Frank Fillmore, president and founder of The Fillmore Group, sat to speak with Wikibon’s Jeff Kelly on theCube to talk about Big Data and data warehousing.

Fillmore noted that unlike other trends in the past, Big Data did not start in the IT world. Instead it started out in the world at large and grew organically to be an IT trend or focus. He felt that transaction processing has matured, and companies are looking for new opportunities to engage their customers at a more intimate level. Companies do not want to wait for a customer to approach them and ask what the company can do for them. They want to approach the customers with catered suggestions and options.

In regards to the presence of traditional data warehouses, Fillmore does not think they will be disappearing anytime soon. There is still a place for it with transactional data that is already established in so many enterprises. Big Data is taking that same data, but puts it in a form where people can use it to make business decisions. He says that we have great control of meta data (data about data) currently. So we are now determining the meta data by the data content itself in real-time, inferring what the data means by its context. Before Big Data we imposed context on data based on its source.

The new challenges of volume, velocity, and veracity that come with Big Data call for the use of different models to handle the data. Fillmore says that is why things like Hadoop and MapReduce have become so popular. The success of Big Data initiatives has allowed enterprises to take advantage of “here and now opportunities” that before they were not able to because traditional data warehouses were too slow to react.

When asked about vendor offerings for Big Data and how they deploy in an enterprise, Fillmore explained that phase one of the Big Data solutions involves an enterprise first testing out value offers that do not have a high cost. They would install new software on existing or old hardware to get an immediate return of answers and insights so that they can see how it would be beneficial. From this experience they will learn the difficulty of maintaining and deploying the solution on a larger scale. That is when phase two comes and they approach an IBM or other vendor for a full enterprise deployment solution that has the needed support.

Fillmore’s final advice to CIOs was to start small when considering Big Data solutions. He stated that it is necessary to reduce the time to value so that they can show quicker results. This will show the value of the investment and allow for further expansion and deployment. See the entire interview below.