The world of web analytics is an exciting field that will continue to see tremendous growth with the coming years as the demand for in-depth analysis of web trafficking data increases. More companies are realizing the importance of using web analytics to their advantage. Whether the website is based on e-commerce, lead generation, or simply delivering content, all companies can use web tracking to more efficiently accomplish their specific goals.
Gathering the data is the first step, and although it is not the easiest it is indeed the most straightforward part of the analysis process. After tracking tools such as Google Analytics or Adobe Site Catalyst are implemented in the source code on your website, data will begin pouring in.
You can access the data by logging into your GA or SC account, and, within the web interface, you can generate countless reports implementing various segments and filters for more customized results. These results can be exported to file formats like Excel workbooks, CSV, and PDF files for immediate use or future reference. You can certainly use Microsoft Excel to create a multitude of charts and graphs from this data, but the real challenge comes when there is interest for deeper statistical analysis and multi-faceted dashboards.
There are a slew of different business intelligence tools out there, but each company should choose one based on their specific needs. Ease of use, display attractiveness, statistical rigor, learning curve, and price are just a few of the factors to consider when choosing your BI tool. In my experience, I have found there are two main categories of tools: those that perform all types of statistical analysis, and those that produce aesthetically pleasing displays. Unfortunately, I have yet to find a tool that accomplishes both with ease and (relatively) reasonable cost to the user.
Without a doubt, the leading software for statistical analysis is SAS (Statistical Analysis Software). SAS is moderately easy to learn and the SAS website (www.sas.com) provides excellent documentation on every procedure. What questions cannot be answered by this documentation can be found elsewhere on the web, as SAS is a widely used tool. Data can be both imported and exported to various file types with ease, making simultaneous use with other tools uncomplicated. SAS has the unique ability to generate complex data subsets and manipulations through just a few lines of simple code. The standard output is quite mechanical looking, and SAS does provide a “ODS graphics” option that makes things look a little prettier, but still there is no “WOW” factor to the graphs. The main drawback of SAS is the price tag – small companies needing only one or two licenses may find it unrealistic to spend several thousand dollars per year on software that might not be utilized to the full extent of the cost.
An alternative to SAS is R freeware which, as the name states, is completely free to all users. The user interface is much plainer than that of SAS, and sometimes the coding has more steps, but I have found that R can still accomplish almost everything that SAS can. There is also a considerable amount of documentation provided by both R (www.r-project.org) and other users on the web. The main problem with R is data exportation, especially to an Excel format. Code must be submitted to “write out” data, but only to a CSV file. You can open this file in Excel, but it takes a few more manipulative steps within Excel to break the comma-delimited files into separate columns within the spreadsheet. It is not terribly labor intensive, but the additional steps can often become tedious when working with many data sets. Additionally, R output is even more computerized-looking than that of SAS and the graphs it produces are extremely one-dimensional. The most appealing aspect of R is definitely the price, which is none. For most users, this probably outweighs the shortcomings of data exportation and display attractiveness.
Another alternative is Minitab Statistical Software. For users not familiar with or willing to learn computer programming, Minitab provides all statistical tests through a point-and-click GUI, similar to Microsoft Excel. You simply import your data and manipulate it within Minitab’s interface. A wide range of options for calculating descriptive information, test statistics, and regression analyses are available. Generating all types of plots, charts, and graphs could not be simpler, and users can easily color code, add trend lines, and change display type with the click of a button. Minitab also provides thorough web documentation and support (www.minitab.com). Minitab has just as much capability for statistical testing as SAS or R, but SAS is the most versatile and R is free, thus Minitab is my third choice.
For companies interested more in producing beautiful reports than sophisticated statistical analysis, there are numerous products that will deliver stunning, detailed, and informative dashboards and displays. To me, the most important considerations in choosing this type of BI tool are depth of detail and ease of use. I have not even begun to scratch the surface of discovering all of the many display tools that exist, but I have found one product that meets all of my expectations.
Tableau is by far the easiest and most user-friendly software of those I have tested. It is also the 2012 winner of the Digital Analytics Association award for New Technology of the Year. It allows you to drag-and-drop different dimensions and metrics to various locations within the interface, showing instantaneous results. Users can create multiple full-screen sheets allowing a high level of detail for each individual display. Several of these sheets can then be consolidated into one dashboard in which synced filters can be implemented for interactive data segmentation. Tableau provides free tutorial videos and weekly webinars for continuous education (www.tableausoftware.com). For users interested in sending automated reports to clients or interacting with dashboards over the web, investing in the pricier Server option is the solution. Otherwise, Tableau offers both Desktop and Professional versions depending on the complexity of the data source being accessed.