How to build a business intelligence course: a complete guide for lecturers

Central question: What makes information genuinely useful for a business decision rather than simply available?

Position Business Intelligence as a managerial capability, not a dashboard product. Students need a shared language for data, information, insight and decision support before they encounter architectures or tools. Connect every BI artefact to a user, a decision, a cadence and a consequence.

What this concept covers

• BI, analytics and decision-support distinctions
• Operational, tactical and strategic decisions
• The data-to-information-to-insight-to-action chain
• BI users, owners and stakeholders
• Decision latency, information quality and business value

Learning outcomes

• Distinguish BI from reporting, data science and transaction processing
• Frame a decision that can be supported by BI
• Identify the user, cadence and required evidence
• Evaluate whether an information product is decision-relevant

How to teach it

• Open with three dashboards and ask what decision each supports
• Rewrite vague management questions into decision statements
• Map information flow from source system to manager action
• Use a short memo requiring a recommendation

Case-style example

Northline Mobility has 18 depots, monthly revenue of $42 million and an on-time delivery rate that has fallen from 94% to 88%. The COO wants one performance dashboard, while depot managers want route-level exceptions. Students specify the users, refresh cadence, five priority measures and an escalation rule. Two defensible designs emerge: executive standardisation or local diagnostic depth. Students discover that BI scope follows the decision, not the amount of available data.

Where students struggle

Students often equate BI with software features. Require them to state who decides what, when, using which evidence and what action follows.

Reading and a quick check

Use the opening chapter of Sharda, Delen and Turban, then ask students to convert one business question into a decision, metric and evidence chain.

Applied teaching activity

Use a case discussion, model critique, dashboard workshop or short decision memo. The objective is to make students commit to a defensible analytical choice without forcing a simulation where it does not fit.

Where it leads

Once students can define the decision, they can ask where the required data comes from and how it should be integrated.

Central question: How should data from different systems be organised so decision-makers receive consistent information?

BI depends on an information architecture that connects operational systems, external data and analytical layers. The course need not become database administration, but students should understand why fragmented definitions, grain mismatches and late data can undermine otherwise polished analysis.

What this concept covers

• Operational sources, APIs, files and external data
• ETL and ELT logic
• Data warehouses, marts, lakes and lakehouse concepts
• Fact, dimension and grain decisions
• Semantic layers and shared business definitions

Learning outcomes

• Map a simple BI architecture
• Identify integration and grain risks
• Explain why shared definitions matter
• Select an architecture appropriate to a decision context

How to teach it

• Give teams a messy system landscape and ask for a target-state map
• Use a revenue metric defined differently by sales and finance
• Show how grain changes what can be aggregated safely
• Keep the emphasis on management implications rather than database syntax

Case-style example

Arbor Retail sells through 62 stores and an e-commerce site. Store sales arrive daily at transaction level, online orders arrive hourly and returns sit in a separate weekly file. Finance reports £18.4 million monthly revenue while Sales reports £19.1 million because one excludes returns and the other uses order date. Students design the source-to-report flow and choose a revenue definition. They discover that reconciliation is an architecture and governance problem before it is a visualisation problem.

Where students struggle

Students may jump to tool choice. Require them to settle source ownership, grain, timing and business definitions first.

Reading and a quick check

Pair the BI data warehousing chapter with a one-page architecture sketch and a short reconciliation exercise.

Applied teaching activity

Use a case discussion, model critique, dashboard workshop or short decision memo. The objective is to make students commit to a defensible analytical choice without forcing a simulation where it does not fit.

Where it leads

Architecture creates the path, but trustworthy BI still depends on data preparation, quality controls and governance.

Central question: When is data good enough to support a decision, and who is accountable when it is not?

Teach data quality as decision risk. Missing values, duplicate customers, inconsistent dates and unowned definitions can change reported performance and therefore change managerial action. Governance gives students a way to assign ownership, escalation and acceptable quality thresholds.

