Mid-exam revision, Lectures 1 to 3

The exam covers everything up to where lessons ended: Simon’s bounded rationality and the behavioral side of decision making. Read the lecture pages, lock in the mnemonics, then write the essays from memory.

How to use this in the last hours

1. Read Lecture 3 fully, then the mnemonics page.
2. Do the essay drills without the word bank first. Turn it on only for the blanks you can’t recall.
3. Any blank you got wrong: go back to that lecture section and say the point out loud once.
4. Finish with the quick quiz.

Lecture 3: Decision making

The last deck taught before the exam. Expect “compare”, “explain” and “discuss with examples” questions here.

Nature of decision making

Decision making is the act of choosing one alternative from among a set of alternatives. First you decide that a decision has to be made, then you identify a set of feasible alternatives, then you select one.

Some experts see it as the most basic of all managerial activities. It is linked most closely with Planning, but is also part of Organizing, Leading and Controlling.

The decision-making process: recognize and define the decision situation, identify alternatives, choose the best alternative, and put it into practice.

Managers make decisions about problems (undesirable situations, e.g. cutting costs by 10%) and opportunities (desirable situations, e.g. learning that profits are higher than projected). It can take a long time to know whether a decision was right.

Effective decisions may aim to optimize (profits, sales, market share, employee welfare), minimize (losses, expenses, employee turnover), or select the best way out of a business, lay off staff, or end a strategic alliance.

Types of decisions

Decision-making conditions

• Certainty: very few organizational decisions, because the environment is complex and turbulent.
• Risk: based on past experience, relevant information, advice of others and own judgment. The decision is “calculated” on which alternative has the highest probability of working. Examples: union negotiations; Porsche choosing an SUV focus vs high-performance sports cars.
• Uncertainty: managers must secure as much relevant information as possible and approach the situation logically and rationally. Intuition, judgment and experience always play major roles.

Classical model vs Administrative model

The Classical Decision Model tells managers how they should make decisions. It assumes managers are logical and rational and that their decisions serve the organization’s best interests. Managers should obtain complete and perfect information, eliminate uncertainty, and evaluate everything rationally and logically. These conditions rarely, if ever, exist.

The Administrative Model, developed by Herbert A. Simon (Nobel Prize in Economics), describes how decisions are actually made. Decision makers have incomplete and imperfect information, are constrained by bounded rationality and tend to satisfice.

Bounded rationality: decision makers are limited by their values and unconscious reflexes, skills and habits. Slide example: American automakers vs foreign automakers.

Satisficing: the tendency to search for alternatives only until one is found that meets some minimum standard of sufficiency, rather than an exhaustive search for the best possible alternative.

Using both: the Classical model shows how managers can at least attempt to be more rational and logical; the Administrative model helps managers understand their inherent biases and limitations.

Rational decision making: 6 steps

Six steps that keep the decision maker focused on facts and logic and guard against inappropriate assumptions and pitfalls.

1. Recognize and define the decision situation. Define precisely what the problem is; develop a complete understanding; carefully analyse the situation.
2. Identify alternatives. Alternatives may be limited by legal, moral and ethical norms, authority constraints, available technology, economic considerations and unofficial social norms.
3. Evaluate alternatives. Each must pass three stages (FSA below). Put “price tags” on consequences.
4. Select an alternative. The real test of decision making. Optimization is the goal because a decision affects several individuals or departments. Several acceptable alternatives may exist; you may not need to reject all others.
5. Implement the chosen alternative. Consider people’s resistance to change. Operational plans help. Unanticipated consequences are still likely.
6. Follow up and evaluate results. Did it serve its original purpose? If not: adopt another identified alternative, or recognize the situation was wrongly defined and start again, or decide it needs more time or a different implementation.

Behavioral aspects of decision making

Decisions sometimes reflect subjective considerations (tastes). Behavioral forces include political forces, intuition, escalation of commitment, risk propensity and ethics.

