Business Environment and Decision Making: Foundations of Decision Support Systems and Business Intelligence

Business Pressures, Responses, and Decision Support Systems

• Core Learning Objectives:

  • Understand the dynamics of today's turbulent business environment and examine how organizations survive and excel by solving problems and exploiting emerging opportunities.

  • Recognize the necessity of computerized systems to support managerial decision making.

  • Examine early classical frameworks that categorize and define managerial decision making.

  • Describe Business Intelligence (BI) methodologies and concepts, detailing their historical and structural relationship to Decision Support Systems (DSS).

• The Business Pressures–Responses–Support Model:

  • Organizations operate within an increasingly complex and turbulent competitive environment, necessitating a shift toward computerized operational and strategic support.

  • Business Pressures: Stem from modern market conditions, consumer dynamics, technological disruption, and societal shifts.

  • Organizational Responses: Actions taken by enterprises to counter environmental pressures and seize opportunities, including strategic planning, agility, collaboration, and process optimization.

  • Decisions and Support: Computerized tools, analytics, and business intelligence frameworks that facilitate integrated analyses, predictive modeling, and rapid managerial decision making.

• Categorization of Business Environmental Factors:

  • Markets:

  • Intense competition in domestic and global arenas.

  • Rapid expansion of global markets.

  • Proliferation of electronic markets blooming across the Internet.

  • Emergence of innovative marketing methods.

  • Expanding opportunities for business process and IT outsourcing.

  • Increasing imperative for real-time, on-demand transactions.

  • Consumer Demand:

  • Escalating consumer desire for customization and personalized products/services.

  • High expectations for superior quality, product diversity, and rapid delivery speed.

  • Shift in customer empowerment, resulting in more demanding and less loyal customers.

  • Technology:

  • Accelerating rates of innovation, yielding rapid introductions of new products and services.

  • Shortened product lifecycles and an increasing obsolescence rate.

  • Growing information overload experienced by executives and operational personnel.

  • Rapid emergence and penetration of social networking, Web 2.0, and advanced collaborative technologies.

  • Societal Factors:

  • Proliferation of government regulations alongside targeted industry deregulation.

  • Demographic shifts: the workforce is increasingly diversified, aging, and composed of a higher proportion of women.

  • Critical focus on homeland security and heightened risks associated with terrorist threats.

  • Compliance mandates stemming from reporting-related legislation, including the Sarbanes-Oxley Act.

  • Heightened public expectations regarding corporate social responsibility.

  • Widespread emphasis on environmental sustainability and ecological impact.

• Organizational Action Modes:

  • Organizations must exhibit multiple behavioral postures: Reactive, Anticipative, Adaptive, and Proactive.

  • Key executive actions include:

  • Employing formal strategic planning mechanisms.

  • Designing and deploying innovative business models.

  • Restructuring and re-engineering internal business processes.

  • Establishing strategic partnerships and business alliances.

  • Upgrading and integrating enterprise corporate information systems.

• Closing the Strategy Gap:

  • A primary objective of computerized decision support is closing the strategy gap.

  • The strategy gap represents the difference between an organization's current performance and its desired performance (as formally defined by its mission, strategic objectives, performance goals, and operational strategies).

Managerial Decision-Making Frameworks and Foundations

• The Nature of Management and Decision Making:

  • Management is defined as a systematic process by which organizational goals are achieved through the coordinated allocation of resources.

  • Inputs: Organizational resources (human capital, financial assets, raw materials, data).

  • Outputs: Attainment of strategic and operational goals.

  • Quantitative Measure of Success:     $\text{Success} = \frac{\text{Outputs}}{\text{Inputs}}$

  • Management is functionally equivalent to decision making ($\text{Management} \equiv \text{Decision Making}$).

  • Decision making represents the deliberate process of evaluating and selecting the best solution from two or more viable alternatives.

• Mintzberg's 10 Managerial Roles:

  • Interpersonal Roles:

  • Figurehead: Performing ceremonial and symbolic duties as the legal or social representative of the enterprise.

  • Leader: Guiding, motivating, and supervising subordinates to align with enterprise goals.

  • Liaison: Establishing and sustaining external contacts and information networks.

  • Informational Roles:

  • Monitor: Continuously scanning the internal and external environment to gather relevant operational and strategic intelligence.

  • Disseminator: Transmitting vital information gathered externally or internally to subordinates and peers.

  • Spokesperson: Communicating official organizational positions and performance to external stakeholders.