What this concept covers

• Completeness, accuracy, consistency, timeliness and validity
• Profiling, cleaning and transformation
• Master and reference data
• Data ownership, stewardship and lineage
• Access controls, retention and quality thresholds

Learning outcomes

• Diagnose common data-quality problems
• Set fit-for-purpose quality rules
• Assign governance responsibilities
• Explain how quality failures affect decisions

How to teach it

• Provide a deliberately flawed dataset and a decision deadline
• Classify issues by materiality, not just count them
• Introduce a RACI-style governance map
• Require a use, fix or stop decision for each defect

Case-style example

Beacon Health has 120,000 patient-service records used to plan staffing. Eight percent have missing service duration, 3% are duplicated and two clinics use different service codes. Operations wants the dashboard tomorrow; the data team wants a two-week cleanse. Students quantify which defects could bias average workload, set an acceptance threshold and recommend whether to publish with caveats or delay. They discover that perfect data is not the same as fit-for-purpose data.

Where students struggle

Students often want either perfect data or immediate delivery. The better question is which defects are material to the decision and what control makes the risk acceptable.

Reading and a quick check

Assign Bernardo et al. (2024) on data governance and quality, followed by a short data-quality triage.

Applied teaching activity

Use a case discussion, model critique, dashboard workshop or short decision memo. The objective is to make students commit to a defensible analytical choice without forcing a simulation where it does not fit.

Where it leads

With trusted data, students can move from preparation into descriptive analysis and performance interpretation.

Central question: What happened, how large was the change and which comparisons make the result meaningful?

Descriptive analytics is the bridge between raw data and managerial interpretation. Students should select useful denominators, time windows and benchmarks rather than reporting every available number. This is also where financial and operating data begin to work together.

What this concept covers

• Summary statistics and distributions
• Rates, ratios and normalisation
• Trend, variance and benchmark analysis
• Financial and operating performance measures
• Segmentation by product, customer, geography and channel

Learning outcomes

• Select appropriate descriptive measures
• Calculate and interpret changes and ratios
• Choose relevant comparisons and benchmarks
• Separate signal from scale effects

How to teach it

• Use the same dataset with three denominators to show how stories change
• Build a compact performance brief before any dashboard
• Combine financial and non-financial measures
• Ask what the metric cannot tell students

Case-style example

Crescent Foods reports revenue up 12% to £96 million, but gross margin falls from 34% to 29%, inventory days rise from 48 to 67 and customer returns increase from 2.1% to 3.8%. Students calculate absolute and percentage changes, benchmark the four indicators and write a 120-word performance diagnosis. One team may emphasise growth; another deteriorating quality. They discover that descriptive BI is comparative interpretation, not a list of KPIs.

Where students struggle

Students may treat growth as automatically positive. Require them to interpret scale, profitability, cash and operational quality together.

Reading and a quick check

Use the descriptive analytics chapter plus a short performance-measurement worksheet.

Where a simulation fits

The Financial Statement Analysis Simulation fits here because students already hold the analytical concepts needed to interpret its data and defend a decision rather than simply follow instructions.

Where it leads

Descriptive measures show what changed. Diagnostic analysis asks why it changed and which drivers matter.

Central question: Which patterns, segments and relationships are plausible drivers of the result, and what evidence would distinguish correlation from explanation?

Diagnostic BI should teach disciplined curiosity. Students explore dimensions, exceptions and relationships, but they must resist turning a visual pattern into a causal claim. The emphasis is on generating and testing plausible explanations, identifying missing evidence and deciding what analysis should come next.

