Session analytics
Microsoft Clarity for ecommerce stores
Clarity records what people did on the storefront and shows it back as session replays, heatmaps and friction signals such as repeated clicks on something that does not respond. Analytics tells you the mobile product page converts less than the desktop one; replays tell you the size guide opens behind the sticky bar. We use it to explain numbers we already trust, and to turn what we see into a ranked list of changes worth testing.
Microsoft Clarity
Session analytics · by Microsoft. A text wordmark, not a logo: we are not affiliated with the vendor and make no claim to its trademarks.
What we do with it
The work inside Microsoft Clarity
Recordings watched against a question
Random session viewing is entertainment. We start from a number that needs explaining, filter to the sessions that produced it, such as mobile visitors who reached the cart and left without starting checkout, and watch until the same behaviour appears repeatedly. The output is a pattern, not an anecdote.
Heatmaps on the templates that sell
Product, collection and cart templates get scroll and click maps read together. Scroll depth shows whether the buying information sits above where most visitors stop. Click maps show taps on elements that do nothing, which is usually an image people expect to zoom or a specification row they expect to expand.
Friction signals as a triage list
Repeated clicking, clicks on dead areas, sessions that bounce straight back to the previous page and script errors are each grouped by template. Sorted by the traffic and revenue behind the template, they become a work queue instead of a curiosity, and the top item is often a small theme fix.
Segments that match the business question
Device, country, traffic source and landing page all filter the same recordings. Linking Clarity to the analytics property lets a report send you to the sessions behind it, which shortens the trip from noticing that paid traffic converts poorly on one collection to seeing what those visitors actually did.
Observations turned into tests
Every finding leaves as a hypothesis with a metric attached: move this element, expect this rate to change, measure it this way. Where traffic supports a controlled test we run one. Where it does not, we ship the change and watch the metric with a clear before and after window.
When it fits
Microsoft Clarity is the right tool when
Stores where mobile and desktop conversion rates differ sharply
That gap is nearly always caused by something specific on a small screen, and it is far quicker to see it in a replay than to infer it from a funnel report.
Merchants without the traffic for reliable split testing
A store with a few thousand sessions a month cannot run meaningful experiments on small changes. Watching real sessions is how those stores find problems worth fixing without waiting for statistical significance.
Teams shipping a new theme, template or checkout change
The first week after a release is when replays earn their place. Behaviour that nobody predicted in staging shows up immediately, and rolling back one section is cheaper than discovering the issue in next month's numbers.
Stores where support describes problems nobody can reproduce
When customers report that a discount field or variant picker misbehaves, finding a recording of it happening turns a vague ticket into a reproducible bug with a device and browser attached.
What to watch
Where accounts drift
Facts about defaults and costs, not criticism. These are the settings we check on every account.
Recordings are kept for a limited window
Session data is retained for a defined period rather than indefinitely, so a replay that proves a point should be saved or clipped when you find it. Longer-term evidence belongs in the analytics property, not here.
Masking needs a deliberate decision
Text can be masked at different levels, and a store handling addresses, order details or anything sensitive should choose that setting on purpose and record why. Defaults are a starting point, not a policy.
It adds a script to the storefront
Any recording tool is another request on the page. Load it with the rest of the marketing tags, keep it out of the critical rendering path and include it when reviewing the performance budget for the theme.
Replays are qualitative evidence
Five sessions showing the same struggle is a strong signal to investigate, not proof of what a change is worth. Sizing the opportunity still needs the rate and the traffic behind the template.
Will a session recording tool slow the storefront down?
It adds a script, so it is not free, but the effect is usually small next to sliders, review widgets and chat tools already on the page. We load it asynchronously and measure the template before and after so the cost is a known number rather than an assumption.
Does it capture personal information from customers?
That depends on how masking is configured and where the script runs. Text masking should be set deliberately, sensitive fields excluded, and hosted checkout pages left to the platform's own rules. The decision belongs in the store's privacy documentation, not in a default setting.
How many recordings should you watch before changing something?
Enough that the same behaviour repeats across different visitors, which is often ten to twenty within a tight segment. If the pattern is not clear by then, the segment is usually too broad and needs narrowing to a single device and template.
Can watching sessions replace split testing?
No, they answer different questions. Replays show why something happens; a test shows whether a change is worth keeping. On low traffic stores replays often carry more of the decision, because a test would take longer than the season it was meant to inform.
Delivered through
Works alongside
Find the leak.
A free Growth Analysis ranks what your store should fix first, by revenue at stake.