Our most recent research stream focuses on the topic of organizational learning. The widely used conception of organizational learning frames it as the process of acquiring, creating, integrating and distributing of information. Clearly visible in that characterization is that the ultimate purpose of organizational learning is to enhance the informational efficacy of managerial decision-making; somewhat less obvious is the implicit assumption that learning is an inherently human endeavour. The latter, however, is gradually being called into question in view of the proliferation of self-learning algorithmic decision engines, commonly known as artificial intelligence (AI). Widely used to automate numerous routine business decisions, such as pricing of automotive insurance policies, eligibility approval for social or health services, or online product recommendations, AI decision engines are capable of independently learning from data in a manner that fits the established conception of organizational learning. And given the already substantial and still rapidly expanding role and value of those technologies to organizational functioning and competitiveness, it is important to expressly include technology-based knowledge creation in the definition of organizational learning.
Organizational Learning in the Age of Data (article)
In information-driven economy, few organizational competencies are as important as the capability to systematically capture, synthesize and disseminate throughout the organization competitively advantageous decision-guiding knowledge. Traditionally, organizational learning has been viewed as a human-centric endeavour, but the rise of big data and advanced data analytic technologies are compelling a fundamental reconceptualization of the scope and modalities of organizational learning. Starting with a high-level overview of the genesis and the current conception of organizational learning capabilities, this research offers a revised and expanded conceptualization of that important organizational competency. Building on the foundation of explicit differentiation between episodic vs. ongoing learning inputs and new vs. cumulative learning outcomes, a new typology of organizational learning modalities is proposed. The new typology of organizational learning explicitly distinguishes between human reason-centric theoretical and experiential learning, and technology-centric computational and simulation learning modalities. By explicitly encompassing artificial intelligence, machine learning and other manifestations of technology-based learning, the proposed organizational learning typology offers a more comprehensive and timely framing of organizational learning. By expressly acknowledging the distinctiveness of human reason- and technology-based learning modalities, business organizations will be able to develop more robust and effective systems and mechanisms to support their goal of remaining competitive in knowledge-based economy.
Evidence-Based Organizational Learning (article)
The growing importance of data-driven decision-making to organizational competitiveness poses a number of vexing organizational learning related questions. In order for organizations to develop methodologically-sound and informationally-complete organizational learning capabilities that go beyond mere accumulation, storage and cataloguing of data and other informational assets, more explicit information amalgamation and synthesis focused frameworks are needed. The Empirical & Experiential Evidence (3E) framework outlined here is built around the idea that since the primary utility of organizational knowledge is to support organizational decision-making, topically-related data and information can be considered decision-guiding evidence. The framework’s evidence synthesis logic parallels the general 6-step process of identifying, assessing, aggregating, weighing, agglomerating and incorporating distinct but related information, but that process is nested within a 3-tier evidence classificatory schema which categorizes all available decision inputs into two broad meta-categories, four more narrowly scoped categories, and twelve even more operationally meaningful sub-categories. The 3E framework also puts forth specific evidentiary insight extraction methodologies that reflect the informational uniqueness of the two meta-categories and four categories.
The Typology of Analytic Knowledge (research in progress)
The essence of developing sound data analytic competency is encapsulated by thoughtful identification of pertinent elements of knowledge, coupled with systematic accumulation of explicit and experiential analytic know-how. Identification of competency-building elements of knowledge requires a sound typology of knowledge - the one developed through our research is graphically summarized here. Our proposed typoology is encapsulated in the three-tier structure of meta-domain, defined as epistemologically-framed broad dimension of data analytic knowledge, domain, conceptualized as practice-framed aspects of data analytic know-how, and knowledge-competency units, or KCUs, defined as distinct units of data analytic knowledge. At the most aggregate, meta-domain level we see three distinct, and broad, dimensions of data analytic knowledge: 1. 'making sense of data', which encompasses the totality of data management, curation, and analytic preparation steps; 2. 'estimation & inference', which encompasses the core elements of statistical parameterization and sensemaking; and 3. 'learning with data', which encompasses the basic elements of statistical, computational and visual analytics.
Developing Analytic Competencies (research in progress)
Whether it is called 'business analytics', 'business intelligence' or 'data science', the broadly framed data analytic skill set draws from multiple disciplines, including statistics and computer science, which necessitates careful and explicit delineation of specific skills and competencies. Building on our typology of data analytic knowledge (outlined above), our research further delves into the process of analytic competency building by considering the progressive and cumulative nature of systematic data analytical process learning built around the three meta-domains of making sense of data, estimation & inference, and learning with data. Though each of the three conceptually and operationally broad meta-domains is comprised of two or more narrower domains of knowledge, each of which is in turn made up of two or more distinct sets of data analytic competencies, as graphically depicted here (right), the acquisition of meta-domains-expressed dimensions of data analytic knowledge can, and in fact should, be seen as being progressively cumulative.
Competency-Based Analytic Learning (research in progress)
Starting with the nearly axiomatic premise that learning efficacy varies across individuals, the self-paced, competency-based learning model seems particularly well suited to teaching data analytics. The critical, and not always expressly acknowledged component of data analytics focused education is the interplay between theoretical and applied knowledge, and yet, strong understanding of the 'what' and 'why' of analytics is essential to developing robust 'how-to' competencies. Moreover, the importance of the conceptual foundation + experiential competency model applies to the entire spectrum of data analytical training, starting with computational data manipulation and feature engineering, carrying into statistical estimation and inference, and ultimately into statistical and machine data analytic approaches and applications.