Analytics 2.0: Big Data, Big Testing, and Big Experiences — Part 2

Stephen DeAngelis

April 4, 2013

In Part 1 of this two-part series, I discussed a Harvard Business Review article written by Wes Nichols, cofounder and the CEO of MarketShare, a global predictive-analytics company headquartered in Los Angeles. [“Advertising Analytics 2.0,” March 2013] In that article, Nichols argues that old analytic techniques are simply too limited to meet the complexities of today’s marketplace. He asserts that analytics 2.0 involves three broad activities: attribution, optimization, and allocation. I noted that marketing technologist Scott Brinker describes this process in a similar manner, stating that marketing’s future involves using big data and big testing to provide the consumer with a big experience. [“The big data bubble in marketing — but a bigger future,” Chief Marketing Technologist, 21 January 2013] As I stated in that post, only analytics 2.0 is capable of dealing with all three “big” activities.

 

For the past several months, my company, Enterra Solutions®, has been working closely with McCormick & Company to help it develop an exciting new service that will help McCormick’s customers get the most out of their food experiences. This new service is called FlavorPrint™ and it is a specific instantiation of how analytics 2.0 can be used in the food and drink industry. There is perhaps nothing more personal than the sense of taste. With each new child, parents become aware of just how personal individual tastes can be. FlavorPrint is going to permit McCormick to provide personalized recommendations concerning recipes, meals, and seasonings that fit an individual’s taste preferences. The FlavorPrint mark (see attached image) will be personalized for each consumer. Over time, the mark will change as customer’s refine their flavor preferences. By clicking on this link, you can sign up for this exciting new service. On that page (a representation of which you can see below), you will see the section that discusses FlavorPrint and the 33 flavors that help define it. Fortunately, you only have to let McCormick know what foods you and the system will automatically relate them to the 33 flavors represented in a FlavorPrint. As the sign up page states, all you have to do is tell McCormick about your favorite foods and the system will “analyze the flavors to create your unique FlavorPrint.”

 

Although the FlavorPrint is designed for an individual, if, during the signup process, you indicate your household’s preferences rather than your personal preferences, the FlavorPrint will reflect household preferences. You might want each family member to get FlavorPrinted just to see how different individual FlavorPrints are each member of the household. The signup process is easy, quick, and fun. All you need to provide is your first name, email, and select a password. You are then presented with a series of dishes or flavors that you can either like, dislike, or skip. Based on that information, a personalized FlavorPrint is generated and you immediately receive recommended recipes that you might enjoy trying. If you want to further refine your FlavorPrint, you can rate more flavors, tell McCormick about your cooking preferences, let them know what you keep in your pantry, and what cooking equipment you use to prepare meals. Trying new recipes will also be less risky if it fits the household’s FlavorPrint profile. As consumers get more comfortable using their FlavorPrints, it will be easier for them to experiment with flavors to enhance their eating experience.

 

As you can imagine, to make this happen a lot must take place behind the scenes. That’s where analytics 2.0 comes into play and where Enterra Solutions is helping McCormick ensure that FlavorPrints are accurate and continuously refined. Enterra® is also helping McCormick provide the “big experience” consumers are looking for. We are doing this using a number of “engines” driven by the Enterra Cognitive Reasoning Platform™. The Platform ingests structured and unstructured data, understands the nature of the data, learns from known and discovered relationships, and takes actions within decision cycles to obtain desired outcomes. The Platform addresses all four of Big Data’s dimensions: Volume, Velocity, Variety, and Veracity.

 

  • Volume: This dimension addresses the size of data. In today’s world, that volume is enormous and getting larger.
  • Velocity: This dimension addresses the timeliness of information. There was a time when quarterly reports were fast enough to keep up with the clock speed of businesses. Today, for truly time-sensitive data, two minutes may be too long.
  • Variety: This dimension addresses the fact that useful data can be messy. It is no longer found solely in neat rows and columns of spreadsheets. Most new data is unstructured and must be filtered and understood to be of value.
  • Veracity: This dimension addresses the trustworthiness of data. If you can’t trust the data, you can’t act confidently act upon the insights obtained from it. As the number of sources of data grows, challenges associated with the veracity of that data also increases.

 

Some analysts add to more “Vs” to that list: Visualization and Value. One of the reasons that McCormick adopted a FlavorPrint icon was to make it easily understood by consumers. If information is presented in a way that is not easily understood, it is little better than having no information at all. Value is not so much a dimension as it is an outcome of the other “Vs” discussed above. In the case of FlavorPrint, value can be found both in the information it contains and from how it can be used to enhance the eating experience. Enterra’s Cognitive Reasoning Platform, brings together two distinct philosophical and technological computing camps (i.e., Mathematical Optimization and Reasoning). This melding of methods allows the Platform to perform rapid computations as well as discover and explore new relationships.

 

Enterra’s Cognitive Reasoning Platform addresses marketplace needs for today’s computer systems to be able to Sense, Think, Act, and Learn® about the environment (industry/domain) in which they operate. Even at the level of “Small Data” it is overwhelming for a human to process, and, at the level of Big Data, it is impossible to manage. Therefore, computers need to evolve from executing tactical instructions to thinking and making sense of the data in a way more similar to humans, but at computer data processing speeds and with a computer’s ability to rapidly process numeric computations. The Platform attempts to do this by pairing an ontology and rules engine, with the muscle of dedicated analytic data processors to perform rapid computations.

 

As Wes Nichols made clear in his article about analytics 2.0, it’s easier to talk about doing it than actually being able to accomplish it. As he puts it, “The opportunity is clear, but so is the challenge.” At Enterra, we are excited about helping develop some of technologies that can rise to the challenge. We are just as excited that the first instantiation of our capabilities are being used by McCormick & Company in its innovative FlavorPrint program.