Microsoft Ads: should Clarity heatmaps drive landing page changes or only conversion data?

A lot of advice treats Clarity heatmaps as a redesign brief, and I'm not convinced. Say a B2B software page gets 200 clicks a month from Microsoft Ads. The heatmap shows nobody scrolls past the pricing block, and the rage clicks land on a logo carousel. The obvious move is to kill the carousel and push pricing up. But 200 sessions is a thin sample, and a heatmap shows where people clicked, not why they left. Conversion rate over the same period is the number that actually pays the bills. I'd only act on Clarity when it points at something the conversion data backs up, like a form field where people drop off. Otherwise you're redesigning a page because of a pretty picture. For those running Clarity on low-traffic pages, how many sessions do you need before you trust a heatmap?

Agreed on the core: conversion data should decide, and the heatmap should only generate hypotheses. But I'm not sure the pricing read holds. "Nobody scrolls past the pricing block" might just mean pricing sits high enough that most people see it on the first screen and rarely get to what's below. That's not necessarily a failure, so moving pricing up could fix something that isn't broken.

The rage clicks are the more useful signal. People often rage-click carousels because the logos look like links or the motion invites a click. Before removing the section, check whether those logos link anywhere. Making them static or linking them to case studies is a cheaper test than deleting a block, and conversion rate will tell you which one helped.

On sample size, there's no clean number. Scroll depth can show a direction with a couple hundred sessions, but specific click patterns need repeated behavior over a few weeks before I'd touch the layout. Are those rage clicks mostly on mobile or desktop?

The carousel test makes sense, and you can sanity-check the pricing idea with data you already have. Compare conversion rate for sessions that reached the pricing block against sessions that didn't. If the people who scroll less convert about the same, position probably isn't the bottleneck, and moving pricing up would just shuffle the page. If the ones who reach pricing convert much better, getting more people there is worth testing.

Does pricing sit above the fold on most of those 200 sessions, or does it depend on device?

Agreed on the split, with one catch: people who scroll down to pricing already have more intent, so a higher conversion rate there mostly reflects who they are, not what the page does. The split shows where the bottleneck might be, but it can't tell you whether moving pricing would help.

For the cutoff itself, I wouldn't use Clarity's scroll depth. It's a page-level percentage, and recording coverage thins out when consent is denied, so the two groups may not be comparable. A better route is a GTM Element Visibility trigger on the pricing block that fires a GA4 event once it's about 50% on screen. Then compare conversion for sessions with that event against sessions without it.

Run the same check on the Microsoft UET side. If the conversion counts for those 200 sessions don't line up between UET and GA4, the split is built on bad numbers, and that needs fixing before you read anything into it.

Agreed on the visibility trigger, and checking UET makes sense. One catch is that Microsoft's UET conversions are usually modeled or delayed when consent is denied, so a 200-session sample can look like a clean split in GA4 and a mess in UET. Pull the raw counts for the same date range in both before you trust either number.