Google Analytics (GA) – is a software from Google that is available as both free and paid version. It is one of the most common programs for analyzing the behavior of the visitor to a website. GA has created much of what is today the industry standard in web analytics. Here you can, among other things, follow measured values ​​such as number of visitors, time on the page, which pages are visited, where the visitors come from before they land on your website (google, facebook, direct, email, etc) and much much more.

GA is an important part of evaluating your ROI or ROAS on marketing.

In e-commerce and other business models where purchases are made directly on the site, GA also offers an additional service called (enhanced) ecommerce tracking and means that transactions can be tracked, including number of products, sku codes, tax, returns etc.

If you want to evaluate how your pages stand against each other, your SEO, perform A/B tests or other CRO (Convesion Rate Optimization), GA also has a part in this. Therefore, GA or other alternative web analytics software is central to your growth team.

There is at the moment four versions of GA; Google Analytics, Universal Analytics (two versions) and Google Analytics 4 (we have a whole post about it here). The three first look the same and most of the difference is behind the curtain so people just refer to them as “Google Analytics” or “GA” – however this new version GA4 is different beast since it has a new GUI and since it will take over as being the only one supported come July 2023.

Google Analytics versions:

  1. The first version is the Classic Google Analytics (GA) which uses the ga.js JavaScript Library
  2. The second version is Universal Analytics (UA or GA2) which uses the analytics.js JavaScript Library
  3. The third version is also Universal Analytics (UA, but really UA2 or GA3) but deployed with the gtag.js JavaScript Library instead
  4. The fourth version is Google Analytics 4 (GA4) that also uses the gtag.js JavaScript Library but instead of hit based model uses a new measurement model called “Event+Parameter” model.