Web Analytics

Analytics records what people do on your site. Its value in marketing is specific: it is the one place where every channel in this course can be compared on the same terms.

Traffic sources — often called acquisition — shows where visitors came from, grouped into channels:

This is the report that answers "is any of this working?" — and it is the reason all the other numbers become interpretable, since 500 visitors means something different depending on whether you paid for them.

• Users and sessions. People, and visits. One person visiting three times is one user and three sessions.
• Engagement. Whether visitors did anything beyond arriving. Modern analytics tools measure this in various ways; the underlying question is the same.
• Pages per session. Whether people explored or left immediately.
• Conversions. The actions you defined as mattering — the only metric that connects to the business directly.
• Landing pages. Which pages people arrive on, which is often not the pages you expect.

Without tagging, analytics guesses where traffic came from, and often guesses wrong — a link clicked in a messaging app or a PDF frequently shows up as "direct".

UTM parameters are labels added to a URL that tell analytics exactly where a click came from:

The three that matter: utm_source (which site), utm_medium (what kind of link), utm_campaign (which campaign). Tag every link you place deliberately — in emails, social posts, and advertisements — and the traffic sources report becomes genuinely accurate instead of approximate.

Analytics with no conversions defined reports activity without outcome. Define the actions that actually matter to you — an enquiry submitted, a purchase, a signup, a call button tapped — before spending time reading anything else.

Analytics numbers are estimates, not exact counts. Ad blockers, privacy settings, consent choices, and people switching devices all mean some activity goes unrecorded. Different tools will disagree with each other, and both will disagree with the advertising platforms.

This is fine as long as you use them for what they are good at: comparing periods, comparing channels, and spotting trends and sudden changes. Treating a number as an exact fact is what leads to bad decisions.