This is my personal experience using 4 different platforms for tracking all the statistics of visitors on the different web sites that I manage. I have and still use the following : AWStats, Google Analytics (GA), Open Web Analytics (OWA) and Piwik. This is not really a feature comparison but rather a look inside each one of them on how they work internally and externally.

A long time ago, in an Internet corner far, far away, there weren’t too many choices for tracking the statistics of visitors on a website. You could read the logs files in text format of your web server, but this was rather a tedious and painstaking job. So quickly people started developing something to resolve that. One of the oldest and most popular at that time was AWStats. At its core, AWStats is the simplest and most straight forward statistics reporting tools. In fact, all of the modern statistics tools, like Google Analytics, Open Web Analytics and Piwik still give the same information, albeit in a different format. The biggest difference with the modern’s one are that they offer more feature, like goals, visitors tracking, alarms, etc. But let me start at the beginning, how each platform works.

AWStats being the oldest and simplest of all, you install it on your web server and using perl and cgi script, it will simply read the logs of your webpage from your HTTP service, like Apache or IIS and will then digest the logs to be displayed in an HTML page for easier reading. It is what I call a log parser. Nothing else. But 10 years ago this was a time saving tool. Being able to look at the statistics of all the visitor in a simple HTML page without having to go through all the logs was really fantastic. You could get your visitor’s IP address, most popular pages being viewed, used browsers, referrers, etc. It really was great (and still is). AWStats simply read the log file and parse the content to another text file for it’s own purpose, really easy to setup and debug. Other platforms existed and are still using the same principle (Webalizer, Analog, etc.). Also for the record, the processing power needed from the server for parsing the logs are pretty low compared to the modern tools, but you still need some. Check out the AWStats demo here (if you can call that a demo).

Then a couple of years later, in 2005 to be more precise, came Google Analytics. In itself, GA is revolutionary because it was one of the first to bring something new to the table that the other statistics platform didn’t offered. I call that visitor tracking – more on that later. The setup compared to AWStats is very different. First of all, by using GA, you agree right from the start to allow Google to be able to compile for their own usage all the statistics of your website’s visitors. Google won’t be able to read directly your private log file from your HTTP service, but instead, you need to install a short portion of HTML code inside each of your website’s page to let Google Analytics get the tracking of all your visitors. All of these are sent back to Google to be stored in a database so that you can access your site statistics through, and only, the GA web interface. Of course, that means sharing your “precious” website statistics with someone else. Google will also block a specific information of your visitors to you. Their IP address, because it’s part of their privacy policy. This is certainly one of the most annoying feature and yet… Not really. After many years, I’ve learned that it’s fun having a specific IP address for the stats, but it’s not that useful (in the general overview for checking the stats, not for security reason I mean). GA does a great job instead for displaying from what country and cities are the visitors coming from. This is mostly more useful compare to having visitor’s IP. And since GA doesn’t read your logs, you also save a lot of processing power for the stats. All is done on the servers of Google, not your own server.

And then GA offered visitor’s tracking. Or what I call, the Holy Grail. You can know with GA on which page exactly does the visitor arrive and from which external site. You can know how much time a visitor spent on each specific page and also on which page he quit your website. This in itself was amazing for marketing people. You can measure with more precision what the visitors are doing on your website. GA will also offer you Goals; You can create some objective for, let say, having 100 visitors to a specific page and you will modify your web site to make sure that no visitors miss that page, track your objective and see if you will reach it. Goals are mostly useful for e-commerce site or page that have ads to monetize the visitor’s click. Again, this is a dream’s come true for a marketing guru. You can take the Google Analytics Tour here.

Now let’s say I’m someone who believe in conspiracy and have a tendency to no trust people in general. If I were that person, I would be freaked out by sharing my precious visitor’s stats with someone else that I don’t even know. But at the same time, Google Analytics is offering me something that no one else does. Or is it? Come into the picture Open Web Analytics (OWA for short, released in 2009) and Piwik (released in 2008). Both are free and open source license. Both platform offers the same thing like GA (tracking, goals, etc.) you get back your visitor’s IP address and you still get to keep the geolocation. But instead, you’ll need to install the platform on your web server and then insert the HTML code inside each of your page to be able to track your visitor (no log parsing). You will also need to have a running database (probably MySQL) to put all of your sensitive visitor’s record inside. So yes, these 2 tools are asking for more setup and configuration, and of course, more processing power from your server. But since I’m paranoid, I no longer have to share with Google my precious statistics!

OWA is in itself a clone of GA. If you have been using GA for a while and want something similar, you don’t like learning a new web interface, then OWA is your first choice. Not everything is the same of course, but almost. You can check a demo of OWA here.

And lastly, the weirdest one of them all, is Piwik. When I first started using Piwik, I really didn’t liked it because, well, I had been using GA since the first day it was released and Piwik boast itself for not presenting the stats like GA. It still offers the same information like GA and OWA but the presentation is really different. And yet, there lies Piwik greatest strength. After a couple of days using it, you get the feel of how it works and to my personal taste, it is now, today, my favorite one. The information is presented in a good layout, you can get in a quick glance your visitor’s tracking and once you have grown accustomed to the interface, you can easily use it. Also, worth mentioning is the upgrade process for Piwik. It is a charm, just like the WordPress blog platform, you simply click on the button for the upgrade through the web interface and bingo, it’s done. OWA still doesn’t offer that. Piwik Demo here.

So after all I said, choosing a web stats platform is really a question of personal choice. If you don’t have too much resources available and are willing to share your visitor’s stats, then GA would be the best choice. If you don’t really care about visitor’s tracking and are really focused on hard and concrete stats, AWStats still does the job today. And if you prefer all of that together without having to share any of it, then you can take your pick between OWA and Piwik, if you are willing to spend all the time and energy required for it.