What is a unique visitor?

What UV measures

Pageviews tell you how much your site was used; unique visitors tell you how many people used it. If one reader views ten articles in a week, that is ten pageviews, possibly several sessions, but one unique visitor.

The catch is that an analytics tool never sees a person. It sees a browser, a device, a network address, sometimes a login. UV is therefore always an estimate built from whatever identifier the tool has, and the choice of identifier decides how accurate it is.

How tools identify a visitor

• First-party cookie or local storage. A random ID stored in the browser. Accurate for one browser until the visitor clears it, uses a private window, or switches device. Needs consent in many regions.
• Login ID. The user's account ID, also called a user ID. The most accurate across devices, but only covers people who sign in.
• Server-side hash. A one-way hash of the site, the IP address, the browser and a secret salt, computed on the server. Nothing is stored on the device. If the salt is kept, the same browser on the same network is recognised on later days; if the salt changes every day, the same person is a new visitor tomorrow.
• Fingerprinting. Combining many browser attributes into a stable ID. It is treated as personal data in most privacy laws and is increasingly blocked by browsers.

A worked example

Anna reads your blog on her phone on Monday and on her laptop on Tuesday, then signs in on the laptop on Wednesday.

• A cookie-based tool counts 2 unique visitors for the week (phone and laptop).
• A server-side hash with a kept salt also counts 2, as long as each device stays on the same network; with a salt that changes daily, it counts 1 visitor per day and 3 visitor-days for the week.
• A tool that links logins can merge the laptop history with her account, and with a phone login too, count her once.

None of these is wrong. They answer the question with different information.

How TapCub counts unique visitors

By default TapCub counts without cookies. The server computes a visitor ID as a one-way hash of a secret site salt, the site, the IP address (for IPv6, only the first 64 bits) and the browser and operating system without version numbers. Only the hash is kept; the IP address is never stored, and nothing is stored on the visitor's device.

In the default balanced mode the site salt is kept, so the same browser on the same network is the same visitor next week, and weekly and monthly UV are real distinct counts for this method. A new network, device or browser still looks like a new visitor. In strict mode the salt rotates every day, so visits cannot be linked across days: UV is exact within a day and a multi-day UV is the sum of visitor-days.

When you identify signed-in users, TapCub ties their sessions to the user, so multi-day and cross-device counts for those people are based on the account rather than an estimate. Realtime and overview reports show UV next to pageviews and sessions.

For the eight reasons two tools never agree on visitors, and a checklist to reconcile them, see why analytics tools show different numbers.

Common mistakes

• Adding daily UV to get monthly UV. The same person visiting on 10 days would be counted 10 times. Use the tool's own monthly figure.
• Expecting two tools to match. Different identifiers and bot filters mean UV rarely agrees within 10–20%. Compare trends, not absolute numbers.
• Reading UV as customers. Office networks and mobile carriers can put many people behind one address; one person can use three devices. Use identified users for anything about customers.

Related guides: New vs Returning Users, Website Visitor Tracking and Analytics Metrics. See also unique monthly visitors.