Analytics on the Cheap: Six Free Stats Packages for the Startup or Small-Business Owner

Many important decisions begin with analysis of website statistics. Statistics packages can give a website owner a different perspective and allow him/her to get more insight into visitors’ behaviors, habits, and the key terms that bring people to the website. Key marketing and business decisions can be founded on website statistics analysis. Once your website is up and you’re looking for traffic, all users should consider a statistics package.

Fortunately, in this day and age, you can find just about everything for a relatively cheap price or even free. There just so happens to be a plethora of statistics packages available, and while the free options aren’t necessarily as comprehensive as the paid statistics software applications, they are worth noting. With Gatineau around the corner (and we can still anticipate its launch), there are other free analytics applications that you should consider using on your site. Here are a list of six that you might want to look into if you are looking for statistics. I have written this guide with a basic overview of each.

1. StatCounter. This free analytics tool is fairly limited (logs up to 100 pageloads unless you upgrade), but in its free form, features a nice friendly graph with the number of visitors who have visited a particular site.

StatCounter also offers a keyword analysis tool, allowing you to see which keywords brought visitors to your site. From this data, you can get more information about which keywords are well optimized and which keywords you wish to convert but are not currently doing much for you.

Further information is available, including lists of the first pages a visitor went to, the page a visitor left from, a list of where the site’s users came from (with a full referral link!), the length of a user visit, which search engine dominated on search from the big four (see below), a list of most recent visitors as well as how they got there (especially useful to determine whether people are clicking through PPC campaigns), the length of a user’s visit, the number of returning visits, a list of operating systems and visitor screen resolutions, and a some other useful tools for a basic analysis.

Here’s an image of what StatCounter calls “Search Engine Wars,” allowing users to see which search engines the traffic came from:

You can also learn more about the users that visit your site with the following Google Maps implementation on the backend:

(Note: The images here reflect the 1000 log package, which was an upgrade to the original.)

StatCounter is an easy on-site script to install. It is simply a Javascript code snippet that you would put on the page(s) that you want to be analyzed.

2. Google Analytics. The Google Analytics package, formerly Urchin, has a large number of features that would overburden an average user. Since this is just a basic overview of the functionality, I am not going to be discussing these features in depth.

Google Analytics offers some very nice graphs, such as the executive overview graph below:

Google Analytics presents data in three columns: Marketing Optimization, Content Optimization, and E-Commerce Analysis.

• Marketing Optimization: Shows the number of visitors total and how many new visitors have discovered the site (versus repeated visitors), the number of visits and pageviews per visit, the number of conversions per visit, the loyalty score of a visitor (have they visited the page more than once? how many times?), the referring source (search engine or site), the geographic location of the visitors, the ISP of the visitors, among other bits of interesting information (including user-defined data). You can also look at how your PPC campaigns are doing (how many page impressions, how many clicks, which keywords targeted the most clicks, etc.) and get keyword recommendations based on keywords present on your site combined with your PPC campaigns.
• Content Optimization: This section of Google Analytics provides additional information on the best-converting keywords and which of these drives traffic for both paid and organic search, in addition to the source of these keywords (such as the search engine that displayed the ads). You can view which pages people landed on by filename and by title tag and you can see how many pages were visited and the length of each visit. You can also track goal conversions to see which were successful and which were abandoned and where. Further, you can get additional information about your visitors: the browsers and operating systems used, the desktop resolution and number of colors displayed, the speed of the users’ connections, and you can find out a variety of browser-specific options (Java/Javascript/Flash enabled?)
• E-Commerce Analysis: For the purposes of this blog post, this section is not being covered. However, this section allows you to see where your visitors are coming from, how they are finding you, which products are being purchased, and which marketing campaigns are contributing to these successful conversions.

In the image below, you can get an idea of how Google displays the navigation in a graphical interface. By clicking on one of the small icons next to a filename, you can see where your visitors navigated from a particular entry page:

The Google Analytics package seems inundating (and it really is), so for a free package, this is a pretty comprehensive solution. It’s also relatively easy to configure; simply placing some JavaScript code on the pages that you wish to track will bring you these results.

3. AWStats. The AWStats statistics package is a server-side solution for those who are able to access and process server logs. Typically, AWStats is installed on the same server that the website is hosted on, but as long as the AWStats can access the web server logs, this should be relatively easy to set up. After it processes the logs, you’ll be able to see a page that looks much like this:

Further breakdowns allow you to see what kinds of visitors visited at certain points in the day, the countries the users visited from, the IP address of the visitors (reverse DNS is available), the robots that accessed the site, the length of the visit, the types of files that were accessed (and how much bandwidth was consumed), the most frequently-visited pages on the site, the operating systems and browsers used by visitors, the number of hits that came from a search engine (see below), the keywords used to access the site, and the HTTP server codes.

This neat functionality allows you to see which search engines have linked to the particular site.

AWStats is a pretty analytics package, and if you have the right permissions, one of the neat things about this is that you can actually update the statistics in real time through the browser. Of all on-server free analytics package, I think that I prefer AWStats above all the others.

4. Webalizer: Like AWStats, Webalizer needs to be installed on a host that is able to access the server logs. When it processes these logs, you get a view of the types of visitors that have come to the site and what they were specifically seeking.

When you click on the month link, you can get a bigger breakdown of the pages on your site that were accessed. For example, you can see the total hits, the response codes served by particular page accesses, the peak hours of website access, the most-accessed URLs (and how often they were accessed), the entry/exit pages, the referral pages, a list of keywords used to find the site, and where your users came from.

Webalizer is a nice package and displays the output in a friendly form. It is pretty self-explanatory as are many of these packages and is recommended.

5. Analog. Analog is a very simple stats package. Results generated by Analog are not as eye-catching as any of the aforementioned four statistics packages. For example, the general summary and month view do not look to be very extensive:

It’s possible that Analog used the graphs just to emulate other statistics packages, but it really doesn’t do much for the user. However, the written statistics are rather helpful, showing users the first two octets of an incoming IP address, a redirected referrer report, a failed referrer report, the list of sites that referred traffic to the particular site, the types of browsers that visitors were using, the operating system used in the particular visit, the list and number of failed and successful status codes returned by the server for a page access, the sizes and types of files accessed, the directories accessed, and a request report of files that have been accessed 20 times or more.

Analog is an on-server statistics package, and while helpful, doesn’t provide as detailed information as AWStats or Webalizer in an as-friendly format. Most hosting providers may provide an on-site statistics package but opt out of Analog due to it lacking some more detailed features.

6. MyBlogLog: The blogging widget software shows a lot different information than any aforementioned packages, and it is also an easy installation with a simple JavaScript snippet embedded onto the pages that wish to be tracked.

MyBlogLog has a basic and premium statistics package. In the basic package, MyBlogLog shows you a list of where users came from, what users viewed, and what they clicked. You can also see what other communities your visitors prefer and, most interestingly, get an idea of what other pages people are interested in.

Here’s what pages were popular on our blog today and the time spent at each:

This is a nice community statistics package which really has some nice and interesting information that is not provided by other statistics packages. The element of community beyond statistics has been discussed before, and despite MBL being under fire recently, it’s still a very valuable package.

Overall, these are some good packages that a user looking for statistics analysis should consider. In fact, you really could use all at once, as none are mutually exclusive of any other. You could then compare results to get an idea of the accuracy of these packages (especially because on-server packages will be as precise as they can be since they’re processing logs directly from the server).

Does anyone have any other statistic package recommendations? Why? What do you think is most accurate? What would you prefer?