Mind Before You Mine
I am re-branding the site to reflect a major theme of the postings. Before you can mine your data you have to mind your data. Like my mom would say “Mind your manners” when caught eating a pickle during Aunt Ethel’s long dinner blessing. I say “Now mind your data” as though your business depended upon it. Enough of the checkout already. Let’s get serious about data driven decisions. Most of the posts stress the fundamentals of web analytics in collecting, processing and reporting quality data that meets business needs.
A data minder, I like to think of myself as a data wrangler, determines what data is needed to support a business decision then goes out and instruments the web site or business systems to collect the data that is needed and develop actionable reports. Why show reports if you cannot do act on the results? A data miner, once the data has been collected, can apply in depth analysis and algorithms to drill down to extract trends and correlations in the data that lead to insights. Sometimes miners get cocky and believe they can extract gems of insight from swamp muck of data.
I can and have done both but find in my consulting that I am continually returning to the fundamentals to ensure that appropriate data policies and governance are in place to insure a sustainable process for collecting data and reporting results in a form that supports timely business decisions. One might be able to perform feats of magic with data that already exists but was collected for a different purpose to generate a pearl of surprising insight. But to do that magic on a daily basis in a timely manner requires a great deal of data minding. Once there is quality data that is reliable and accurate, then data mining, which often includes statistical modeling or non-linear mapping, becomes much more effective.
I am Timothy Kraft. As an Astrophysicist, who has spent most his career as a technologist, I often have a unique take on things. The discipline of astrophysics requires bringing together disparate disciplines to build a unified picture from very little data. I have spent the last 12 years applying this perspective to web analytics and business intelligence.
Often job descriptions for web analysis consultants stress a combination of disciplines including business, marketing, user experience, IT, and web design as well as knowledge of a growing number of domains including advertising, search, ecommerce, social media, mobile, on top of traditional media. All this in addition to having deep data analytic skills and experience in statistics, economics, and lately psychometrics.
Working at this nexus of competing needs and views is a natural operating environment for me and a necessary aspect of analytics. Not that I am a expert in all these, but as a knowledge engineer, before becoming a web analytic developer, I am able to work with domain experts in all these disciplines and bring them together in a unified view and integrated plan of action. I also have Virgo rising so love getting into weeds and doing log audits, tagging sites, developing scripts, or building models.
I have had the opportunity and privilege of participating in developing Web Analytics as discipline at critical turning points in its maturing – when web analytics went from counting server side hits to observing visitor behavior in the client side DOM; from first reporting conversions to using them later to optimize ROI leading to CPA (Cost Per Acquisition) based SEM; from analytics being a cost center to developing a P&L Statement for data that modeled the direct revenue benefit of having the data against cost of collecting and analyzing the data — justifying, with Ops Finance approval, expenditures of 10s of millions of dollars in data acquisition and processing.
Now we see the emergence of Voice of Customer and Customer Experience Management as ancillary to the Web Analytics that is offered by the major vendors requiring more planning, more coordination and more commitment of resources. This has lead to the discussion of maturity models that attempt to out line what the natural progression of capability should be or could be or must be. This has focused the discussion to ecosystems for web analytics and business intelligence. It is at this point of synthesis that I believe I have some helpful experiences and points to contribute to the analytic community.