What this concept covers

• Drill-down and slice-and-dice analysis
• Segmentation and exception analysis
• Correlation and association
• Root-cause hypotheses
• Variance decomposition and driver trees

Learning outcomes

• Generate diagnostic hypotheses
• Use segmentation to locate performance differences
• Interpret correlations cautiously
• Recommend additional evidence before acting

How to teach it

• Begin with one surprising KPI movement and hide the explanation
• Let teams request two additional cuts of data
• Use driver trees to structure competing explanations
• End with a what-we-know, infer and still-need memo

Case-style example

Marlow Telecom sees monthly churn rise from 2.4% to 3.6% across 210,000 subscribers. Churn is 6.8% among customers with two or more service incidents, 4.1% among month-to-month customers and 2.0% among annual-contract customers. Students build a driver tree, calculate segment contributions and propose two tests before intervention. One conclusion focuses service reliability; another contract design. They discover that BI narrows hypotheses but does not prove causality by itself.

Where students struggle

Students often overstate causality from cross-sectional patterns. Make them label evidence as observation, hypothesis or validated explanation.

Reading and a quick check

Pair Hjelle et al. (2024) with a diagnostic dashboard exercise and an evidence-quality checklist.

Where a simulation fits

The Financial Statement Analysis Simulation fits here because students already hold the analytical concepts needed to interpret its data and defend a decision rather than simply follow instructions.

Where it leads

Once students can diagnose performance, they need to communicate it visually without hiding uncertainty or overloading the user.

Central question: How can a dashboard make the important signal easier to see without distorting the underlying evidence?

Dashboard design is where BI becomes visible, so it is tempting to teach mainly aesthetics. A stronger course treats visualisation as an information-design decision: chart choice, scale, hierarchy, annotation, interaction and context should reduce cognitive load and support a defined action.

What this concept covers

• Chart selection and visual encodings
• Scales, baselines and comparison design
• Information hierarchy and cognitive load
• Filters, drill-through and interaction
• Annotations, uncertainty and accessibility

Learning outcomes

• Choose charts appropriate to analytical tasks
• Critique misleading or overloaded displays
• Design a dashboard around a user decision
• Explain visual trade-offs and accessibility choices

How to teach it

• Run a dashboard critique before showing design rules
• Give each group the same dataset but a different decision user
• Limit the dashboard to one screen and six visuals
• Require an oral defence of every chart and interaction

Case-style example

Harbour Hotels tracks 14 properties with occupancy, ADR, RevPAR, cancellations and guest satisfaction. Executives want portfolio trends, while general managers want daily exceptions. Students receive 18 candidate visuals and must design a six-visual dashboard with one filter and two annotations. The executive design favours comparison; the operator design favours alerts. They discover that a good dashboard is role-specific, and that adding more visual elements can reduce decision quality.

Where students struggle

Students often optimise for visual novelty or density. Mark them on decision relevance, comparison quality, accessibility and justified omission.

Reading and a quick check

Assign the dashboard visualisation study by Hjelle et al. (2024), then use the SMU F&B case as a design application.

Applied teaching activity

Use a case discussion, model critique, dashboard workshop or short decision memo. The objective is to make students commit to a defensible analytical choice without forcing a simulation where it does not fit.

Where it leads

Dashboards become more useful when measures are deliberately linked to strategy through KPIs and scorecards.

Central question: Which measures should management use to steer performance, and how can a KPI avoid rewarding the wrong behaviour?

A KPI is not simply an important number. It has an owner, definition, target, frequency, threshold and intended response. This concept connects BI to management control, incentives and operational decision-making while showing why local optimisation can damage enterprise performance.

What this concept covers

• Metric versus KPI distinctions
• Leading and lagging indicators
• Targets, thresholds and tolerances
• Balanced scorecards and strategy maps
• Metric ownership and behavioural consequences

Learning outcomes

• Design a decision-linked KPI
• Distinguish leading and lagging measures
• Identify gaming and incentive risks
• Build a compact scorecard with explicit actions

How to teach it

• Improve a weak KPI such as increase sales
• Use a KPI dictionary with owner, formula, target and action
• Compare finance, operations and marketing perspectives
• Debate whether a higher KPI is always better

Case-style example

Orion Audio is choosing between three product lines. Product A has a 42% contribution margin but a £3.8 million fixed-cost base; Product B has a 31% margin and faster payback; Product C has lower margin but the highest NPV at £9.4 million. Students design a five-KPI scorecard for the CEO, then explain how CFO, COO and CMO priorities would change the weighting. They discover that KPI selection encodes managerial priorities and can produce different defensible decisions.

