Article review of “Thirty years with the balanced scorecard”

Article review of “Thirty years with the balanced scorecard”

 

How to present this article in Harvard reference format?

Tawse, A. & Tabesh, P., 2023. Thirty years with the balanced scorecard: What we have learned. Business Horizons, 66, pp.123-132. Available at: https://doi.org/10.1016/j.bushor.2022.03.005 [Accessed date].

Replace "[Accessed date]" with the date you accessed the article.

 

What are the key research issues the article wants to address?

The article identifies four key unresolved research questions regarding the balanced scorecard (BSC) and its impact on firm performance:

1.              What are the sources of ambiguity in the relationship between BSC adoption and firm performance?

2.              What is the empirical evidence regarding the relationship between BSC adoption and firm performance?

3.              What factors influence the relationship between BSC adoption and firm performance?

4.              How can practitioners maximize the benefits of BSC adoption?

These questions aim to clarify the validity and value of the BSC as a strategy implementation and performance measurement tool.

 

Describe two main theories employed in this article.

The article employs the following two main theoretical perspectives regarding the Balanced Scorecard (BSC):

1.                                  Strategic Fit and Causal Linkage Theory: Central to the BSC’s theoretical foundation is the concept of fit between strategic goals and the controllable drivers captured by BSC metrics across its four dimensions (financial, customer, internal business processes, and learning and growth). Kaplan and Norton emphasized that identifying the causal connections between these measurable drivers and long-term strategic goals is critical to BSC effectiveness. For example, aligning key process and learning metrics with a cost-leadership strategy enables effective strategy implementation by channeling organizational efforts toward desired outcomes. The establishment of a formal strategy map visually capturing these cause-and-effect relationships further strengthens this mechanism .

2.                                  Organizational Learning Theory: The BSC also operates as a platform for organizational learning, enhancing performance through feedback and the improvement of employee mental models regarding the organization's future direction. Studies cited in the article show the BSC facilitates both single-loop learning (improving information accessibility and relevance) and double-loop learning (strategy evolution), and supports vision dissemination. This learning process enables the continuous revision and improvement of organizational strategies, contributing to better alignment between strategy and operations .

These theories support the notion that the BSC improves performance by linking strategy to actionable metrics and by fostering continuous learning within organizations. However, the article also notes criticisms that question the empirical support for the causal relationships and the BSC’s flexibility in dynamic environments .

  

Highlight 2 main primary findings reported in this article.

 Two main primary findings reported in the article are:

1.              Positive but Moderate Overall Impact of BSC Adoption on Firm Performance: The meta-analysis of 11 empirical studies found that the overall relationship between Balanced Scorecard (BSC) adoption and firm performance is positive, with an aggregated effect size of 0.433. This indicates that BSC adoption contributes positively to organizational outcomes, but the impact is moderate rather than strong.

2.              The Importance of Causal Linkage and Measurement Method: The effectiveness of the BSC is significantly enhanced when causal linkage between BSC measures and strategic goals is explicitly established, typically through the use of a strategy map. The meta-analysis showed that causal linkage increases the effect size by 0.321, supporting the theoretical foundation of the BSC. Additionally, the study found that subjective measures of performance (e.g., surveys) tend to show a much higher perceived impact of the BSC (effect size = 0.747) compared to objective financial measures (effect size = 0.188), suggesting that while BSC adoption is viewed as highly beneficial internally, its direct influence on objective financial metrics may be more limited.

 

Describe 3 main claims of the article in terms of Toulmin's model of argument.

 

Using Toulmin's model of argument (which includes claim, grounds (evidence), warrant (reasoning), backing, qualifier, and rebuttal), three main claims of the article are:

Claim 1: The Balanced Scorecard (BSC) adoption positively contributes to firm performance.

·                                                Grounds: Meta-analysis of 11 empirical studies showing an aggregated effect size of 0.433 supporting a positive relationship between BSC adoption and firm performance .

·                                                Warrant: Combining multiple studies increases statistical power and reliability of results, providing a more accurate estimate than individual studies alone .

·                                                Backing: Prior literature on BSC’s role in strategy implementation and performance measurement  .