  • Decisional Roles:

  • Entrepreneur: Initiating change, identifying opportunities, and launching innovative projects.

  • Disturbance Handler: Managing unexpected operational disruptions, crises, and conflicts.

  • Resource Allocator: Determining the distribution and scheduling of personnel, funding, equipment, and organizational assets.

  • Negotiator: Representing the organization in major transactions, contracts, and dispute resolutions.

• The Four-Step Scientific Decision-Making Process:

  • Step 1: Define the Problem (or Opportunity): Clarify organizational challenges or untapped business potential.

  • Step 2: Construct a Model: Formulate a conceptual, mathematical, or computational representation that accurately reflects the real-world problem environment.

  • Step 3: Identify and Evaluate Solutions: Generate alternative candidate solutions within the constructed model and evaluate their expected performance.

  • Step 4: Compare, Choose, and Recommend: Compare potential solutions against defined criteria, select the optimal alternative, and recommend it for execution.

• Information Systems Support Capabilities:

  • Supporting group communication, remote work, and synchronous/asynchronous collaboration.

  • Enabling improved data management and query optimization across enterprise units.

  • Structuring and querying enterprise data warehouses and handling high-velocity Big Data.

  • Providing advanced mathematical, algorithmic, and analytical support.

  • Overcoming innate human cognitive limitations regarding processing speed, working memory, and information storage.

  • Facilitating structured enterprise knowledge management.

  • Supplying "anywhere, anytime" pervasive mobile and cloud decision support.

• The Gorry and Scott-Morton Decision Support Framework (1971):

  • Combines Simon's (1977) degree of structuredness with Anthony's (1965) taxonomy of managerial control.

  • Degree of Structuredness (Simon, 1977):

  • Highly Structured (Programmed): Routine, repetitive problems governed by definitive, standardized decision rules.

  • Semistructured: Decisions with both structured and unstructured phases, requiring both automated algorithms and subjective human judgment.

  • Highly Unstructured (Nonprogrammed): Novel, ill-defined, complex challenges devoid of established pre-existing routines or procedures.

  • Types of Control (Anthony, 1965):

  • Operational Control: Ensuring specific task-level activities are executed effectively and efficiently.

  • Managerial Control (Tactical Planning): Securing resources and ensuring their effective utilization to fulfill organizational objectives.

  • Strategic Planning: Establishing top-level, long-range organizational policies, goals, and resource trajectories.

• The 9 Decision Grid Cells (Gorry and Scott-Morton Taxonomy):

  • Cell 1 (Structured Operational Control): Accounts receivable, accounts payable, order entry.

  • Cell 2 (Structured Managerial Control): Budget analysis, short-term forecasting, personnel reports, make-or-buy decisions.

  • Cell 3 (Structured Strategic Planning): Financial management, investment portfolio management, warehouse location analysis, distribution system design.

  • Cell 4 (Semistructured Operational Control): Production scheduling, inventory control.

  • Cell 5 (Semistructured Managerial Control): Credit evaluation, budget preparation, plant layout design, project scheduling, reward system design, inventory categorization.

  • Cell 6 (Semistructured Strategic Planning): Building a new plant, mergers & acquisitions, new product planning, compensation planning, quality assurance design, human resource policies, inventory strategic planning.

  • Cell 7 (Unstructured Operational Control): Buying software, approving commercial loans, operating a help desk, selecting a cover design for a magazine.

  • Cell 8 (Unstructured Managerial Control): Labor and business negotiating, recruiting an executive, buying hardware systems, political lobbying.

  • Cell 9 (Unstructured Strategic Planning): Research and development (R&D) planning, new technology development, corporate social responsibility planning.

Decision Support Systems (DSS) and Business Intelligence (BI)

• The Concept and Evolution of DSS:

  • Classic Definition (Gorry and Scott-Morton, 1971): Interactive computer-based systems that assist decision makers in utilizing data and models to solve unstructured and semistructured problems.

  • DSS couples the intellectual resources of individuals with the computational capabilities of the computer to elevate decision quality.

  • "Decision Support" functions as a historical umbrella term encompassing a wide variety of computerized decision-support tools.

  • Over time, standalone DSS concepts expanded, giving rise to enterprise-scale Business Intelligence.

• Definition and Core Scope of Business Intelligence:

  • Coinage: The term "Business Intelligence" was popularized by the Gartner Group in the mid-1990s, though its conceptual origins date back decades earlier.

  • Umbrella Definition: An umbrella term combining architectures, tools, databases, analytical tools, enterprise applications, and methodologies.

