Web analytics

Understand what happens on your website

Four things get counted, and everything in the reports is built from them. Once the relationship between those four is clear, the rest of web analytics stops being confusing.

Start freeRead the guides14-day trial, no card
Traffic38 online
Visitors12,482+12.4%
Pageviews31,904+18.1%
Bounce rate38.2%-3.8%
Average visit2m 14s+9.2%
Aug 8Aug 15Aug 22Aug 29Sep 5

Top pages

  • /4,120+9.8%
  • /pricing2,318+22.4%
  • /docs/quickstart1,640+4.1%
  • /blog/tracking-without-cookies1,204+61.7%

Top sources

  • google / organic5,412+14.2%
  • Direct / none2,806-1.2%
  • reddit.com / referral1,140+96.4%
  • x.com / referral884-18.3%

Every metric carries the change against the previous period of the same length.

01

Four things get counted, and everything else is built from them

One person visits on Monday and reads three pages. They come back on Tuesday, read two more and start a signup. That is one visitor, two visits, five pageviews and one event — a visit ends after 30 minutes of inactivity, which is what makes the second day a second visit. Visitors and pageviews then move independently: a quiet week with long visits raises pageviews while visitors fall, which is why the two are always reported side by side rather than as one number.

One person, two days/traffic
  • Visitors1
  • Visits2
  • Pageviews5
  • Events1

02

Every metric splits by every dimension

Visitors, visits, pageviews, bounce rate, average visit and events can each be broken down by page, source, campaign, country or device. A total on its own rarely says anything; the useful move is splitting it until you find the part that changed. Campaigns are read from UTM parameters into their own breakdown, so a newsletter link never gets mixed in with organic arrivals from the same domain.

Visitors by source/traffic
  • google / organic5,412
  • Direct / none2,806
  • reddit.com / referral1,140
  • x.com / referral884

03

One report per question, rather than one surface to configure

There is no report builder to learn and no dashboard to assemble before the product says anything. Five reports, each named after the question it answers, and each of them reporting against the previous period of the same length so a number never appears without context.

Reports/overview
  • Where are visitors coming from?Traffic
  • What is happening right now?Realtime
  • Where do people give up?Funnels
  • Do they come back?Retention
  • Who behaves differently?Segments

What these numbers cannot tell you

A visitor is identified without cookies, from a key that rotates daily, so nobody can be followed from one day to the next. That is what keeps the script cookie-free, and it also means a returning visitor is measured for a group rather than as a named person. Country is resolved from the request and then discarded. Nothing here is sampled: every report is computed over every pageview received, so a small segment stays trustworthy.

Questions about web analytics

These are on the guides index too, alongside every other question the site answers.

What is web analytics?

Measuring how people find and use a website. At minimum that means counting visitors and pageviews. In practice it also means knowing which sources send people, which pages hold attention, where they abandon a flow and whether they return.

What is the difference between a visitor and a visit?

A visitor is a person in the reporting window. A visit is one browsing session by that person, ending after 30 minutes of inactivity. One visitor across two days produces one visitor and two visits.

Does VisitWatch sample data?

No. Reports are computed over every pageview received rather than estimated from a subset, so small segments stay trustworthy.

Free14 days, no card

See your own four numbers

One line of script and the counting starts. Nothing to model first, and no reports to assemble.

14-day free trial. No credit card required.