Where students struggle

Students often choose measures that are easy to calculate rather than measures that govern the intended decision.

Reading and a quick check

Use the managerial analytics chapters and require a KPI dictionary plus one example of a perverse incentive.

Where a simulation fits

The Managerial Accounting Simulation fits here because students already hold the analytical concepts needed to interpret its data and defend a decision rather than simply follow instructions.

Where it leads

Once KPIs define current performance, forecasting asks what is likely to happen next and how uncertainty should be communicated.

Central question: How should a BI course use forecasts without creating false confidence in uncertain future outcomes?

Predictive analytics extends BI beyond monitoring. The lecturer-facing objective is not to train specialist machine-learning engineers, but to help students frame a prediction problem, choose suitable baselines, interpret forecast error and connect predictions to decisions.

What this concept covers

• Forecasting questions and target variables
• Train, validation and test logic at a conceptual level
• Regression, classification and time-series intuition
• Error measures, confidence and calibration
• Model drift and changing business conditions

Learning outcomes

• Frame a predictive business question
• Interpret forecast error and uncertainty
• Compare a model with a simple baseline
• Explain how prediction quality affects a decision

How to teach it

• Use a naive baseline before introducing a complex model
• Choose an error metric based on business cost
• Compare model accuracy with actionability
• Include a forecast revision when new data arrives

Case-style example

BluePeak Energy forecasts weekly demand for 40 sites. A new model reduces mean absolute percentage error from 14% to 9%, but underestimates the top 5% of peak-demand weeks by an average 11%. Students compare the model with a seasonal baseline, calculate the likely capacity shortfall and recommend whether to deploy it for staffing and procurement. They discover that average accuracy can improve while decision risk worsens in the cases management cares about most.

Where students struggle

Students may treat a more accurate model as automatically better. Require them to connect the error profile to the cost of the decision.

Reading and a quick check

Assign Chatterjee et al. (2023) on analytics, forecasting and performance, followed by a forecast-decision memo.

Where a simulation fits

The Portfolio Management Simulation fits here because students already hold the analytical concepts needed to interpret its data and defend a decision rather than simply follow instructions.

Where it leads

Predictions inform choices. Prescriptive analytics makes the trade-off explicit by selecting actions under constraints.

Central question: How should managers choose among competing actions when resources, risk and objectives conflict?

Prescriptive analytics turns insights into resource allocation. Students should distinguish a model objective from a managerial objective, recognise constraints and test sensitivity. A mathematically optimal answer can still be strategically weak if the model omits risk, sequencing or implementation capacity.

What this concept covers

• Decision variables, objectives and constraints
• Optimisation and capital rationing
• Scenario and what-if analysis
• Sensitivity and robustness
• Model limits and managerial overrides

Learning outcomes

• Formulate a constrained decision problem
• Interpret an optimisation recommendation
• Test sensitivity to key assumptions
• Defend or override a model output with evidence

How to teach it

• Use a limited-budget allocation problem
• State objective and constraints in plain language
• Stress-test one assumption at a time
• Mark the defence of an override as carefully as the calculation

Case-style example

Ridgeway Manufacturing has £10 million to allocate across eight projects. Four individually have positive NPVs, but funding all attractive projects requires £14.2 million. One portfolio produces £3.1 million total NPV with high execution concentration; another produces £2.9 million with greater diversification and faster payback. Students compute both portfolios and recommend one under a board risk limit. They discover that capital allocation combines optimisation with risk appetite and implementation judgement.

Where students struggle

Students often assume the objective function captures everything that matters. Ask what is outside the model and whether it is decision-critical.

Reading and a quick check

Use the prescriptive analytics chapter and a short capital-rationing exercise before simulation.

Where a simulation fits

The Capital Budgeting Simulation fits here because students already hold the analytical concepts needed to interpret its data and defend a decision rather than simply follow instructions.

