Data Management as a Business Discipline – Part 2: Theorems and Principles

In the blog “Why Data Management Is Today’s Most Important Business Discipline”, I challenged the business and IT communities to reframe the conversation about data management; transform data management from an IT practice into a commercial discipline focused on leveraging data (and analytics) to deliver business and operational services results.

If the data is “the most precious resource in the world(thanks to “The Economist”), then corporate leadership must crop their expectations of data management as a business discipline to guide the organization in leveraging data to drive and generate new customers, sources, products, services, and business value.

But what is a business discipline and what do we need to create to realize this vision of data management as a business discipline?

A Discipline consists of systematic research, observation, measurement and experimentation resulting in the assimilation of learning into laws, theorems, concepts, principles, practices, frameworks and formulas to enable consistent application and continuous improvement from application real of this discipline.

According to the article “6 attributes of an academic discipline» written by Joshua Ki, a Academic discipline consists of the following features:

1 – Search object. “The disciplines have a specificity research object (e.g. finance, economics, mathematics, computer science), even if the subject of the research may be shared with another discipline.

2 – Accumulated specialist knowledge. “The disciplines have a body of accumulated specialist knowledge referring to their research object, which is specific to them and generally not shared with another discipline”.

3 – Theories and concepts. “The disciplines have theories and concepts who can effectively organize accumulated specialist knowledge.

4 – Terminology. “Disciplines use terminologies or a specific technical language adapted to their research object.

5 – Research methods. “Disciplines have developed research methods according to their specific research needs.

6 – Institutional event. “Disciplines must have institutional event in the form of subjects taught in universities or colleges, respective academic departments and related professional associations.

I think Data Management ticks each of these boxes by a Academic discipline perspective. And I will add #7 of a Commercial Discipline perspective:

7 – “The disciplines have documented best practices, learnings and stories associated with the successful and unsuccessful application of this discipline in real world situations to create a measurable and sustainable environment business and operational value.”

For a commercial discipline to be relevant and ensure the commitment of business and IT management, this discipline must create value. And how can we guide these value creation conversations?

Yes, we can create a series of play cards that summarize the value creation theories and concepts (Discipline Requirement #2) that our data management business discipline must adopt.

Also, think how fun it would be to pull out these playing cards in meetings with your company’s stakeholders or your data and analytics teams to remind yourself that these data management efforts should be focused on valuable creation.

Let’s review these maps and start imagining how you could use them to create a value-driven culture (through data and analytics)!

Figure 1: The Royal Court of Value Creation Driven by Data and Analytics

  • Value Engineering Framework breaks down the organization’s strategic business initiative into its supporting business components (stakeholders, use cases, decisions, key performance indicators) and data and analytics requirements.
  • Nanoeconomics is an economic theory of behavioral and performance propensities predicted by an individualized entity (human or peripheral).
  • Analytical profiles continuously codify, share, reuse, and refine the predicted propensities, patterns, trends, and relationships for the organization’s key human and peripheral assets or entities.
  • Use cases are a group of decisions around a common KPI or metric in support of specific business initiatives that have quantifiable business or operational value.

Figure 2: The lieutenants of data-driven and analytics-driven value creation

  • Economic multiplier effect of data, based on the economic multiplier effect, formulates the calculation of the accumulated attributable value of a dataset from the reuse of that same dataset in multiple use cases.
  • Learning economies measure the effectiveness of an organization’s value creation from continuous learning and adaptation to continuous changes in the business and market environment and ecosystem.
  • Schmarzo’s Economic Digital Asset Valuation Theorem highlights three effects that result from the sharing, reuse, and continuous improvement of an organization’s data and analytical assets.
  • The Think Like a Data Scientist methodology is collaborative ideation, a value-centered, human-empowered “scientific method” that seeks to unleash the predictive intuition of subject matter experts.
Slide 3-2

Figure 3: Footsoldiers of data-driven and analytics-driven value creation

  • Prioritization matrix facilitates collaboration between business and data science stakeholders to identify use cases that have both significant business value and reasonable feasibility of successful implementation
  • Canvas for developing hypotheses captures data science requirements, including business goals, KPIs and metrics to measure success, key stakeholder decisions, potential ML features, and costs of false positives and false negatives
  • AI utility function evaluates multiple variables and KPIs across multiple dimensions to guide actions or decisions the AI ​​model can take or take to achieve desired goals.
  • Key Performance Indicators (KPIs) are quantifiable measures used to assess an organization’s progress and ultimate success in achieving its business or operational objectives and goals.

Figure 4: The Wildcards of Data and Analytics Driven Value Creation

  • Features are the mathematically transformed variables that AI/ML models use during training and inference to make predictions and guide actions or decisions
  • Analytical scores are the results of analytical models factoring several variables and characteristics at predict the likelihood of a behavior or action expressed as a single number. For example, a credit score predicts the likelihood that a person will repay their loan.

I hope this article has provided both fun and a dose of creativity on how your organization could leverage value creation playing cards to start reframing data management conversations and expectations. – that data management is not just an IT practice, but is becoming the most important business discipline in the 21st century.

Now I just need to find an affordable way to turn them into real (affordable) playing cards that I can hand out at my workshops (that would be very cool!!).

Ramon J. Espinoza