• Political forces, coalitions: an informal alliance of individuals or groups formed to achieve a common goal (stockholders, directors, parliament blocs). Impact can be positive or negative. Managers must know when to use them, check they act in the organization’s interest, and control negative effects.
• Intuition: an innate belief about something without conscious consideration; doing something because it “feels right”. Based on years of experience in similar situations; lets managers skip an a-to-z process occasionally.
• Escalation of commitment: staying with a decision even when it appears to be wrong (Pan Am holdings). Guard against sticking too long, but don’t bail out too soon.
• Risk propensity: the extent to which a decision maker is willing to gamble. Organizational culture is a prime ingredient.
• Ethics: relationships of the firm to employees (closing a department to save money), of employees to the firm, and of the firm to other economic agents.

Lecture 2: BI, analytics and decision support

Frameworks and definitions. Names and years matter here: Simon 1977, Anthony 1965, Gorry and Scott-Morton 1971, Gartner mid-1990s.

Business Pressures-Responses-Support model

Companies are moving aggressively to computerized support of their operations (BI). Pressures come from today’s competitive climate; organizations take responses to counter them; computerized support facilitates the responses.

Organizational responses: be reactive, anticipative, adaptive and proactive. Employ strategic planning, new business models, restructure business processes, join business alliances, improve corporate information systems.

Closing the strategy gap: a major objective of computerized decision support is to close the gap between current performance and desired performance (as expressed in mission, objectives, goals and strategy).

Managerial decision making

Management is a process by which organizational goals are achieved using resources. Inputs are resources, output is goal attainment, and success = outputs / inputs. Decision making is selecting the best solution from two or more alternatives.

Mintzberg’s 10 managerial roles (3, 3, 4)

Hook: “Funny Lions Laugh, Many Dogs Sing, Every Duck Really Naps”.

Four-step decision process (scientific approach)

1. Define the problem (or opportunity).
2. Construct a model that describes the real-world problem.
3. Identify possible solutions to the modelled problem and evaluate them.
4. Compare, choose and recommend a potential solution.

How IS supports decisions: group communication and collaboration, improved data management, managing data warehouses and Big Data, analytical support, overcoming cognitive limits in processing and storing information, knowledge management, anywhere anytime support.

Gorry and Scott-Morton framework (1971)

A grid of two dimensions:

• Degree of structuredness (Simon, 1977): highly structured (programmed), semi-structured, highly unstructured (nonprogrammed).
• Types of control (Anthony, 1965): strategic planning (top level, long range), management control (tactical planning), operational control.

Hook: “Simon Structures, Anthony Controls.”

DSS and BI

DSS (Gorry and Scott-Morton, 1971): interactive computer-based systems that help decision makers use data and models to solve unstructured problems. DSS couple the intellectual resources of individuals with the capabilities of the computer to improve decision quality. DS became an umbrella term and evolved into BI.

BI is an evolution of decision support. Then: Executive Information System. Now: “Everybody’s Information System”. BI adds visualizations, alerts and performance measurement. The term emerged from industry.

Definition: BI is an umbrella term combining architectures, tools, databases, analytical tools, applications and methodologies. It is a content-free expression (means different things to different people). Major objective: easy access to data and models so managers can analyse. BI transforms data to information (and knowledge), to decisions, and finally to action.

History of BI

BI architecture: four components

Business value of BI applications: customer segmentation, propensity to buy, customer profitability, fraud detection, customer attrition, channel optimization.

DSS vs BI

Analytics taxonomy (INFORMS): descriptive, predictive, prescriptive.

Lecture 1: Course overview

Definitions and lists. Learn the four analytics questions word for word.

BI and BA

BI is a broad category of applications, technologies and processes for gathering, sorting, accessing and analysing data to help business users make better decisions.

BA is iterative, methodical exploration of organizational data with emphasis on statistical analysis to drive decision making.

BI focuses on descriptive analytics; BA extends to predictive and prescriptive. Applications: performance management, CRM, supply chain (inventory), risk management (fraud detection), strategic planning (market expansion).

