In Log-Based Analytics, Money Does Talk
Open source log-based web analytics software makes it easier to make sense of web server logs, but proprietary log-based web analytics software typically takes analytics to another level.
Over the years, the line between open source or free software and proprietary (sometimes referred to as commercial or paid) software has become fuzzier and fuzzier. Still, in log-based web analytics, proprietary products usually do have an edge over open source in three areas: detail, reporting, and privacy.
In the early days, proprietary log-based web analytics software was quite popular. Over time, log-based web analytics programs have lost popularity, for reasons we’ll explore in another blog. Long established favorites, such as WebTrends, have moved away from the log-based tracking model, leaving few products in the market.
Still, they do exist. Products like Deep Log Analyzer fill a niche need for stronger web analytics products at affordable prices, yet even that has been changing lately. Price aside, the key benefits of these products tend to include:
Where most open source log-based analytics settle for restating the statistics in tables and charts, proprietary log-based analytics attempt to compete with alternative packages by “connecting the dots.” You may recall that web server logs record visitor hostname or IP address information, as well as the date and time of the requests. Proprietary products seek out and present visitor behavior from this information, offering statistics on popular paths within a site, frequency of return visits, bounce rates, and more. This additional detail is extremely valuable in learning the effectiveness of a site’s design and performance.
Bar graphs, pie charts, and statistical tables are handy, but all too often with open source products, you have to go through a lot of hoops to get to see the report you need. Sometimes, you’re forced to create your own report in a spreadsheet, using statistics found in an online log analyzer, and moving that data to the spreadsheet can be a challenge in its own right.
This is another point addressed by many proprietary web analytics products. Pre-configured reports are often made exportable into a variety of useful formats, including printed, Excel, or CSV. By being exportable, the reports can be shared with others offline, or further utilized in spreadsheets and other documents.
Privacy or security is many times sited as a major reason why website operators prefer proprietary log-based analytics. Web server logs are rarely stored in publicly accessible locations, yet many open source log-based analytics solutions are intended for server-based operation, and produce their reports on pages that are easily accessed by the public, if they know where to look. Because proprietary products are often used on standalone machines in the home or office, the reports are naturally protected from prying eyes. Site operators simply download periodic copies of the log files, typically using file transfer protocol (FTP), and import the file into the analytics program. They then tell the program to crunch the numbers; only rarely can such a program read and report on log information quickly, particularly if the log is large. Once the calculations are completed, however, a full-range of analytics information is available, as privately and securely as the desktop computer itself.
Logs have their limits
Still, with all of this extra ability, these products all share a significant limitation: they can only report on what is collected by the web server. This means you can only draw just so many conclusions from a server log. Also, questions have been raised about the accuracy of log-based analytics. We will look at these issues in our next blog.
Michael Pedone is the President / CEO of eTrafficJams.com, a search engine optimization and website marketing company <http://www.etrafficjams.com> located in Clearwater, Florida that specializes in getting targeted, eager-to-buy traffic to your site. You can catch him blogging at: <http://www.etrafficjams.com/blog/>.