Web analytics

What web analytics does

Web analytics captures and evaluates what visitors do on a website. It answers questions such as: where do sessions come from, which pages are seen, at which point do users drop off, which channel produces orders. That makes it the basis for almost every decision in online marketing, from budget allocation to the question of which sentence in the checkout is actually understood.

The key metrics

• Sessions and users: How many visits and how many people behind them. The two are not the same, one person can return several times.
• Engagement rate: Share of sessions with genuine activity, measured by duration, page views or events.
• Conversion rate: Share of sessions ending in a purchase, always measured against the matching base.
• Channel: Attribution of the session to a source, with the caveat that attribution is never unambiguous.
• Cart and checkout abandonment: The most important report in a shop, because it points directly at lost revenue.

Why the data is rarely complete

Without consent only technically necessary cookies may be set. A large share of visitors declines, and another share blocks scripts at browser level. The result is underreporting that is not random: people who use ad blockers often behave differently from the average. Absolute numbers are therefore unreliable, while trends and ratios within the same system remain usable.

Two countermeasures are common. First, server side tracking, where collection runs through your own server and is harder to block. Second, modelling, where the provider estimates missing values. Both improve the situation but do not replace clean measurement.

What a clean setup looks like

At the start stands a measurement plan, not a tool choice. The plan records which question is to be answered, which event is needed for it and how that event is defined. Only then does implementation begin. The most frequent mistake is a purchase that fires twice, because both the order submission and the thank you page send an event.

After that, verification belongs in the daily routine. Once a week, sessions, orders and revenue in the analytics tool should be reconciled against the figures in the shop. If revenue deviates by more than a few percent, the measurement is broken and every report built on it is worthless.

How it differs from product analytics

Web analytics looks at the path through the site. Product analytics looks at behaviour inside an application, for instance which feature gets used. In a shop the border is fluid, because account, wish list and cart already have application character. In practice this means that questions about feature usage need additional collection that web analytics does not provide.