·                                                Qualifier: The impact is relevant but moderate.

·                                                Rebuttal: The effect size is moderate, and some studies show mixed results; variation exists depending on implementation details and measurement approaches .

Claim 2: The establishment of causal linkage between BSC measures and strategic goals enhances the effectiveness of BSC in improving firm performance.

·                  Grounds: Meta-analysis indicates that causal linkage increases the effect size by 0.321, showing stronger BSC effectiveness when strategy maps are used to establish causality .

·                  Warrant: Strategy maps clarify the cause-and-effect relationships, improving managerial understanding and alignment with strategic goals, which improves implementation outcomes  .

·                  Backing: Strategic management literature emphasizing the importance of linking measures to strategy for effective implementation  .

·                  Qualifier: Causal linkage significantly strengthens the BSC–performance relationship.

·                  Rebuttal: Many organizations do not implement causal linkage, potentially reducing BSC effectiveness .

Claim 3: Subjective performance measures tend to show a larger positive impact from BSC adoption than objective financial measures.

·                  Grounds: Meta-analysis results showed effect size of 0.747 for subjective performance measures versus 0.188 for objective financial data.

·                  Warrant: Subjective measures capture firm-specific strategic impacts better but may be biased due to confirmation bias of managers reporting benefits .

·                  Backing: Psychological literature on confirmation bias and methodological critiques of subjective performance measures.

·                  Qualifier: BSC is perceived to have a large impact subjectively, but direct objective effects are less pronounced.

·                  Rebuttal: Subjective assessments carry the risk of bias; objective measures may not fully capture nuanced strategic benefit.

                                                                                                                  

Describe 2 main research limitations of the study.

Two main research limitations of the study are:

1.              Small number of empirical studies included in the meta-analysis: The quantitative synthesis was based on only 11 peer-reviewed empirical studies that quantitatively measured the BSC adoption–firm performance relationship with sufficient data to calculate effect sizes. While the meta-analytic approach improves reliability by aggregating findings, the small sample size limits the generalizability and robustness of conclusions drawn from the analysis.

2.              Heterogeneity across study settings and methods: The studies included in the meta-analysis were conducted in a wide variety of industries (e.g., banking, energy, telecom) and used diverse methodological approaches (e.g., surveys and experiments). This broad diversity enhances generalizability but does not capture industry-specific nuances or contextual factors influencing BSC effectiveness, which may lead to an incomplete understanding of how different environments moderate the BSC–performance relationship.

A note on AI-augmented learning: for Accounting and Finance students doing assignment with generative AI

A note on AI-augmented learning: for Accounting and Finance students doing assignment with generative AI

 

What is AI-augmented learning? Illustrate the topic in the context of an Accounting and Finance student producing a case study [of a listed company in Hong Kong] with secondary data and then applying the balanced scorecard approach for this company.

AI-augmented learning is an approach where students work with AI as a thinking and analysis partner—using it to retrieve, structure, interrogate, and test ideas—while remaining responsible for judgment, interpretation, and final decisions. In accounting and finance education, this means AI supports case research, data handling, and framework application (such as the Balanced Scorecard) without replacing the student’s critical reasoning.

What is AI-augmented learning?

AI-augmented (or “augmented intelligence”) learning treats AI as a capability amplifier: it combines human knowledge, judgment, and creativity with AI’s speed, pattern recognition, and information-processing power. The pedagogical aim is not to offload thinking to AI (automation) but to keep the learner “in the loop,” using AI to generate feedback, alternatives, explanations, and structured drafts that the student then evaluates and revises.

In practice, AI-augmented learning in higher education shows up as:

• Personalised tutoring and Socratic questioning (e.g., Khanmigo-style assistants that probe understanding rather than give answers).
• Research and literature-support tools that help locate, summarise, and compare sources while the student checks accuracy and relevance.
• Simulation and case-based learning where AI plays roles (analyst, auditor, mentor) at defined points to challenge student work and highlight omissions.
• Formative feedback on drafts, calculations, or interpretations so students can iterate before final submission.

This aligns with evidence that AI can improve engagement and understanding when used for interactive, tailored support rather than just task execution.

