Anatomy is the study of the structure and internal workings of an entity.
There are likely many ways that organizations can determine the value of their data. But I’m lazy, so I only teach and employ one that I know works – a use case approach that attributes the value of the organization’s data in the application of that data to optimize the organization’s key business and operational use cases.
One cannot determine the value of one’s data in isolation of the business.
There are multiple benefits from the use case approach to determining and exploiting the (economic) value of the organization’s data:
- The use case approach establishes value – not technology, not data, not analytics – as the common linkage between the business and data analytic stakeholders.
- The use case approach drives the organizational collaboration and alignment in identifying, validating, valuing, and prioritizing the sources of organizational value creation – the business and operational use cases.
- The use case approach allows organizations to embrace a “minimum viable model” approach to developing and iteratively improving the business and operational outcomes while the organization incrementally builds out the organization’s data and analytic capabilities and assets.
- The use case approach activates the Schmarzo Economic Digital Asset Valuation Theorem, that highlights the three economic benefits associated with the economics of data and analytic assets (Figure 1).
So, the Use Case Approach in driving business and data analytics stakeholder alignment and collaboration around value depends upon one wee consideration – how does one define a use case? Let’s see if we can use this blog to clarify the definition of a use case.
What is a Use Case?
Use Case is cluster or aggregation of Decisions around a common KPI (Key Performance Indicator) in support of targeted Business Initiative or Business Challenge
Decisions are the heart of a use case, but its decisions in context of the targeted business initiative or business challenge. Without the context, then it’s very hard (near impossible) to optimize, automate, and create semi-autonomous decision-making.
Decisions are a powerful value enabler because decisions are (Figure 2):
- Easily Identifiable: Every business or operational stakeholder knows what decisions they are trying to make because they have been trying to make those decisions for years, if not decades.
- Actionable. Unlike a question (which is a valuable tool for exploration and clarification), a decision infers an action (including a decision not to act).
- Source of Attributable Value. Organizations can determine the quantifiable and attributable value from making improved business and operational decisions.
- Optimizable with Data Science. Data Science teams specialize in applying analytics to the organization’s data to optimize decisions (that’s kind of their thang).
- Collaborative. Decisions drive collaboration between the Business and Data Science stakeholders to identify the variables and metrics (features) that might be better predictors of performance and behaviors.
Figure 2: Power of Use Cases = Decisions + KPIs
A key part of the “Thinking Like a Data Scientist” methodology that I teach at universities (and am teach this spring at Iowa State University) and in client workshops is embracing the diverse perspectives of the different stakeholders around the decisions that need to be made in support of the organization’s targeted business initiative or business opportunity.
Stakeholders arethe personnel or organizations who either impact or are impacted by the targeted business initiative or business challenge.
From the perspective of each key stakeholder, we want to:
- Brainstorm the different Decisions that each stakeholder needs to make in support of the targeted business initiative or business challenge (think flip charts, post-it notes, and too many donuts in the same room)
- Identify the Desired Outcomes from those decisions; that it, what does good or success look like from the perspective of the different stakeholders, and
- Validate the KPIs and Metrics against which the different stakeholders will measure the effectiveness of their decisions.
It is from these diverse and sometime conflicting stakeholders’ perspectives that we get a thorough and holistic scoping of a use case (Figure 3).
Figure 3: Anatomy of a Use Case
A broad and diverse range of KPIs and metrics will be critical when we establish the AI Utility Function that the AI models will use as they seek to continuously optimize these use cases (Figure 4).
Figure 4: Role of the AI Utility Function
The AI Utility Function assigns values to certain actions that the AI model or agent can take. An AI agent’s preferences over possible outcomes can be captured by a function that maps these outcomes to a utility value; the higher the number the more that agent likes that outcome.
Thoroughly defining the KPIs and metrics that comprise the AI Utility Function is the difference between AI models that destroys humanity like Terminators versus models that guide thoughtful decision making like Yoda.
Updated Use Case Design Template
We document all this work within the Use Case design template. I’ve updated the Use Case Design Template that is part of my Value Engineering (Thinking Like a Data Scientist) framework. I’ve added more content to the design template including key stakeholders and their desired outcomes, overall use case ideal outcomes, privacy and governance considerations, and the ramifications of use case failure (Figure 5).
Figure 5: Use Case Design Template from Schmarzo’s Value Engineering Framework
By the way, I’d love for folks to test this baby out. And if you are interested in testing this design template, please send me a message and I’ll send you a PowerPoint version of the template.
Yea, as always, #BetterTogether.