Home » Business Topics » Data Strategist

Critical Role of Analytic Profiles in Developing Data Products

  • Bill Schmarzo 
Side view of information security analyst looking at charts near
Side view of information security analyst looking at charts near graphs on computer monitors in office

The tech industry is abuzz with hyped up pontifications and bold predictions of the business-changing potential of Data Products.  I could not be happier as it’s a topic I have explored in several blogs (see the end of this blog for a list of my blogs on Data Products…yea, I know, get a life).

A recent article from the Harvard Business Review titled “A Better Way to Put Your Data to Work” highlights the potential business benefits of Data Products including:

  • New business use cases can be delivered as much as 90 percent faster.
  • Total cost of ownership, including technology, development, and maintenance costs, can decline by 30 percent.
  • The risk and data-governance burden can be reduced.

The article lays out the “Data Product Approach” for building data products (Figure 1).

Slide1-1

Figure 1: Data Product Approach According to HBR

Nice graphic Figure 1, but Figure 1 is wrong.

There are two important concepts that Figure 1 gets wrong:

  • What Figure 1 calls “Data Products” are really Analytic Profiles.  Analytic Profiles are key-value stores that capture, codify, and share the individualized human and device predicted behavioral and performance propensities (analytic scores) that are used to optimize the organization’s business and operational use cases.
  • And the Use Cases on the far right of Figure 1 can be developed and delivered as Data Products, where Data Products are a category of domain-infused, AI/ML-powered apps designed to help non-technical users manage data-intensive operations to achieve specific, meaningful, relevant business and operational outcomes.

Figure 2 provides my updated “Data Product Approach” roadmap.

Slide2-1

Figure 2: Corrected Data Product Approach According to Bill Schmarzo

Let me clarify the critical interplay between Analytic Profiles and Data Products further, and how this interplay is key in helping organizations to unleash the business value of your data.

What are Analytic Profiles?

More than centralizing the capture all the historical operational and engagement data about your organization’s key entities (e.g., 360-degree views of customers, suppliers, products, and employees), organizations need to discover, codify, and share the actionable insights (predicted behavioral and performance propensities of what these entities are likely to do next) to drive more accurate, more relevant, higher precision Business Outcomes.

Analytic Profiles is a key-value data model for capturing, codifying, sharing, and continuously refining the individualized human and device predicted behavioral and performance propensities. These individualized analytic insights can include analytic scores (like credit scores), predictive indicators, segments, features, and business rules that codify the behaviors, preferences, propensities, inclinations, tendencies, interests, associations, and affiliations of the organization’s key human or device business entities (Figure 3).

Slide3-1

Figure 3: Analytic Profiles Capture Key Business Entity Analytic Insights or Predictive Propensities

We leverage the Analytic Scores captured in the Analytics Profiles to power Precision Decisions to optimize the organization’s key business and operational Use Cases (Figure 4).

Slide4

Figure 4: Use Case Optimization via Precision Decisions drive from Analytic Profiles

These Analytic Profiles provide the foundation for the creation, management, scaling, and continuous learning and refinement of your Data Products.

What are Data Products?

Data Products are a category of domain-infused, AI-powered apps designed to help non-technical users manage data-intensive operations to achieve specific business outcomes.  Data apps use AI to mine a diverse set of customer and operational data, identify patterns, trends, and relationships buried in the data, and make timely predictions and recommendations. Data apps track the effectiveness of those recommendations to continuously refine AI model effectiveness.

Data Products are engineered to deliver well-defined business or operational outcomes.  For example, Figure 5 shows an orchestrated series of Data Products (Prospect Profiling, Prospect Targeting, Customer Acquisition, Customer Cross/Up-sell, Customer Retention) that support an organization’s customer acquisition and retention value chain processes.

Slide5

Figure 5: Customer Acquisition and Retention Value Creation Processes

In Figure 5, Data Products are underpinned by Analytic Profiles (e.g., Customer, Product, Service, Store, and Campaign Analytic Profiles) that capture, re-use, and continuously refine the individualized predicted behavioral and performance propensities (analytic scores such as Customer Potential Lifetime Value Score and Customer Credit Risk Score in Figure 5) that are used to optimize the organization’s Customer Acquisition and Retention Value Chain process.

Summary: Importance of Analytics Profiles and Data Products

Like a great pitcher – catcher battery in baseball, Analytic Profiles are the key enabler for delivering more relevant, more timely, more personalized Data Products. Analytic Profiles and Data Products work together to continuously learn and adapt (leveraging AI / ML) to refine the Precision Decisions that optimize the Data Product’s business and operational outcomes at scale (Figure 6).

Slide6

Figure 6: Collaborative learning process between Analytic Profiles and Data Products

Finally, the integration of AI / ML into the Analytic Profiles-to-Data Products lifecycle can accelerate the (1) application and (2) subsequent capture of new customer, product, service, and operational insights yielding a semi-autonomous Data Product management process that can scale and adapt with minimal human intervention.

Extra-credit Homework!

For extra credit homework, here is the list of blogs that I have written on Data Products.  There might be a test at the end, so take notes!

Tags:
Tags:

Related Content