Analytics
Microsoft Clarity: free, unlimited, gone in 30 days
Microsoft Clarity is free forever with no traffic limit, no site limit, no user limit and no sampling — and it deletes your session recordings after 30 days. Every paid heatmap tool meters how much you can record. Clarity meters how long you can keep it, and on that axis it is the least generous tool in the category. That single fact decides whether free is the right choice for your site.
What This Covers
The short version What free actually buys The meter is time, not volume Two clocks, 9x apart Why small sites should invert the usual advice The sampling nobody mentions What Clarity will not tell you Two rules that can disqualify your site What we could not verify FAQThe short version
Read in September 2026 from Microsoft's Clarity documentation on Microsoft Learn and from clarity.microsoft.com.
What free actually buys
Clarity's free tier is not a trial or a freemium tier with a conversion funnel behind it. Microsoft's FAQ answers the pricing question in two sentences: "Clarity is a free service forever. You never encounter traffic limits or are compelled to upgrade to a paid version." We looked for a paid plan; there is not one. The absence of limits is broader than most people expect, and worth stating precisely because this is the part Clarity genuinely wins:
- Traffic. "There are absolutely no limits on the number of sites per account" and "there are no traffic limits on these sites." Microsoft says Clarity processes more than a petabyte of data from over 100 million users a month.
- Sampling. "No sampling (that is, tracking only a portion of your site's traffic)." Paid analytics tools frequently sample at volume; Clarity says it does not.
- Team members. "There are no limits to the number of members you can add to each project" — though only 10 invites can be pending at once.
That combination is genuinely unusual: put Clarity on every site you own, invite whoever you like, never see a bill. Nothing below contradicts it. The argument here is only that the thing being rationed is somewhere else.
Quoted from Microsoft's Clarity FAQ on Microsoft Learn, read September 2026.
The meter is time, not volume
Every comparison of heatmap tools lines them up on sessions per month, because that is how the paid ones are sold. On that axis Clarity wins by an unbounded margin and the comparison ends. Line the same tools up on how long they keep a recording and the order reverses completely.
| Tool | Cheapest plan | Volume cap | Recordings kept |
|---|---|---|---|
| Microsoft Clarity | $0 | None published | 30 days |
| Lucky Orange (Build) | $32/mo billed annually | 3,500 sessions/mo | "Starts at 60 days" |
| Crazy Egg (Starter) | $29/mo billed annually | 5,000 pageviews/mo | 6 months |
Vendor-published figures read September 2026 from luckyorange.com/pricing and crazyegg.com/pricing. Lucky Orange lists 90, 180 or 365 days as paid storage upgrades. Crazy Egg's higher tiers state 2 years. Our arithmetic on the rest.
The free tool with no volume limit has the shortest memory in the category. Crazy Egg's entry plan keeps recordings about six times longer than Clarity — roughly 182 days against 30 — for $29 a month, or $348 a year. Lucky Orange's entry plan keeps them twice as long before you buy any storage upgrade.
What makes this more than trivia is who it applies to. Lucky Orange's cheapest plan caps you at 3,500 sessions a month and Crazy Egg's at 5,000 tracked pageviews. A site below those caps gets nothing from Clarity's unlimited volume, because it was never going to exceed a paid tool's limit anyway. Most readers of this site sit in exactly that band, so their real choice is not free-versus-paid on capacity — it is 30 days for nothing against six months for $348 a year, and the answer depends entirely on whether they would ever open a four-month-old recording. Above those caps the sum flips hard and Clarity becomes very hard to beat, because paid tools charge steeply for volume and Clarity charges nothing for any amount of it.
Two clocks, 9x apart
Clarity runs three retention policies, not one, and the gap between them is where people get caught out:
| Data type | Retention | What it drives |
|---|---|---|
| Playback data | 30 days | Session recordings |
| Click data | 9 months | Heatmaps, dashboard metrics |
| Labelled or favourited sessions | 9 months | Recordings you deliberately saved |
Quoted from Microsoft's Clarity data retention documentation, read September 2026. Microsoft states that data on Clarity servers, including backups, "is deleted after the retention period and cannot be recovered."
Nine months is about 274 days, so the aggregate window is roughly 9.1 times longer than the recording window — meaning that for about 89% of the period your heatmap covers, the recordings that would explain it no longer exist. You can see that 300 people clicked a dead spot last March. You cannot watch a single one of them do it.
The two are connected at a technical level, which is why this bites. Microsoft's heatmap troubleshooting page states that "Heatmaps are picked from recordings" — so the aggregate number and the evidence behind it age out on different schedules.
The operating rule this produces is a cadence, not a setting. If you review Clarity less often than once every 30 days, you are structurally guaranteed to miss recordings — a monthly review on the 1st has already lost everything from the first days of the previous month. A weekly look means every recording is seen roughly four times before it expires. This is the single highest-value habit change on this page and it costs nothing.
