by Chris Matty, CEO of Versium
Both marketing and statistics are age-old industries with a myriad of scholarly works and entire century-old industries using these disciplines as their foundations. Think old school. Casinos and the profitability derived from predictive stats are probably the best example of this marriage, yet this bond expands across real estate, banking, energy and politics to name a few.
On the surface, you might think the two have run their mutual course with little to no opportunity for innovation, however, a relatively new entrant now threatens to transform this bountiful relationship into a menage a trois! Recent advances in computer technology and processing paradigms have created the new age of ‘Big Data’ and newfound efficiencies that come with it.
A new gold rush is upon every enterprise to find the forge that will enable them to begin mining shiny new brilliances of insight. Several new tools and technologies are at the forefront with the promise to help harvest this value.
Unfortunately, many of these tools require a substantial understanding and time investment – and that’s only if you remember your college statistics or econ classes. On top of that, the “Four V’s” (Volume, Velocity, Variety, and Veracity) also mean you must be comfortable with unusual programming paradigms more unfriendly than a graphing calculator.
There are resources available, but the few (i.e. data scientists) who are capable of working with these tools are expensive and very hard to find, given the industry explosion. And that’s where Big Data takes many enterprises today: Big Expenses and Big Headaches.
While the Borg-like machine learning algorithms begin to gnaw on the seat of the CMO’s office, the first defensive questions surface including: “How many $150k+/year data scientists will be needed? What will my hardware / software investment run for my Hadoop cluster to log and analyze this data?
Should I put it all on the cloud? Should I outsource? What can I really do with all this data?” But is any of this really addressing the needs of the business, or it is just part of the journey through the big expense hills?
Within the big expense hills lies another insidious creature: the big complexifier. These creatures insist that every problem the enterprise has can and should be solved by some magical mathematical algorithm. Nate Silver’s “The Signal and the Noise” discusses some of the shortcomings of big data and statistics – financial ratings and the 2007 financial collapse for example.
The truth is many enterprises may not actually need to bury themselves within the Big Data mountain. Instead, they probably only need to meet a “pair of the V’s,” as mentioned above. And for a large class of these organizations, the real signal they ought to focus on is their customer and that’s it. Unfortunately, they only know who their customer is based on how that customer interacts with their enterprise. But these are real people, and they interact in the real world outside of the enterprise all the time.
Massive volumes of data is created by consumers every day as they interact in the real world. They buy products, subscribe to magazines, talk to their friends, visit websites, donate to causes and support political views. This type of real-life data is tremendously powerful in predicting behavior. People with kids behave and act far different from those who don’t.
Those who drive a hybrid have buying patterns much different from those who drive sports cars. A person or family who just purchased a house instantly has needs and desires different from when they were renting.
In today’s era of social media, the platforms that enable effective social networking and communication with friends generate massive volumes of interest and social behavioral data. Analyzing and understanding all of these types of data points or consumer attributes can help an enterprise to truly understand their customer. This type of understanding and insight can help with more granular segmentation, optimized marketing communication and uncovering avenues to maximize profit and ROI.
So here we are again back to the Big Data and big expense problem. How do companies derive benefit from this type of real life data? The good news is that the age old statistical models can be applied to these new data sets to generate powerful and more accurate predictive models. Certainly a wireless carrier might gain sound insight by analyzing trends in customer service complaints – possibly predict customers who might be ready to cancel (i.e. Churn). But what happens when you look at this same enterprise data after it has been combined with real-life data? Well, something very powerful.
The ability to predict consumer behavior gets far more accurate. The mind can understand that a person with kids is much different from one who does not. Think about a computer taking over hundreds of data points and combing these points (signals) to analyze such attributes like who’s more likely to cancel a service. What if a company could predict who is or isn’t likely to respond to a discount offer? Why provide a discount to those who would buy at full price anyway? It becomes possible to know who is most socially active and more importantly who are advocates versus detractors so that social campaigns can be most effective.
The analysis of this real-life data can be used to predict everything from energy consumption, to who will pay their bills on time to what type of cross-selling opportunities will generate the greatest conversion. Relatively small volumes of historical enterprise data are required to train these powerful models that utilize real life data.
As the era of Big Data continues to evolve, these types of innovative approaches to processing new data sets will be used to drive material value in enterprises based upon specific business cases that are intended to drive profit maximization. Those who learn to efficiently navigate the Big Data hills will be far ahead of the game and able to do so while effectively mitigating the big expenses and big complexities, and certainly new tools will arrive that promise to help.
Chris Matty, CEO of Versium, has been involved with numerous early stage technology ventures, in the Internet, mobile and SaaS industries where he has held executive positions. He was a founding executive of InfoSpace; GM of the Merchant Services Division, and SVP of Business Development and Sales – and helped grow revenue to $200M+. He has held executive positions at numerous start-ups including: BigTip, DepotPoint, PayScale, SmartWorks.
Visit Versium at http://versium.com/.