Where it leads

Prescriptive analysis creates recommendations, but widespread BI adoption depends on self-service capability, data literacy and governance.

Central question: How can organisations give users more analytical autonomy without creating inconsistent metrics and uncontrolled data use?

Modern BI increasingly shifts analysis from specialist teams to business users. This creates speed and local ownership but also duplication, metric drift and governance risk. Students should learn that self-service BI is an operating model involving roles, training, certified data, support and guardrails.

What this concept covers

• Self-service BI use cases and user segments
• Data literacy and analytical competency
• Certified datasets and semantic models
• Centres of excellence and support models
• Adoption, trust and change management

Learning outcomes

• Classify self-service BI scenarios
• Assess user capability and governance needs
• Design lightweight guardrails
• Recommend an adoption approach for different user groups

How to teach it

• Create novice, casual-user and power-user personas
• Decide which activities should be centralised or self-service
• Audit two conflicting dashboards for metric drift
• Build an adoption plan with training and support

Case-style example

Vantage Insurance has 420 managers using 96 self-built dashboards. Monthly active-policy figures differ by up to 7% across teams, while central BI requests take 12 working days. Students choose between tighter central control and a governed self-service model with certified datasets, training and publishing rights. They estimate the service-level and governance trade-offs. They discover that the real design problem is allocating analytical responsibility, not choosing between freedom and control.

Where students struggle

Students may present self-service as democratisation with no operating cost. Make them specify support, certification and escalation responsibilities.

Reading and a quick check

Assign Passlick et al. (2023) and Deja et al. (2021), then ask students to propose a three-tier self-service model.

Applied teaching activity

Use a case discussion, model critique, dashboard workshop or short decision memo. The objective is to make students commit to a defensible analytical choice without forcing a simulation where it does not fit.

Where it leads

Self-service creates scale. The next concept asks how AI and advanced analytics change what BI systems can generate and what users must verify.

Central question: Where does AI genuinely improve BI work, and where does automation increase the need for verification and human judgement?

AI can accelerate query generation, narrative summaries, anomaly detection, forecasting and dashboard creation. The educational risk is allowing polished automated outputs to replace reasoning. Students should use AI as an analytical assistant while retaining responsibility for data provenance, assumptions, validation and decision consequences.

What this concept covers

• Augmented analytics and natural-language querying
• Automated insights and anomaly detection
• Machine-learning integration with BI
• Generative AI for analysis and narrative
• Human oversight, verification and model risk

Learning outcomes

• Identify appropriate AI-enabled BI use cases
• Verify AI-generated analytical claims
• Explain model and data risks to decision-makers
• Design a human-in-the-loop review step

How to teach it

• Give students an AI-generated summary containing unsupported claims
• Trace every claim to evidence
• Compare human and AI narratives from the same dashboard
• Require a permitted-use declaration and validation checklist

Case-style example

LumaBank deploys an AI assistant that summarises 55 branch dashboards each morning. In a pilot, 92% of summaries are directionally correct, but 4% misstate a denominator and 4% infer causes not present in the data. Students design a review workflow, decide which summaries can be auto-published and specify escalation thresholds. One group favours speed; another tighter control. They discover that AI changes the cost of producing insight but not accountability for its accuracy.

Where students struggle

Students may confuse fluent explanation with validated analysis. Require traceability from claim to data and a clear statement of uncertainty.

Reading and a quick check

Assign the NIST Generative AI Profile (2024) alongside the AI chapters of Sharda, Delen and Turban.

Applied teaching activity

Use a case discussion, model critique, dashboard workshop or short decision memo. The objective is to make students commit to a defensible analytical choice without forcing a simulation where it does not fit.

Where it leads

The final concept combines governance, ethics and action so that BI recommendations are defensible beyond technical accuracy.

Central question: When should an organisation act on an insight, and what ethical, privacy, security or implementation constraints should change that action?