Value of BI and analytics

• Operational efficiency: find bottlenecks (manufacturing yield).
• Customer experience: personalized marketing (targeted recommendations).
• New revenue: market trends (new product development).
• Risk mitigation: detect anomalies (credit risk assessment).
• Strategic decisions: comprehensive view (market share analysis).
• Real-time monitoring (supply chain disruptions).

Components of BI systems

Data sources and management

• Internal: transactional systems, operational databases, ERP. External: market research, social media, public datasets, third-party providers.
• ETL: extract from sources, transform to fit needs, load into warehouse.
• Data quality management: accuracy, completeness, consistency via cleansing and validation.
• Master data management: single authoritative version of critical data (customers, product catalogs).

Storage and processing

• Data warehouse: centralized repository of integrated data, optimized for querying and analysis.
• Data mart: subject-specific subset of a warehouse (sales, finance).
• OLAP: multidimensional analysis (sales by region, product and time).
• In-memory databases: data in main memory for real-time analytics.
• Big data technologies: Hadoop, Spark for structured and unstructured data.

Analytics and visualization tools

BI platforms, data mining tools, advanced analytics (predictive modelling, ML), self-service BI (non-technical users without IT), mobile BI.

Data warehousing process

Collect, clean, integrate, design schema (star, snowflake), implement ETL, maintain (updates, tuning, capacity planning). Hook: “Clever Children Invent Super Easy Machines”.

Data mining techniques

Visualization strategies

Chart selection (bar for comparisons, line for trends), colour schemes, interactive elements (drill-down, filters, tooltips), storytelling, design principles (clarity, whitespace, consistency), customization for audiences (executive vs analyst).

BI problem-solving approach

Problem definition, data collection and preparation, exploratory data analysis, advanced analytics, results interpretation, implementation and monitoring.

Real-world applications: sales optimization (dynamic pricing), churn prevention, supply chain optimization, fraud detection, marketing effectiveness (A/B testing), HR analytics (turnover).

Star schema vs snowflake schema

From the data warehouse design slides and the data modelling reading. Both are dimensional models; the difference is whether dimension tables are normalized.

The building blocks

The lecture slide: a star schema is a design particularly suited to ad hoc queries, in which dimensional data are separated from fact data. It has two types of tables: dimension tables and fact tables.

Relationship: each dimension table is linked to the fact table one-to-many (one product appears in many sales rows). On the slide this shows as 1 at the dimension and m at the fact table.

Hook: Facts are numbers you add up; dimensions are words you filter and group by.

Star schema

Each dimension holds all its attributes in one denormalized table. Product name, category and department all sit in Dim_Product, so “Beverages” is repeated on every beverage product row. Every dimension is one join away from the fact table.

Snowflake schema

A snowflake schema extends the star schema by normalizing dimension tables into multiple levels of related tables. Category and department move into their own tables, so “Beverages” is stored once. The diagram branches outward like a snowflake. Reaching category now needs an extra join.

Side-by-side comparison

Same question, two schemas (SQL)

Question: total sales in ZMW by product category.

Star: one join

SELECT p.category,
       SUM(f.amount_zmw) AS total_sales
FROM   fact_sales f
JOIN   dim_product p ON f.product_key = p.product_key
GROUP BY p.category;

Snowflake: two joins

SELECT c.category_name,
       SUM(f.amount_zmw) AS total_sales
FROM   fact_sales f
JOIN   dim_product  p ON f.product_key  = p.product_key
JOIN   dim_category c ON p.category_key = c.category_key
GROUP BY c.category_name;

Related term worth one line in an essay: a galaxy schema (fact constellation) has several fact tables sharing the same dimension tables.

All mnemonics

Cover the right-hand side and recite. If you can expand every hook, you can write every essay.

Essay drills

Each drill is a full model answer with key terms removed. Type the missing word, then check. Your answers are saved in this browser.

Quick quiz

Pick an answer to see whether it is right and why.