Many search marketers today rely on analytics packages like Onminutre or Google Analytics to run reports and analyze data. These applications, especially Ominture, are great for real deep dive analysis of how users get to a website and the million different things they do while visiting. More times than not, it seems to me that if you consider the basic functions of analytics tools specially packaged for online marketing, these tools are really nothing more than basic statistical tools that have been in use across various industries since the 1950′s. The part that I think is missing is the fact that search marketers know nothing about basic statistical tools or how to use them and rely too heavily on the automated report generation and analysis of online marketing analytics software.
The basic statistical tools I would suggest becoming familiar with are tools like Control Charts, Pareto Charts and Process Capability for measuring performance of a Pay-Per-Click (PPC) campaign and ANOVA, Regression Analysis and Design of Experiments (DOE) for the analysis of the PPC campaigns that were previously measured. These 6 tools are the core tools that support the measurement/analysis phase of a continuous PPC cycle. They allow you to understand how well an ad group or a campaign or an account is performing, identify what works and what needs to be improved and insight as to how to make improvements when necessary. I also find that at times running PPC data through these types of reports can also provide more detailed information about performance than a similar report from an online marketing analytics software report.
The additional benefit to becoming familiar with knowing how basic statistical tools like the ones mentioned above work, you can then begin to understand and speak the language of other industries that you may find yourself working with. Aligning internal and external online marketing processes, reporting and analysis with major retailers, financial, pharmaceutical, and service industries that have quality programs in place and are typically trained to use these tools is an important measure for success.
What’s the Best Statistics Software to Use?
Glad you asked. The most widely used software for statistical analysis is Minitab. It is also typically the most expensive. In recent years other software applications such as SigmaXL has chipped away at Minitab’s market share by offering a similar package at a lower cost. These are both fine applications to use if you have the budget, but they also lean towards the automated, less hands-on feel that Omniture provides which semi-defeats the purpose of this article.
The software that I have found that has a more hands-on approach is the Statistical Process Control (SPC) software by QI Macros. The ‘software’ is actually an ‘all-in-one’ statistical toolkit that integrates into Excel. Not only is does it conveniently plug into Excel and provide you with a solid learning foundation of how these basic (and more complex) SPC tools mentioned above work, it is considerably less expensive compared to Minitab. In addition to the measurement and analysis tools, QI Marcos SPC software also includes planning and control tools that can be incorporated to support continuous online marketing processes.
I have already had some success in combining basic statistical process tools into my reporting as an ‘added bonus’ along with the Omniture reports. With this success and to better align myself with future client expectations I plan to continue integrating more basic stats tools into the core Search Quality Management (SQM) methodology I am developing. It wont happen overnight but eventually, Pareto Charts and ANOVA’s and several other quality statistical tools will become requirements to online marketing and not just something extra from another industry.