A Business Intelligence course should end with accountability. Data can be accurate yet inappropriate to use, an algorithm can be predictive yet unfair, and a dashboard can show a problem without telling managers what to do. Students integrate evidence quality, stakeholder impact, governance and implementation into a final recommendation.

What this concept covers

• Privacy, access and purpose limitation
• Bias, fairness and representativeness
• Security and sensitive information
• Decision rights, escalation and accountability
• From insight to intervention, monitoring and learning

Learning outcomes

• Identify ethical and governance risks in a BI recommendation
• Assess whether data use is proportionate and defensible
• Design an action and monitoring plan
• Communicate limitations and residual uncertainty

How to teach it

• Use a high-performing but ethically ambiguous segmentation model
• Require an impact assessment before recommendation
• Ask who has authority to act and who bears the downside
• End with a board-style decision memo and oral defence

Case-style example

Eastgate Lending identifies a customer segment with a 2.7x higher missed-payment rate using 38 variables, including postcode and device behaviour. Excluding those variables reduces model lift by 16% but lowers proxy-discrimination concerns. Students compare the two models, define a permissible decision use and design monitoring metrics for false positives and complaints. They discover that the strongest predictive model is not automatically the strongest business decision when legitimacy and stakeholder impact are included.

Where students struggle

Students may treat governance as a final compliance checklist. Make it part of the decision logic from data collection through action and monitoring.

Reading and a quick check

Use Sharda, Delen and Turban on ethical and managerial considerations plus NIST AI RMF material.

Applied teaching activity

Use a case discussion, model critique, dashboard workshop or short decision memo. The objective is to make students commit to a defensible analytical choice without forcing a simulation where it does not fit.

Where it leads

This closes the BI lifecycle: frame the decision, build trustworthy information, analyse it, communicate it, act and learn from the result.

10-week course structure

A 10-week version should combine architecture with governance and dashboard design with KPI teaching. Keep prediction, prescriptive choice, self-service/AI and the final governance capstone so the lifecycle does not collapse into reporting.

14-week course structure

A 14-week version creates space to separate architecture, data quality, diagnostic analysis and visual design, and to use more than one applied simulation with a proper post-activity debrief.

The Financial Statement Analysis Simulation places each student in the role of a Financial Analyst assessing ChippyCo. Students connect the Income Statement, Balance Sheet and Cash Flow Statement, calculate and interpret profitability, liquidity, efficiency and solvency ratios, respond to updated information across five reporting periods, and finish with a supported share-price prediction. It is the closest current Finsimco fit to BI because students turn structured quantitative and qualitative data into an evidence-based judgement.

Why it fits here

• Students already know descriptive and diagnostic analysis, so the activity can focus on interpretation rather than formula recall.
• The repeated reporting periods create a natural BI cycle: measure, compare, interpret change, revise a view and explain why.
• Individual submissions provide attributable evidence, which helps separate calculation from interpretation.

What students do

• Analyse the three primary financial statements and connect evidence across them.
• Calculate and interpret profitability, liquidity, efficiency and solvency ratios.
• Review updated financial statements, company developments and analyst/annual reports across five periods.
• Revise a share-price view when new evidence changes the performance story.
• Use the final reflection view to compare share-price history, answers and class outcomes.

Recommended course placement

• At the end of Session 4 after descriptive analytics.
• Across Sessions 4 and 5 if the lecturer wants a fuller diagnostic debrief.
• Before a dashboard or KPI assignment so students have a rich performance dataset to interpret.

Knowledge students should have before the simulation

• Three-statement financial statement basics
• Ratio interpretation
• Trend and benchmark logic
• Difference between calculation accuracy and decision judgement

Delivery format

Suggested debrief questions

• Which ratio or trend changed your view most, and why?
• Where did a correct calculation still allow more than one interpretation?
• Which qualitative development changed the meaning of a reported number?
• What additional dashboard view would have reduced uncertainty?
• How would you separate calculation error from interpretation error in grading?

