AI/IoT/Analytics

EMC World Day 1: Big Data Business Model Maturity Index and Analytic Profiles

Bill Schmarzo By Bill Schmarzo May 5, 2015

EMC-GS-May-the-4th1-300x145Day one of EMC World is over and I have survived.  Not so sure about days 2 and 3, but so far I’m hanging in there.  I did my presentation today and got a couple of thought-provoking questions afterwards.

What Phase of the Big Data Business Model Maturity Index is the hardest?

I was asked which of the phases of the Big Data Business Model Maturity Index (BDBM) is the most difficult (see Figure 1).

Image1

 

I think it’s clearly the Business Insights phase for the following reasons:

  • There is the greatest amount of organizational inertia in moving from Monitoring to Insights.  It requires both IT and the Business stakeholders to “think differently” about data (cost to be avoided or asset to be hoarded?), about analytics (descriptive analytics or predictive/prescriptive analytics?), and about the role of data and analytics in the organizational decision-making processes (HIPPO’s make the decision or do what the data tells you?).
  • It is also the phase that does not have an immediate financial benefit or ROI.  The financial benefits or ROI isn’t realized until the Business Optimization phase, but the Insights phase is key to gleaning actionable insights about your business entities that you can use to optimize key business processes or key business initiatives.

In order to overcome the inertia of Monitoring and move to the Insights phase, we guide our customers to target key business initiatives and focus on delivering financial ROI in a 9 to 12 month window.  Our Big Data Vision Workshop process is designed to help organizations identify those high-value and high-feasibility use cases.  Following this, the Proof of Value (not Proof of Concept) engagement is designed to demonstrate the ROI of big data use case and prove out the analytic lift that would advance the organization into the Optimization phase.

What comprises an Analytic Profile?

In my presentation, I tend to talk a lot about the power of Scores in delivering metrics that are potentially better predictors of performance.  Scores are important in supporting the decisions you are trying to make and the actions or outcomes you are trying to predict (see Figure 2).

Image2

Figure 2: Analytic Profile Example

However, there are other elements that comprise an Analytic Profile.  Let’s look at what Bill Schmarzo’s Analytic Profile might be from the perspective of Starbucks.

  • Demographic Information.  This is the basic information about me such as name, home address, work address, age, gender, marital status, income level, value of home, length of time in current home, education level, number of dependents, etc.
  • Behavioral Information:  Now we’re starting to get interesting, as we want to create behavioral insights that are relevant for the business initiatives that Starbucks is trying to support.  Depending upon the targeted business initiative (customer retention, customer up-sell, customer advocacy, new store locations, channel sales, etc.), here is some behavioral information that Starbucks might want to capture about me:  favorite drinks in rank order, favorite stores in rank order, most frequent time-of-day to visit a store, most frequent day-of-week to visit a store, recency of store visit, frequency of store visits in past week / month / quarter, how long do I stay at which stores (“Pass Thru” or “Linger”), etc.
  • Classifications.  Now we want to try to create some “classifications” about Bill Schmarzo life that might have impact on my key business initiatives such as:  Lifestage classification (Long marriage, kid in college, kid at home, weight/diet conscious), Lifestyle classification (heavy traveler, heavy chai tea drinker, light exerciser), Product classification (morning coffee/oatmeal consumer, afternoon frap/treat consumer), etc.
  • Rules.  We might also want to capture some rules or propensities about Bill’s usage patterns that we can use to support my business initiatives, including: propensity to buy oatmeal when he buys coffee when traveling in the morning, propensity to buy a cookie/pastry when traveling in the afternoon, propensity to buy product in the channel, etc.
  • Scores.  We also may want to create scores (as I discussed in the “Thinking Like a Data Scientist” blog series and in my Monday presentation) to support decision-making and process optimization.  Scores that we might want to create (again, depending upon the targeted business initiative) could include Advocacy Score (which measures my likelihood to recommend Starbucks and make positive comments for Starbucks on social media), Loyalty Score (which measures my likelihood to continue to visit Starbucks stores and buy Starbucks products), Product Usage Score (which is a measure of how much Starbucks product I consume – and revenue I generate – when I visit a Starbucks store), etc.

Eager to see what Day 2 brings! Follow me on twitter for more updates throughout the day @schmarzo

 

United Universe

Bill Schmarzo

About Bill Schmarzo


Read More

Share this Story
Join the Conversation

Our Team becomes stronger with every person who adds to the conversation. So please join the conversation. Comment on our posts and share!

Leave a Reply

Your email address will not be published. Required fields are marked *