  • Described as a "content-free expression" because it signifies different technological configurations to different audiences.

  • Primary Objective: Provide intuitive access to enterprise data and models, granting business managers the capability to perform deep analyses.

  • Value Conversion Sequence: Facilitates the transformation chain:     $\text{Data} \longrightarrow \text{Information/Knowledge} \longrightarrow \text{Decisions} \longrightarrow \text{Action}$

  • Delivers "one version of the truth" across the enterprise by harmonizing and consolidating disparate data sources into a single unified format.

  • Shifted organizational paradigm from historical Executive Information Systems (EIS) designed exclusively for top leaders to "Everybody's Information System" (modern BI deployed enterprise-wide).

• Chronological History of Business Intelligence:

  • 1970s: Management Information Systems (MIS) reporting; marked by static, batch-oriented, periodic printed reports.

  • 1980s: Executive Information Systems (EIS); provided high-level visibility, dynamic drill-down capabilities, and executive dashboard views for top leadership.

  • 1990s: Emergence of On-Line Analytical Processing (OLAP), ad-hoc query tools, multidimensional data cubes, and dynamic reporting; Gartner Group formally coined the term "BI".

  • 2010s: Integration of Artificial Intelligence (AI), predictive data mining, text mining, web-based executive portals, cloud dashboards, Big Data infrastructures, and social media analytics.

  • 2020s and Beyond: Evolution toward pervasive, autonomous, real-time streaming analytics, deep machine learning integration, and edge intelligence.

• Interconnected Capabilities in the BI Ecosystem:

  • Central enterprise data infrastructure comprises ETL processes, Metadata repositories, Data warehouses, and specialized departmental Data marts.

  • Functional presentation and analysis layers include:

  • Spreadsheets (e.g., Microsoft Excel).

  • Traditional Decision Support Systems (DSS).

  • Executive Information Systems / Executive Support Systems (EIS/ESS).

  • Multidimensional On-Line Analytical Processing (OLAP).

  • Dynamic querying and financial reporting tools.

  • Enterprise scorecards, digital cockpits, and performance dashboards.

  • Automated workflows, alert generation, and notification systems.

  • Data mining, text mining, predictive analytics, and forecasting modules.

  • Information broadcasting tools and integrated enterprise web portals.

• DSS versus BI: Comparison and Synthesis:

  • Similarities: Both share fundamental architectural principles, rely on database systems, utilize modeling techniques, and emphasize data-driven decision enhancement.

  • Differences:

  • Direct vs. Indirect Support: DSS traditionally provided direct support for dedicated problem-solving tasks, whereas BI often provides indirect support by democratizing data access and broad reporting.

  • Target Audience: DSS historically focused on technical analysts and individual decision makers; BI targets operational employees, knowledge workers, and enterprise managers simultaneously.

  • System Development: DSS commonly relied on custom, in-house prototypes and tailored algorithms; BI systems emphasize commercial off-the-shelf (COTS) enterprise suites.

  • Origins: DSS originated within academia (theoretical modeling and behavioral science); BI was conceived, named, and commercialized by industry practitioners and technology analysts.

Architecture and Implementation of Business Intelligence

• The Four Major Architecture Components of BI:

  • 1. Data Warehouse (DW): The architectural foundation containing clean, integrated historical and subject-oriented data derived from multiple operational systems and external sources.

  • 2. Business Analytics: An integrated suite of software tools designed for manipulating, mining, synthesizing, and analyzing data stored within the data warehouse.

  • 3. Business Performance Management (BPM): An enterprise portfolio of metrics, key performance indicators (KPIs), and governance methodologies for tracking, monitoring, and evaluating organizational progress toward strategic goals.

  • 4. User Interface (UI): Presentation layer (e.g., visual dashboards, balanced scorecards, interactive enterprise portals) enabling end users to intuitively inspect and drill into data.

• High-Level Environments of BI:

  • Data Warehouse Environment: Maintained primarily by technical IT staff. Activities involve organizing, standardizing, summarizing, and loading heterogeneous data sources into centralized enterprise data warehouses and targeted data marts.

  • Business Analytics Environment: Leveraged by business users and data analysts. Involves interactive querying, analytical modeling, data manipulation, and extracting analytical results.

  • Performance and Strategy Environment: Governed by managers and corporate executives. Directs enterprise performance via BPM strategies, KPI tracking, and operational alignment.

  • Future Integration: Increasing convergence with intelligent systems (artificial intelligence, machine learning, autonomous systems) feeding directly into browser, portal, and dashboard interfaces.