Illustration: Accounting & Finance student case study on a HK listed company

Consider an Accounting and Finance student tasked with producing a case study on a Hong Kong listed company (e.g., MTR Corporation, HKEX: 66) using secondary data, then applying the Balanced Scorecard (BSC). An AI-augmented workflow might look like this.

1. Scoping the case and selecting the company

Student tasks

• Define the purpose: e.g., “Evaluate how well Company X’s strategy translates into measurable performance across financial and non-financial dimensions.”
• Set criteria for company choice: listed in HK, sufficient public disclosures (annual report, ESG/sustainability report), clear strategy narrative.

AI-augmentation

• Use an AI research assistant to:
• Generate a shortlist of HK-listed firms with strong public reporting (based on known indices like Hang Seng, known ESG reporters).
• Summarise each candidate’s business model, key segments, and recent strategic themes from annual/ESG reports.
• The student then:
• Verifies claims against the actual reports (HKEX news, company website).
• Chooses the company and refines the research question.

This keeps the student in control of selection and framing while AI accelerates scanning and synthesis.

2. Gathering and organising secondary data

Student tasks

• Identify data sources: annual reports, interim reports, ESG/sustainability reports, investor presentations, HKEX filings, reputable news and analyst commentary.
• Extract key figures: revenue, profit margins, ROE, capex, employee numbers, customer metrics, safety or service-quality indicators, ESG KPIs.

AI-augmentation

• Use AI to:
• Create a data-extraction template aligned with the BSC perspectives (Financial, Customer, Internal Process, Learning & Growth).
• Summarise long documents (e.g., “Extract all KPIs mentioned in the 2024 Annual Report and classify by BSC perspective”).
• Draft a structured evidence table: year, KPI, value, source page, notes on definition.
• The student then:
• Checks extracted numbers against the original PDFs.
• Resolves ambiguities (e.g., different definitions of “operating profit”).
• Adds context (one-off items, accounting policy changes).

Here AI reduces mechanical reading and structuring work; the student ensures accuracy and interprets accounting choices.

3. Applying the Balanced Scorecard framework

The Balanced Scorecard links strategy to performance via four perspectives:

• Financial: profitability, growth, risk, capital efficiency.
• Customer: satisfaction, retention, market share, service quality.
• Internal processes: operational efficiency, innovation, safety, cycle times.
• Learning & growth: employee capability, culture, systems, digitalisation.

Student tasks

• Map the company’s stated strategy (from MD&A, CEO statements) to BSC perspectives.
• Select 2–4 KPIs per perspective that best reflect that strategy.
• Analyse trends (3–5 years) and relationships (e.g., does investment in training correlate with service improvements and then financial results?).

AI-augmentation

• Use AI to:
• Propose an initial BSC map: “Given this strategy text, suggest plausible BSC objectives and KPIs for a transport/utility company.”
• Draft cause-and-effect logic: “Explain how improved employee training could lead to fewer service disruptions and higher customer satisfaction, then better financial performance.”
• Generate visual descriptions for a strategy map (student later draws it in PowerPoint or similar).
• Run “challenge” prompts: “Identify weaknesses or missing links in this BSC design for Company X.”
• The student then:
• Critiques AI suggestions against the actual business model and industry realities.
• Adjusts KPIs (e.g., adds on-time performance for a transport firm; adds digital adoption metrics if relevant).
• Writes the interpretive narrative, using AI-drafted text only as a starting point.

This mirrors role-based AI frameworks where AI acts as analyst/mentor at defined points, but students must detect omissions and construct their own assessment.

4. Analysis, insight, and recommendations

Student tasks

• Interpret patterns: where does performance support or contradict the stated strategy?
• Identify tensions: e.g., cost-cutting vs. service quality; rapid digitalisation vs. employee capability gaps.
• Formulate recommendations: strategic, operational, or measurement-related.

