Various sites send traffic to our website. The referring site is part of the default data an analytics tool collects. To add more context to a visit, we need to embed meaning in the page that the analytics system can later extract. This is called page tagging or tagging content for web monitoring.
1.The browser requests a page from the web server.
2. The web server returns the page, which contains a reference to a piece of JavaScript provided by the analytics service.
3. The JavaScript is retrieved from the service. In some cases, it may already be cached on the browser or built into the page itself.
4. The analytics service returns the JavaScript to the browser.
5. The browser executes the JavaScript, which collects information about the page request.
6. The JavaScript puts this information into an HTTP request, often for a tiny image. All of these attributes are sent to the analytics service.
7. The service responds with a small object (which can be ignored).
The magic of page tagging happens in step 5, in which the JavaScript collects information to send back to the analytics service. This information includes:
-Information about the technical environment, such as screen resolution, referring URL, or operating system.
-Information about the visitor, such as a cookie that uniquely identifies her and lets the analytics service stitch several pages together into a visit.
-Information within the page itself that provides business context. For example, a retailer might tag a page with “shoes” if it offers shoes for sale. Or it might indicate whether the ad the user saw was part of a specific campaign.
By recording not only technical data on the visit, but also business context, the analyst can then segment outcomes with this context. He can see whether one campaign works better than another, or whether users who buy shoes ultimately buy jackets.
It’s also a full-time job, with large organizations often having entire departments devoted to tagging, experimentation, and reporting. Analytics is still changing, particularly around the integration of other data sources and around the move away from page-centric websites.
