Oracle Big Data at Work Seminar

8-8-13
Our owner and principal, Mark Gavoor, attended an Oracle sponsored seminar in Chicago: Big Data at Work.

Big Data Analysis or Large Data Analytics is a topic and emerging discipline that has a significant amount of buzz around it. The basic principle and goal is simple. With most businesses operating on stable ERP systems, such as Oracle, they have amassed a wealth, an incredible amount of data. Beyond basic functional reporting, the goal of Big Data Analysis is to see how this data can be used to to gain deeper insights into almost any aspect of the business to drive increases in revenues and productivity. The goal is to turn the massive amounts of data into information that can help drive the the business faster and in new directions. Note that we are not talking about a mountain of data but more like a mountain range of data.

Most Big Data examples one might read in the business process involves sales. Retailers have a wealth of information due to customers using credit cards for most purchases and retailers, like CVS and Walgreens, having customer loyalty card programs. The retailers have a wealth of information on specific customer buying history. Using analytics, they can see patterns and preferences and offer promotions tailored to the individual created in their ERP and delivered via email. But, this methodology can work in all business functions especially in finance and the supply chain. In this posting, we shall restrict ourselves to sales and marketing.

There is a great case that was widely publicized a few years ago when this discipline was emerging. A big box retailer had learned through Big Data Analysis that when women switch to unfragranced personal care products, they tended to shop for maternity clothes shortly thereafter. The shopping for nursery furniture, diapers, baby accessories, and diapers followed that. So, they had an analytic routine that started sending emails touting all their maternity wear and nursery offerings when they saw a woman switch from fragranced to unfragranced personal care items. Sales in the maternity clothing and nursery department sales increased. Brilliant. This is an beautiful example of Big Data Analysis.

Continuing on with this example, a man called the manager of the local big box store. He was furious. The mans sixteen year old daughter was suddenly getting emails offers for maternity clothes and nursery furniture and accessories. He complained to the manager who offered to look into this and make it stop. The man called the store manager back a few days later and the manager immediately apologized again and said he was still waiting to hear back from corporate on this matter. The man said that he should be the one apologizing... his daughter just informed him that she was pregnant. This shows the predictive power of analytics.

In a gross simplification, there are four major steps to Big Data Analysis project.

[1] First, there must be a question posed. The question could be very specific or, if one is willing to do more exploring, the question could be more open ended. e.g. How can we increase the sales of maternity and nursery goods?

[2] The second step is to mine the data i.e. to create a smaller data set needed to answer the question by reducing the mountain range to a specific mountain or large hill. Let's look at the buying patterns of women who bought maternity clothes and nursery goods. Let's look at the previous 9 months of buying data. Even though this was a paring of all the big box retailer data, it was probably still a very large data set from a statistical analysis perspective.

[3] The third step is to do the statistical analysis. We assume the big box store did a large correlation study looking for the goods, factors, that could be used as predictors. Voila! They got the answer outlined above.

[4] Lastly, can what was learned be put into "production" as our big box retailer did? When in production, at least one question needs to be answered: did we get good results. Per our example, did maternity clothing and nursery sales increase?

More to follow on this. Probably from this conference.

Thanks Oracle.