What web analytics is, and how the process works

Web analytics is collecting, analysing and reporting data about your website, so you can decide based on what happens rather than what you assume.

Concretely it means measuring actions: how many people arrive, which pages they see, how long they stay, what they do. But measuring is only half of it. Without a sense of what you want to know, it only produces numbers.

Web analytics is more than gathering data and turning it into a report. In practice it looks like this:

1. Define your goals. What are you trying to achieve? More revenue, happier customers, more awareness. This comes before everything else.
2. Translate goals into KPIs. What do visitors have to do to move those goals? A KPI tree helps make that explicit.
3. Collect data. Only now. Usually through a piece of JavaScript on your site that passes each action to your analytics tool.
4. Process and report. Give the KPIs context. A number without a comparison isn't an insight.
5. Form a strategy. What do your engagement figures say about how people feel about your content, and what will you do about it?
6. Experiment. Form a hypothesis and test two versions against each other. That's how you know a change works instead of hoping it does.

Your website is where you sell, publish or capture leads. That makes it the place where you can learn the most about your audience.

• You learn who arrives. Which pages are popular, how long people stay, what resonates with which audience.
• You measure conversions. When they happen, how often, and where people drop out before they do.
• You understand your marketing. Which channel brings the most traffic is one thing; which brings the most interested visitors is the question that counts.
• You see which content works. And with it, ideas for what the next piece should be.

Web analytics splits into two main categories. Off-site looks at behaviour beyond your own site, to gauge your potential audience and your position in the market. On-site measures what happens on your site: forms, purchases, clicks.

Most vendors do the latter through page tagging, a piece of JavaScript on your pages that acts as a sensor and sends data to the tool.

We focus on on-site analytics. Those tools fall roughly into four categories, distinguished by the question they answer.

• Marketing analytics. Where do visitors come from and how do they use your site or app? The basis for channel decisions.
• User experience analytics. How do people experience your design? Quantitative and qualitative, with A/B tests, multivariate tests, feature toggles and content personalisation.
• Product analytics. How do people use your product or service? Product teams track engagement and behaviour over time here.
• B2B lead analytics. Which organisations visit your site? These tools match IP addresses to a company database, which feeds your sales funnel.

Web analytics gives you the data to improve your site deliberately instead of by feel.

It has downsides too: you can drown in metrics, data is never fully accurate, and privacy limits what you may collect. So pick a tool that fits what you actually want to know, and measure less than you can.