• The Three-Step ETL (Extract, Transform, Load) Process:

  • 1. Extract:

  • Captures raw transactional data from diverse internal and external source systems.

  • Integrates data across operational transaction databases, relational databases, legacy systems, external file repositories, and unstructured data pools.

  • 2. Transform:

  • Cleanses corrupted, incomplete, or duplicate records.

  • Standardizes inconsistent data definitions, character encodings, and numerical units.

  • Applies explicit business rules and logic to harmonize conflicting data.

  • Profiles, audits, and validates data to resolve data quality defects.

  • 3. Load:

  • Writes processed, clean, and structured data into target enterprise data warehouses and localized data marts.

  • Prepares and structures database indexes and dimensional tables for rapid analytical access and querying.

• Business Value of BI Analytical Applications:

  • Customer Segmentation: Grouping customer portfolios into behavioral or demographic tiers for personalized targeting.

  • Propensity to Buy: Scoring individual likelihood to purchase supplemental goods or upgrade services.

  • Customer Profitability: Identifying exact margins delivered by specific accounts, separating high-yield from loss-inducing clients.

  • Fraud Detection: Flagging abnormal transactional patterns in real time to prevent illicit activity.

  • Customer Attrition (Churn): Predicting customers who exhibit high probabilities of defecting to competitors.

  • Channel Optimization: Selecting the most cost-effective and responsive communication and distribution pathways.

• Phases of a Typical BI Project Lifecycle:

Business Analytics Taxonomy and Hierarchy

• The INFORMS Analytics Taxonomy:

  • Formulated by the Institute for Operations Research and the Management Sciences (INFORMS).

  • Categorizes business analytics into three progressive disciplines: Descriptive, Predictive, and Prescriptive.

• Detailed Taxonomy Dimensions:

  • Descriptive Analytics:

  • Guiding Questions: "What happened?", "What is happening?"

  • Key Enablers: Business reporting systems, dynamic visual dashboards, enterprise scorecards, and data warehousing.

  • Business Outcomes: Well-defined business problems and clearly identified market opportunities.

  • Predictive Analytics:

  • Guiding Questions: "What will happen?", "Why will it happen?"

  • Key Enablers: Data mining algorithms, text mining, web and social media mining, and statistical forecasting techniques.

  • Business Outcomes: Accurate projections of future organizational states, trends, customer behaviors, and market conditions.

  • Prescriptive Analytics:

  • Guiding Questions: "What should I do?", "Why should I do it?"

  • Key Enablers: Mathematical optimization, dynamic computer simulation, quantitative decision modeling, and heuristic expert systems.

  • Business Outcomes: Best possible business decisions, maximized payoffs, and optimized automated transactions.

• The Four-Tier Analytics Hierarchy:

• Detailed Analytics Classifications:

  • Descriptive & Diagnostic Analytics:

  • Descriptive: Represents the traditional cornerstone of enterprise BI; utilizes OLAP, slice-and-dice queries, data visualization, and reporting to document historical occurrences.

  • Diagnostic: Seeks root cause attribution; provides drill-down, roll-up, and granular data inspection capabilities to isolate root causes (e.g., investigating why conversion rates spiked or slumped).

  • Predictive Analytics:

  • Definition: The rigorous deployment of statistical and computational algorithms to identify historical patterns and forecast the probability of future outcomes.

  • Primary Techniques: Linear and logistic regression analysis, supervised machine learning algorithms, deep neural networks, and association-rule data mining.

  • Prominent Applications: Sales lead scoring, customer churn prediction, inventory demand forecasting, and credit card fraud detection.

  • Prescriptive Analytics:

  • Definition: Advanced analytical discipline that determines, evaluates, and recommends the optimal course of action among feasible alternatives.

  • Primary Techniques: Mathematical programming (linear, integer, non-linear), constrained optimization models, stochastic simulation, and decision tree logic.

  • Exemplar Domain (Revenue Management):

    • Utilized heavily for perishable assets (airline passenger seats, hotel rooms, rental cars) where unallocated capacity expires.

    • Adjusts pricing dynamically in real time based on demand elasticity forecasts to maximize overall yield.

    • Hotels use revenue management models to adjust room prices per night.

  • Maturity Benchmark: Prescriptive analytics represents the highest tier of analytical maturity and operational sophistication.

Big Data Analytics and Advanced Architectures

• Defining Big Data and The Core Dimensions:

  • Big Data describes datasets whose scale, distribution, diversity, and velocity exceed the storage, management, and processing capacity of conventional database architectures.