Why small sites should invert the usual advice
The standard advice for a new site is to watch session recordings, because with little traffic you can afford to watch them all. The retention policy makes that backwards. Recordings expire 30 days after they are made, so the number available at any moment is about one month of traffic — and it never grows. A blog doing 200 sessions a month holds roughly 200 recordings today and 200 a year from now; waiting does not accumulate evidence. Click data behaves the opposite way: kept 9 months, it builds to about nine months of traffic, which on that same blog is nine times more evidence than the recordings ever hold at once.
| Monthly sessions | Recordings on hand (steady state) | Sessions behind a mature heatmap |
|---|---|---|
| 200 | ~200 | ~1,800 |
| 1,000 | ~1,000 | ~9,000 |
| 3,500 | ~3,500 | ~31,500 |
Our arithmetic, applying Microsoft's published 30-day and 9-month retention windows to a steady traffic rate. Real figures vary with traffic growth and with how many pages that traffic is spread across.
So on a small site, heatmaps reach a usable sample and recordings never will. Use the heatmap to find the page and the element; use recordings while they exist to answer a specific question about it. And if you are still building the traffic that makes any of this meaningful, the groundwork in building a display-ad blog and finding low-competition keywords matters far more than your analytics setup.
The sampling nobody mentions
Clarity advertises "no sampling" and that claim is about collection — it tracks all your traffic rather than a slice. Retention is a different matter, and Microsoft's FAQ is precise about it: "Clarity retains recordings for 30 days from the time of recording. However, Favorite recordings and randomly selected sample of recordings are retained for up to 9 months."
Past 30 days, what survives is whatever you deliberately saved, plus a random sample you do not choose and cannot rely on. There is no export and no way back: deleted data "cannot be recovered."
Which makes favouriting the entire archive strategy. If a recording shows something you will want to cite in three months — a checkout failure, a confusing nav pattern, a mobile layout breaking — label it inside the first 30 days or accept a coin toss. Labels are themselves kept 9 months. It is a manual habit with a hard deadline, and nothing in the interface will warn you that it has passed.
What Clarity will not tell you
This page used to carry a title promising to double your click-through rate. That claim was not ours to make, and it is worth retracting explicitly because the vendor contradicts it: Microsoft's own article on calls to action says "The best way to understand which CTAs are delivering your objectives is to test them" and publishes no expected improvement figure whatsoever. Clarity is a diagnostic instrument, not a treatment. Treat any specific multiplier as marketing.
There is a second, subtler gap. Clarity's frustration metrics are its best-known feature, and Microsoft publishes no numeric threshold for any of them:
- Rage clicks are flagged when a user "clicks multiple times in a clustered area in rapid succession." How many clicks, how close, how fast — unstated.
- Dead clicks are clicks that get "no feedback in a reasonable amount of time." What counts as reasonable — unstated.
- Quick backs are measured against "the dwell time for that page", a value that varies per page and is not shown to you. Microsoft adds that "In future releases, this metric adapt to your specific site."
- Excessive scrolling is scrolling "higher amounts of vertical scrolling than the expected average" — an average Clarity does not publish.
Quoted from Microsoft's semantic metrics documentation, read September 2026.
None of that makes the metrics useless, but it means you cannot calibrate them, reproduce them, or compare a Clarity rage-click count against another tool's. Use them to rank your own pages against each other over time — a comparison the hidden thresholds do not spoil — and do not treat the absolute number as a measurement. The same caution applies to Clarity's Copilot summaries, which carry Microsoft's own documented warning about the "risk of generating ungrounded content" and takeaways that "shouldn't be treated as expert recommendations." Take that literally.
Two rules that can disqualify your site
Two published conditions decide whether Clarity is usable for you at all, and neither appears in the roundups.
Audience age. Microsoft states plainly: "Clarity shouldn't be used on any websites/apps targeting users under the age of 18." That is a flat restriction on the kind of site, not a setting.
Consent, and the data-quality trap behind it. In the EEA, UK and Switzerland, Clarity requires an explicit consent signal before it sets cookies. The legal side is unsurprising; the measurement side is what catches people. Without consent, Microsoft says Clarity "cannot track a continuous user journey. For instance, if a user visits page A and then page B during the same visit, Clarity treats these as two separate sessions."
That is not just lost recordings — it silently inflates your session count and deflates your pages-per-session, with no error message. If your European traffic is material and your consent signal is not wired up, the dashboard is not merely incomplete, it is wrong in a specific and predictable direction. Microsoft confirms that using Clarity without consent does not violate its terms, so nothing will stop you; the numbers just quietly stop meaning what you think. Clarity supports a JavaScript consent API and platforms such as CookieYes, with Google Consent Mode support listed as "coming soon".
