This article gives you an overview of different technologies for web analytics. We look at the possibilities and limitations of the different technologies. The three technologies that we will look at in this article are: server-side, client-side and hosted solutions. For each of the three there are numerous providers who have their own take on what terms like visits, sessions, page views and so on, mean. Feel free to take a look at the Siteimprove definition of web analytics terms. In this article we will look at the more general features of each of the three technologies.
Server-side tracking
The most common way to do server-side tracking is to do log file analysis. The one big advantage of log file analysis based tracking is that you don’t have to make changes to your web programming or server. This is typically a standalone program that needs access to your web server logs, and then does all the computing from the logs. Using the web server logs also means that you have to keep your logs for as long back as you want the tracking to cover. Since the web server normally logs every action including individual elements on a page like images and other graphics, logs take up quite a lot of space.
Another disadvantage of server-side tracking is that it is the server that is responsible for gathering the tracking information. This has the effect that pages cached in the users web browser or a web proxy will not be counted because the page wasn’t fetched from the server. This also means that the tracking is only able to track what info the server is able to gather about the user, and not further info about user screen resolution and other data like that.
The web server logs will also contain entries from whenever a search engine spider or another web crawler has fetched pages. So it is up to the log analysis program to sort these entries out, and keep track of which spiders exist.
Server-side tracking summed up
While server-side tracking software often is inexpensive or even free, and easy to install, it typically means acquiring new hardware to store the log files and run the software from. At the same time the software needs to be up to date, to be able to exclude web crawlers from the statistics. Furthermore the statistics will lack user info, and not be able to count users who load a cached version of the website.
Client-side tracking
Normally client-side tracking is done by putting a small JavaScript on all the pages that needs to be tracked. The script will be loaded every time the page is loaded, whether the page is loaded directly from the server or from a proxy or web cache. This is one advantage of using client-side tracking. Another advantage that comes from using JavaScript is automatic exclusion of most web crawlers from the statistics since the crawlers don’t execute scripts.
Scripts also have the advantage of being executed from the user computer, which allows them to gather much more information about the user. This includes putting cookies on the client computer that can be used for tracking returning visitors and more. Scripts collect more information from each visitor than server logs, but the collected data take up less space, because it only keeps one entry per page seen and not an entry for each element on the page.
Like server-side tracking client-side tracking also has it disadvantages. One of them is that you have to install the script on all the pages of your website, while this might sound overwhelming, for most websites it involves editing in one or two templates. Another disadvantage of script-based tracking is that some users disable execution of scripts or the use of cookies in their browsers. These users will not be part of the statistics. Like server-side tracking client-side tracking also involves acquiring both software and hardware to run the tracking on, both of these demand manpower for maintenance.
Client-side tracking summed up
Client-side tracking provides more accurate tracking data than server-side tracking. It also gives more details on the users system. Client-side tracking software is typically more expensive than server-side and like server-side involves a hardware investment and manpower for maintenance. Client-side tracking is vulnerable to users disabling scripts or cookies in their browsers, because they will not be counted.
Hosted tracking
Hosted tracking works like client-side tracking by putting JavaScript on each page of the website, but instead of having a local JavaScript, this JavaScript fetches the entire JavaScript from the analytics provider’s server and also sends the tracking information back to the analytics provider’s servers. This gives hosted tracking the same advantages as client-side tracking and more. Hosted tracking involves no hardware or software investments or manpower to maintain these. The tracking software is automatically updated whenever the analytics provider releases a new version. Most hosted tracking solutions provide a web interface to see your statistics, this gives you the possibility to view your analytics from anywhere with web access.
Since hosted tracking is script based it has the disadvantages of needing installation on all pages of the website like client-side tracking, and is also vulnerable to users who have disabled scripts and cookies in their web browsers.
Another concern that hosted tracking providers often meet is that a script that is fetched from an external server, and puts cookies on the users' computers, will use third party cookies, but most solutions provide the cookie from the websites domain and therefore they are first party cookies and there is no reason to be concerned.
One last thing that might be an issue or disadvantage with some providers of hosted tracking is the question of data ownership and data confidentiality. So make sure you ask your provider who owns the data that they collect on your behalf, and ask your provider what happen to the data when you terminate the agreement.
Hosted tracking summed up
Hosted tracking is similar to client-side tracking but with the added advantage of automatic updating. Even though the hosted tracking normally has a monthly fee, this fee is typically less than the total cost of acquiring and maintaining hardware and software for server-side or client-side solutions.
SiteAnalyze hosted web analytics
At Siteimprove we have chosen to provide our SiteAnalyze web analytics as a hosted tracking solution. We have chosen this model because we think it outperforms the other models. In SiteAnalyze the clients owns the data and all tracking data will be deleted upon termination of the agreement. Another reason we think that a hosted solution is the best, is that every time we have made improvements or added features, the clients have access to them right away, without the need to install or upgrade anything.