  • The 3 Vs Framework:

  • Volume: The massive scale, sheer physical magnitude, and immense transactional footprint of collected data.

  • Velocity: The rapid speed at which data is continuously generated, ingested, transferred, and processed (batch vs. near-instantaneous streaming).

  • Variety: The wide range of incoming data types, spanning structured records, semi-structured files (JSON, XML), and unstructured content (text, audio, video, clickstream data).

  • Additional Critical Dimensions:

  • Veracity: The inherent cleanliness, quality, reliability, consistency, and trustworthiness of the data.

  • Value: The net business benefit and economic return extracted from the collected data through advanced analytics.

• Big Data versus Traditional Business Intelligence:

  • Big Data does not replace traditional BI; rather, it extends enterprise analytics capabilities into high-volume, multi-structured, and streaming domains.

• Core Big Data Frameworks and Technologies:

  • Hadoop Ecosystem:

  • HDFS (Hadoop Distributed File System): Fault-tolerant, distributed file storage system designed to execute on commodity server clusters.

  • MapReduce: Distributed software framework for parallel processing of massive data blocks across distributed nodes.

  • Hive and Pig: High-level abstraction languages providing SQL-like queries (HiveQL) and procedural data-flow scripting (Pig Latin) over distributed clusters.

  • HBase: Scalable, distributed, column-oriented NoSQL database layered directly over HDFS.

  • Apache Spark:

  • In-memory cluster computing engine that avoids costly disk I/O bottlenecks.

  • Executes computational workloads significantly faster than legacy MapReduce.

  • Houses native modules for streaming analytics (Spark Streaming), graph processing (GraphX), and machine learning (MLlib).

  • NoSQL Database Architectures:

  • Designed to overcome the horizontal scalability and rigid schema constraints of relational databases.

  • Document Stores: Store flexible, semi-structured document payloads (e.g., MongoDB, CouchDB).

  • Key-Value Stores: High-performance hash maps optimized for rapid read/write operations (e.g., Redis, DynamoDB).

  • Column-Family Stores: Store sparse datasets organized by dynamic columnar families (e.g., Apache Cassandra, HBase).

  • Graph Databases: Dedicated structures representing nodes, edges, and properties for complex relationship analysis (e.g., Neo4j).

• Cross-Industry BI and Analytics Applications:

  • Online Retail (Amazon.com):

  • Web clickstream path analysis to track real-time shopping behavior.

  • Behavioral customer segmentation for hyper-targeted merchandising.

  • Collaborative filtering engines delivering personalized product recommendations.

  • Automated digital marketing campaign management.

  • Dynamic pricing algorithms and real-time inventory forecasting.

  • Financial Services:

  • Quantitative credit risk analysis and portfolio exposure calculation.

  • Real-time automated transaction monitoring for fraud detection.

  • Granular customer profitability and lifetime value modeling.

  • Healthcare:

  • Clinical patient outcome analysis and predictive readmission prevention.

  • Hospital resource optimization (bed occupancy, nurse-to-patient staffing, surgical suite scheduling).

  • Treatment effectiveness analytics comparing longitudinal medical therapies.

Questions and Discussion Prompts

• Prompt: The Equivalence of DSS and BI:

  • Question: Is DSS equal to BI ($\text{DSS} = \text{BI}$)?

  • Discussion Points: While BI evolved directly out of DSS and shares similar architectural components and data focus, they diverge in scope, target audience, and historical development. BI serves as an enterprise-wide umbrella term popularized by commercial industry vendors to democratize reporting across entire organizations, whereas DSS originated within academic research as specialized, direct analytical modeling systems built for specific, complex decision-making tasks.

• Prompt: Application of Predictive Analytics across Industries:

  • Question: What predictive analytics applications are relevant to your specific industry?

  • Discussion Points: Contextual evaluation of how machine learning, regression, and data mining can be applied to solve domain-specific operational bottlenecks, anticipate customer churn, automate lead scoring, forecast demand curves, and identify anomalous or fraudulent behavior.

• Prompt: Architectural Selection of NoSQL Databases:

  • Question: When would an enterprise choose a NoSQL database over a traditional relational database (RDBMS)?

  • Discussion Points: Criteria include scenarios where data schemas are highly dynamic or polymorphic, data volumes exceed petabyte scales requiring horizontal scale-out architectures across commodity clusters, applications demand low-latency read/write throughput without rigid ACID transaction constraints, or data is natively represented as unstructured documents, key-value caches, or complex relationship graphs.