Quoted from Microsoft's Clarity FAQ, read September 2026.
What we could not verify
- Hotjar, entirely. hotjar.com/pricing now 308-redirects to contentsquare.com, which we could not fetch. No Hotjar price, session limit or retention period appears anywhere on this page. Hotjar is the tool most often compared against Clarity, and its absence from our table is a gap rather than a judgement. Mouseflow was also unreachable.
- Clarity's own pricing page. clarity.microsoft.com/pricing rendered no content to us. The "free service forever" wording comes from Microsoft's FAQ on Microsoft Learn instead, which is first-party and unambiguous.
- Microsoft's published benchmarks. clarity.microsoft.com/benchmark returned no data to automated fetch, so no industry median for rage clicks, dead clicks or scroll depth appears here. If you want a comparison figure for your own numbers, that page is the place to look manually.
- The randomly retained sample. Microsoft says a "randomly selected sample of recordings" survives to 9 months but publishes no sample rate. We make no claim about what share of your recordings that is, which is precisely why favouriting rather than hoping is the advice above.
- Metric thresholds. As set out above, no numeric definition of a rage click, dead click, quick back or excessive scroll is published. We state the wording and no number.
- Whether the 9-month window was recently shortened. Several current third-party comparisons quote 13 months. Microsoft's live documentation says 9 months and we use that, but we could not find a first-party changelog confirming a change, so we describe the 13-month figure as outdated rather than as a documented reduction.
These are September 2026 snapshots of documentation that changes without announcement. The 30-day and 9-month retention windows are the load-bearing facts on this page; a change to either would move the conclusion, not just a number.
The one-line version
Clarity gives away unlimited volume, which a small site does not need, and rations the thing it does need — so review weekly, favourite anything you will want later, and only pay a rival if six-month-old recordings are genuinely worth $348 a year to you.
What AdSense actually requires Keyword tools under $50FAQ
Is Microsoft Clarity really free?
Yes, and unusually so. Microsoft's FAQ states that Clarity is "a free service forever" and that "you never encounter traffic limits or are compelled to upgrade to a paid version." There is no paid tier to upgrade to. The same page states there are "absolutely no limits on the number of sites per account," no limits on team members per project, and "no sampling," meaning Clarity tracks all your traffic rather than a portion. We found no volume cap of any kind. The catch is not a hidden charge or a usage ceiling — it is time. Session recordings are deleted 30 days after they are made, the shortest recording window of any tool we priced.
How long does Microsoft Clarity keep session recordings?
Thirty days. Microsoft's data retention page states that playback data is retained for 30 days, and its FAQ repeats that recordings are kept "for 30 days from the time of recording." Two longer windows are often confused with it: click data, which heatmaps are built from, is retained 9 months, and sessions you label or favourite are also kept 9 months. Microsoft states that data on its servers, including backups, "is deleted after the retention period and cannot be recovered." Note that several widely shared comparisons still quote 13 months for the longer window — Microsoft's current documentation says 9.
Should I pay for a heatmap tool instead of using Clarity?
Only if you need recordings to last longer, because that is the one axis where paid tools beat Clarity. Crazy Egg's cheapest plan is $29/mo billed annually and states 6 months of recordings storage — roughly six times Clarity's window — but caps you at 5,000 tracked pageviews a month. Lucky Orange's cheapest is $32/mo billed annually with storage that "starts at 60 days" and a 3,500 monthly session cap. Below those caps you are not buying capacity at all, since Clarity's unlimited volume is worth nothing to a site that would never hit a paid tool's limit. You are buying time, at about $348 a year in Crazy Egg's case. Above those caps the sum flips and Clarity becomes very hard to beat.
Can Microsoft Clarity double my click-through rate?
No tool can promise that, and Microsoft does not. Its own article on calls to action says "The best way to understand which CTAs are delivering your objectives is to test them," and publishes no expected improvement figure at all. Clarity is a diagnostic instrument rather than a treatment: it shows where people click, how far they scroll, and where they click things that do nothing. Any change in click-through rate comes from what you do about that. Be sceptical of any page quoting a specific multiplier — including the older version of this one — because the vendor publishes no such number.
Does Microsoft Clarity work on a low-traffic blog?
The heatmaps work well and the recordings mostly do not, which reverses the usual advice. Because recordings expire after 30 days, the number available at any moment is about one month of traffic and it never grows — a blog doing 200 sessions a month holds roughly 200 recordings today and 200 a year from now. Click data is kept 9 months, so it accumulates to about nine months of traffic: on that same blog, roughly nine times more evidence than the recordings ever hold at once. On a small site, use heatmaps to find the page and the element, and dip into recordings for specific questions while they still exist.