AI-augmentation

• Use AI to:
• Summarise key findings from the student’s own notes into bullet points.
• Generate alternative interpretations: “Give three different explanations for the decline in operating margin despite revenue growth.”
• Stress-test recommendations: “List potential risks and unintended consequences of recommending higher capex in digital systems.”
• The student then:
• Selects and refines the most convincing arguments.
• Grounds recommendations in evidence from the reports and BSC analysis.
• Ensures the voice and critical stance are their own.

This preserves human judgment while using AI to broaden the set of considered explanations and risks.

5. Academic integrity and learning value

To keep the work genuinely AI-augmented (not AI-automated):

• The student:
• Defines the research question, company choice, and analytical lens.
• Verifies all data and accepts responsibility for accuracy.
• Makes the final interpretive and evaluative judgments.
• AI is used as:
• A research and structuring assistant.
• A challenger and idea generator.
• A formative feedback tool before finalising the case.

This approach reflects current thinking that AI in education should enhance, not replace, the cognitive work that builds expertise.

 

Identify 3 things a student doing AI-augmented MUST not do; otherwise, using generative AI can be considered as dishonest academic behavior and is also not good for genuine intellectual learning.

For AI-augmented learning to remain honest and educationally valuable, students must avoid these three behaviours. Crossing any of them typically counts as academic misconduct and also undermines genuine intellectual development.

1. Do not submit AI‑generated work as your own original output

Students must not:

·        Paste AI‑written text, analysis, code, tables, or diagrams into an assignment and present it as if they wrote it themselves.

·        Paraphrase AI output just enough to “hide” it, then claim the ideas and wording are theirs.

·        Use AI to produce whole sections (literature review, case analysis, recommendations) without meaningful human authorship.

Why this is dishonest and harmful:

·        It misrepresents the student’s knowledge and abilities, which is the core of plagiarism/cheating policies.

·        It bypasses the cognitive work (reading, reasoning, structuring arguments) that builds expertise, so learning gains are minimal.

·        Many policies explicitly state that submitting AI‑generated content as your own is academic misconduct, even if the tool is not mentioned by name.

AI‑augmented alternative: use AI to brainstorm, outline, or draft working material, then substantially rewrite, extend, and integrate it with your own analysis, data, and voice—and disclose AI use where required.

2. Do not use AI in ways that violate the instructor’s rules or the assessment’s learning objectives

Students must not:

·        Use generative AI on an assignment or exam where the instructor has forbidden it (including “no AI at any stage” rules).

·        Employ AI to gain an unfair advantage (e.g., getting model answers, full solutions, or tailored exam help) when the task is designed to assess unaided competence.

·        Ignore explicit limits such as “AI allowed for brainstorming only” or “no AI for data analysis or writing.”

Why this is dishonest and harmful:

·        Institutions treat breaches of stated AI rules as academic integrity violations, equivalent to other forms of cheating.

·        It undermines the learning objectives the assessment is meant to measure (e.g., independent problem‑solving, critical evaluation).

·        It creates an unfair advantage over peers who follow the rules.

AI‑augmented alternative: check the syllabus and assignment brief; if unclear, ask the instructor. Where AI is permitted, use it only for the allowed purposes and at the allowed stages, and be ready to explain and defend your work orally if asked.

3. Do not fail to disclose, cite, and verify AI‑assisted content (including fabricated references)

Students must not:

·        Use AI to generate ideas, structure, text, data summaries, or references and then provide no acknowledgement that AI was used (when AI use is permitted/expected to be disclosed).

·        Copy AI‑produced citations or references without checking whether the sources exist and are accurately described.

·        Submit work containing AI “hallucinations” (fake cases, non‑existent articles, incorrect statistics) without verifying them.

Why this is dishonest and harmful:

·        Non‑disclosure of substantial AI assistance can be treated as plagiarism or misrepresentation of originality.

·        Fabricated or unverified references are a form of scholarly misconduct; students are responsible for the accuracy of everything they submit.

·        Relying on unchecked AI output erodes research skills and critical evaluation, which are central to higher‑education learning.

AI‑augmented alternative: where allowed, briefly state how you used AI (e.g., “used to draft an initial BSC map and suggest KPIs; all analysis and final text are my own”), cite AI outputs per your institution’s guidance, and personally verify every fact, figure, and reference before submission.