Skip to content
Tutorials Center
My learning

Topics

Lesson 23 of 38 ยท Analytics and Attribution

Analytics Fundamentals and Marketing KPIs

Learn how web analytics works: events, sessions, users, dimensions and metrics, funnels, and how to choose KPIs that guide decisions rather than flatter you.

Analytics turns clicks and visits into decisions. Without it, marketing is guesswork. With too much of it, you drown in dashboards and still cannot answer a simple question: is what we are doing working? This lesson gives you the vocabulary and a method for choosing numbers that matter.

How tracking works

A tracking tag (a small piece of JavaScript) on your site or app sends an event to an analytics tool whenever something happens: a page view, a click, a form submission, a purchase. Each event carries parameters (the page address, product name, value). The tool stores events and groups them into sessions (a period of activity by one visitor) and users (an identifier, usually a cookie or a sign-in ID, that links sessions).

Modern tools such as Google Analytics 4 are event-based: nearly everything, including page views, is an event. Other tools (Matomo, Plausible, Adobe Analytics, Mixpanel, PostHog) use similar ideas.

From action to report

Action A visitor clicks, scrolls or buys
Tag Script records an event
Storage Events are grouped into sessions
Report You slice by dimension and metric
1 / 2

Dimensions and metrics

Every report combines two kinds of columns:

  • A dimension is a category: channel, country, device, page, campaign.
  • A metric is a number: sessions, users, conversions, revenue.

"Sessions by channel" has the dimension channel and the metric sessions. Most analysis is choosing the right dimension to slice a metric by.

Common metrics, and their traps

Metric Meaning Watch out for
Sessions Visits Bots inflate it; one person can create many
Users Distinct people (approximately) Cross-device people count twice
Engagement rate Share of sessions that were engaged (e.g. lasted over a set time, had a conversion or several views) Definitions differ between tools
Bounce rate Share of single-page, no-interaction sessions A blog reader who finds the answer is not a failure
Conversion rate Conversions / sessions (or users) Choose and keep one denominator
Revenue per visitor Revenue / visitors Skewed by a few large orders

Different tools define the same word differently. Write the definition next to every KPI so a number means the same thing next quarter.

Vanity metrics versus actionable metrics

A vanity metric looks good but does not change a decision: total page views, followers, impressions. An actionable metric ties to a result and tells you what to do next: cost per lead, trial-to-paid rate, revenue per channel. Ask of every metric: "If this number doubled, would we act differently or earn more?" If not, move it down the dashboard.

KPIs per funnel stage

Pick one or two KPIs per stage of the marketing funnel:

  • Awareness: reach, branded search volume, share of voice.
  • Consideration: engaged sessions, email sign-ups, content downloads.
  • Conversion: leads, sales, cost per acquisition, conversion rate.
  • Retention: repeat purchase rate, churn, net revenue retention, NPS.

Each KPI needs an owner, a target, a time period and a data source.

A funnel in numbers

JS Funnel drop-off report

Press Run (or Ctrl + Enter) to see the console output here.

No output.

The weakest step is where conversion optimization should start.

Segmenting: the average lies

An overall conversion rate can hide big differences. Compare by channel, device, new versus returning, and country.

JS Segment conversion rates

Press Run (or Ctrl + Enter) to see the console output here.

No output.

Data quality checklist

  1. Exclude internal traffic and known bots.
  2. Define conversions once and test them (complete a test purchase and confirm it appears).
  3. Use consistent campaign naming and UTM tags.
  4. Watch for duplicates, such as a purchase event firing on every reload of the thank-you page.
  5. Respect consent choices; see privacy and compliance. Some visitors will not be tracked, so tool numbers are estimates.

Common mistakes

  • Treating any tool's number as the exact truth.
  • Comparing periods without allowing for seasonality.
  • Reporting too many KPIs. Five to eight are plenty for a team.
  • Changing metric definitions without noting it.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. In 'sessions by country', what is 'country'?

    Dimensions are categories; metrics are numbers.
  2. 2. Which is the best example of an actionable metric?

    It ties spend to results and shows where to shift budget.
  3. 3. Why is it risky to read only the overall conversion rate?

    Segmenting by device, channel or audience reveals the real problems and winners.
  4. 4. What should accompany every KPI?

    Clear definitions keep numbers consistent and actionable.

Summary

  • Analytics tools record events and group them into sessions and users.
  • Reports slice metrics (numbers) by dimensions (categories).
  • Prefer actionable KPIs over vanity metrics; keep a few per funnel stage.
  • Segment before concluding; averages hide differences.
  • Keep data clean and respect consent.

Next: UTM Tracking and Campaign Naming