How to use is it real? Which changes are bigger than noise
Every headline number, channel, top page, country and device against the period before, with a plain label: Clear change, Probably real or Could be noise.


What it does
quietstats compares the range with the same number of days before it. Counts (visitors, pageviews, visits, and visitors per channel, page, country and device with 10+ across both periods) use a Poisson test; bounce rate uses a two-proportion test. Clear change means about 3 standard errors or more; Probably real means 2 or more. The biggest signals come first.
Why it matters
Small sites swing a lot by chance. Knowing which changes are real stops you celebrating or panicking over noise.
How to use it
- Open Deep dives → Analysis lab → Is it real? (or go to /app/lab/isitreal?site=YOUR_SITE).
- Choose 7, 30 or 90 days; long-range trends read best at 90.
- Download any table with its CSV link; small numbers show as “<5” there too.
How it works
- Open your dashboard, go to Deep dives and scroll to Analysis lab.
- Under Traffic over time, tap Is it real?.
- Pick 7, 30 or 90 days (and any options at the top), read the “How to read this” notes, and download any table as CSV.
Privacy
- Collected: Nothing new: it reads the pageviews (and goals, events or speed checks) you already have
- Not collected: Anything about a single visitor
- Not collected: Links between days: visitor ids still reset at midnight
- No cookies, and nothing stored on the visitor’s device.
- Reports show totals only: never a visitor list, IP address or single visit.
Limits
- Statistics describe chance, not causes; something else may have changed too
- Up to 92 days and 200,000 pageviews per screen
- Read from the database only when you open it, then kept for 10 minutes
Plans
Every plan, including Free. quietstats is free during launch: the Free plan covers 1 site and 5,000 pageviews a month, and no card is needed. See pricing.
Questions
Is this an A/B test?
No. It compares two periods. For two versions you tag yourself, use A/B tests.
Why are small pages missing?
Rows need 10+ visitors across both periods; below that, the test cannot say much.
Related
Feature page: Is it real? Which changes are bigger than noise → What it is, why it matters, privacy and how other tools do it.