The Managerial Accounting Simulation moves from product-level economics to a constrained capital-allocation decision. Each participant compares three consumer electronic products using contribution margin, break-even, margin of safety, ROI and NPV, then allocates a $100 million budget across five opportunities through CFO, COO and CMO perspectives before making a final CEO allocation. For a BI course, it is valuable because students must select and interpret measures, reconcile conflicting executive priorities and turn metrics into a management decision.

Why it fits here

• The activity shows that different functions can read the same evidence differently because they optimise for different outcomes.
• Students cannot rely on one KPI: they must combine cost, profitability, risk, operational feasibility and market potential.
• Individual calculations, selections, reasons and allocations create a strong basis for debrief and assessment evidence.

What students do

• Calculate contribution margin, break-even measures, margin of safety, ROI and NPV for three products.
• Choose the product that offers the strongest overall opportunity and write a rationale.
• Review a $100 million capital constraint and five investment opportunities.
• Make separate CFO, COO and CMO allocations, then reconcile them in a final CEO decision.
• Compare rankings, selected products and recorded reasons with other participants in the output stage.

Recommended course placement

• After Session 7 once KPI and scorecard logic is established.
• Before Session 9 to create a bridge from metrics into prescriptive allocation.
• As a summative applied exercise when the individual rationale is paired with a short post-simulation defence.

Knowledge students should have before the simulation

• Contribution margin and break-even logic
• Basic NPV and ROI interpretation
• KPI trade-offs and scorecard thinking
• How functional priorities can conflict inside one decision

Delivery format

Suggested debrief questions

• Which KPI drove your product choice and which important measure did you deliberately not prioritise?
• How did the CFO, COO and CMO recommendations differ, and what explains the difference?
• What made the final CEO allocation defensible rather than merely an average of three views?
• Where would a dashboard help and where would it oversimplify the trade-off?
• What evidence should an individual student add to make the simulation usable for summative assessment?

The platform records decisions and comparative outcomes that can support academic judgement. It does not replace lecturer grading, and team-level simulations do not by themselves establish which individual student made which argument.

Grading with the Financial Statement Analysis Simulation

Students analyse financial statements, calculate and interpret ratios, respond to updated information and submit an individual share-price view.

• Accuracy of ratio calculations and use of the three statements
• Quality of interpretation across reporting periods
• Whether the student changes a view when evidence changes
• Ability to separate calculation error from interpretation error
• A short individual reflection or oral defence explaining the final recommendation

Grading with the Managerial Accounting Simulation

Students compare product economics and then allocate capital through CFO, COO, CMO and CEO perspectives.

• Calculation accuracy across contribution margin, break-even, ROI and NPV
• Quality of the written product-selection rationale
• Consistency between functional perspective and allocation
• Quality of the CEO reconciliation of conflicting metrics
• A post-simulation memo explaining trade-offs and rejected alternatives

Grading with the Portfolio Management Simulation

Teams use CAPM and mean-variance concepts, construct a portfolio across 25 global companies, rebalance across changing quarters and evaluate risk-adjusted performance.

• Portfolio construction and diversification logic
• Use of beta, expected return, covariance and volatility
• Quality and timing of rebalancing decisions
• Whether risk-adjusted performance matches the assigned Hedge Fund or Pension Fund mandate
• Individual report or viva outside the platform to create attributable evidence

Grading with the Capital Budgeting Simulation

Each student appraises competing projects using NPV, IRR, profitability index and payback, then allocates a fixed $10 million budget.

• Accuracy of project-appraisal calculations
• Quality of project selection and capital-rationing logic
• Value created by the chosen portfolio relative to alternatives
• Written reasons and sensitivity to qualitative project risks
• Individual assumptions note explaining why the submitted allocation is defensible

Moderation and free-riding

• Publish one common rubric before the activity.
• Moderate samples across grade bands using the same evidence pack.
• Do not use leaderboard position as a grade proxy.
• Add an individual component wherever the simulation is team-based.
• Record documented technical disruption separately from academic judgement.