What Is Google Analytics?
Google Analytics is Google’s free web analytics platform. It tracks who visits your website, where they came from, what they did once they arrived, and which actions led to conversions. The current version, Google Analytics 4 (commonly called GA4), replaced the older Universal Analytics in July 2023 after a long transition period. GA4 uses an event based data model rather than the session based model that powered Universal Analytics for over a decade.
If you run any kind of business website, you should have GA4 installed. It is free, it is the most widely supported analytics platform on the web, and it is the foundation that the rest of your measurement stack plugs into. Most other tools, from heatmaps to attribution platforms, expect GA4 to be present and assume its data is clean.
Why Did Google Move From Universal Analytics to GA4?
Three reasons converged. First, browser privacy changes broke the third party cookie based tracking that powered Universal Analytics. Apple’s Intelligent Tracking Prevention, Firefox’s Enhanced Tracking Protection, and the eventual end of third party cookies in Chrome made the old model unworkable. Second, the session based model did not handle apps, server side data, or modern customer journeys cleanly. Users move between mobile web, mobile app, and desktop in a single buying decision, and the old model could not reliably stitch those moments together. Third, machine learning improved enough that Google could deliver predictions and modeled conversions inside GA4, which Universal Analytics never offered.
The transition was painful for most teams because GA4 reports look and behave differently. Many of the metrics changed names and definitions. Engagement rate replaced bounce rate as the primary engagement signal. The reporting interface was rebuilt around explorations rather than fixed reports. Most clients we audit had GA4 installed but never actually configured properly, which means the data they are looking at is not telling them what they think it is.
What Should You Actually Track in Google Analytics?
The basics start with traffic by acquisition source. GA4 splits this into Default Channel Group categories: Organic Search, Paid Search, Direct, Email, Organic Social, Paid Social, Referral, and Display. Watching these by week tells you which channels are healthy and which are slipping. Engagement metrics come next: engagement rate, average engagement time, pages per session. These reveal whether visitors are actually reading or whether the traffic is empty volume.
The most important configuration step is setting up key events for the actions that map to revenue. A purchase, a signup, a lead form submission, a phone call, a download. GA4 will not measure conversion rate properly until these are configured, and most accounts we audit have either too few key events or too many. Ecommerce sites should also implement the full enhanced ecommerce event sequence: view item, add to cart, begin checkout, purchase. Each event unlocks different reports and powers the funnel analysis features that are some of the most useful parts of GA4.
What Are the Common Mistakes Teams Make With GA4?
The most common mistake is treating GA4 as Universal Analytics with a different interface. The data model is genuinely different and reports that look familiar are not always measuring what they used to. Engagement rate is not the inverse of bounce rate. Sessions are calculated differently. Direct traffic includes scenarios that Universal Analytics counted elsewhere. Teams that try to compare GA4 numbers to historical Universal Analytics data without understanding the redefinitions usually conclude that traffic crashed when really only the metric definition changed.
The second common mistake is not setting up the GA4 to Google Ads link, which means conversion data does not flow into the ads platform and bid optimization runs blind. The third is forgetting to configure data retention, which defaults to two months for some properties and silently truncates the longer term comparisons most reports need. The fourth, and the most serious, is leaving spam traffic unfiltered. Every GA4 property accumulates referral spam, and a few bot heavy weeks can distort year over year comparisons in ways that take months to spot.
How Do You Use Google Analytics in Practice?
Install GA4 through Google Tag Manager rather than hardcoding the snippet, because changes are easier to manage that way. Configure key events for the actions that actually matter to revenue. Build a small set of standard reports for the questions you ask weekly rather than relying on GA4’s exploration interface for every check. Connect GA4 to Google Search Console for organic search data and to Google Ads for paid data, so the picture is unified rather than scattered across tools.
Most teams underuse GA4 because nobody set it up correctly in the first place. We configure GA4 properly inside our Analytics service so the data is honest from day one. For a focused look at the most misunderstood GA4 metric, read how to diagnose GA4 engagement rate before rebuilding anything and the broader engagement rate and bounce rate guide. For related concepts, see Bounce Rate, Engagement Rate, and Attribution Model. The bottom line: GA4 set up well is one of the highest leverage hours of work in marketing. Set up badly, it is just decoration.