sell read half the page and leave? A number can tell you a conversion rate dropped. It cannot tell you a thumb missed a button.

The honest version of your problem is simpler and more uncomfortable. You don't have a data problem. You have a watching problem.

"We don't have enough traffic to optimise" is the wrong problem

Here is the line I hear most from founders sitting on an underperforming site: we don't have enough traffic to know what's wrong yet. It feels responsible. It is also backwards.

The belief comes from A/B testing — split two versions of a page, send traffic to both, wait for a statistically significant winner. That genuinely does need volume you probably don't have, and we've made that case before. But A/B testing is the wrong tool for finding what's broken. It is a tool for choosing between two things you already suspect are fine.

Finding what's broken needs almost no traffic at all. Back in 2000, Jakob Nielsen published the research that still anchors usability work today: testing with just five users uncovers roughly 85% of a site's most evident usability problems, and ten users gets you to about 95%. The discovery curve is steep early and flattens fast. You do not need a thousand sessions to learn that your contact form breaks on Safari. You need to watch the fifth person hit the wall the first four hit.

So the small business that thinks it's too small to optimise has it inverted. You have too little traffic to test. You have more than enough to watch.

What a recording shows that a number never will

Numbers aggregate. Recordings are specific, and the specifics are where the money leaks. A few patterns show up again and again once you start watching.

Rage clicks. A visitor clicks the same spot four or five times in a second because nothing is happening. The tools flag these for you automatically. Almost always it is an element that looks clickable but isn't: a styled box that should be a link, a button wired to nothing on mobile, a menu that won't open. Your buyer is trying to give you money and the door won't open.

Dead clicks. Quieter, just as expensive. Someone clicks an image, a phone number, a "free guide" graphic, and gets no response, so they decide the page is broken and stop trusting it. You will never see this in a conversion report. It is obvious in a thirty-second replay.

Scroll depth. A heatmap colours your page by how far people actually get. Founders are regularly caught out to learn that the offer they agonised over, three screens down, is seen by maybe one visitor in five. Everyone else made their decision above the fold, and the rest of the page may as well not exist.

The phone. Most of your visitors are on a phone. Most owners only ever check their own site on a laptop, on fast wifi, logged in. Watch ten mobile recordings and you will find the gap between the site you think you have and the site your buyers actually meet.

None of these is a redesign. Each is a small, specific fix — and you can only fix what you have seen.

A session recording is a bug report your buyer files without knowing it

Here is the reframe that changes how you treat the whole exercise. A conversion problem is not a marketing mystery. It is a bug.

When an engineer gets a bug report, they don't stare at an aggregate error-rate chart and guess. They reproduce the failure and watch it happen, step by step, until the cause is obvious. That is exactly what a session recording is: your buyer reproducing the bug for you, in full, without being asked, then leaving before you read it.

This matters because the alternative to watching isn't rigour. It's vibes. Redesigning a page on a hunch is vibes. Acting on a dashboard that shows what changed but never why is vibes with a chart attached. At SharpHaw we say decisions aren't vibes-based — every change is measured against a number. Watching real sessions is how you earn that number honestly: you see the problem, you fix the one thing, you measure whether enquiries actually move. That is tracked from click to client, not click to dashboard.

Your dashboard counts who left. A recording shows you the door they couldn't open.

The free tool, the consent, and the honest caveat

You don't need budget for this. Microsoft Clarity is genuinely free: unlimited session recordings and heatmaps, no caps, no credit card. It flags rage clicks and dead clicks for you automatically. Hotjar is the other common choice, EU-based and privacy-focused. Either one installs in minutes.

Two things to get right, because this is where doing it properly separates you from doing it carelessly.

First, consent. You are recording how real people use your site, so in Europe this counts as personal data, and it needs consent before it runs. Since late 2025, Clarity won't even operate in the EU, UK or Switzerland without cookie consent wired up. If you read our piece on the cost of the cookie banner, you've already handled that part correctly. Good tools mask sensitive fields like passwords and card numbers by default. Leave that masking on.

Second, the caveat the tool vendors skip. The five-user rule finds the evident problems, not every problem, and watching is qualitative: it tells you what to fix, not how much the fix will pay off. That is fine. You are not trying to prove a 4.2% lift to a board. You are trying to find the button that doesn't work before another month of buyers walk into it.

Watch ten visitors this week

You can do the whole thing in an afternoon. No project, no agency, no waiting for traffic to build.

1. Install one tool (Clarity or Hotjar) on your highest-intent page first: the one your ads or your best content sends people to.

1. Wire up consent properly before it records, and confirm sensitive fields are masked.

1. Wait for ten to twenty real sessions on that page. At normal small-business traffic that is days, not months.

1. Watch them end to end, sorted by rage clicks first, then the longest sessions that didn't convert. Write down every moment you wince.

1. Fix the single most common failure, ship it this week, then watch the next ten to see if the wincing stops — and whether your enquiries move with it.

Do that once and you will never trust a dashboard alone again. The site stops being a black box and becomes a thing you can watch people use, and improve, every week, on the evidence in front of you.

That's the loop. Plan. Build. Iterate. It compounds.

If you'd rather have someone watch with the trained eye of an engineer who has fixed hundreds of these, that's what we do. Book a 30-min call — bring your worst-performing page, and leave with a fix-it list. You can also see how the Analytics module inside SharpOS puts behaviour, not just numbers, in front of you, and what a conversion-first website looks like when it is built to be watched.