9. Value Governance, Performance, and Strategic Analytics

Introduction and Context

Information technology governance becomes meaningful only when it demonstrates how technology decisions create, protect, and sustain organizational value. Investment approval, architecture standards, cybersecurity controls, service management, and data policies are not ends in themselves. They are mechanisms through which an organization seeks outcomes such as revenue growth, operational resilience, regulatory compliance, improved citizen services, research capability, or better customer experience. Value governance connects these mechanisms to measurable strategic results.

This chapter examines the analytical and managerial systems used to govern value: value frameworks, objectives and key results (OKRs), balanced scorecards, key performance indicators (KPIs), dashboards, maturity assessments, benchmarking, benefits realization, and causal analysis. The central concern is not simply whether an activity was completed or whether a target was reached. It is whether the selected measures represent strategically important outcomes, whether performance can be interpreted fairly, and whether decisions based on the evidence improve organizational value.

Measurement is therefore a governance activity. Poor measures can encourage local optimization, conceal risk, reward undesirable behaviour, and create the illusion of control. Strong measurement systems make assumptions explicit, distinguish outputs from outcomes, identify trade-offs, and support timely intervention. At doctoral level, performance analytics must also be treated as an epistemic problem: what is being claimed, what evidence supports the claim, what alternative explanations exist, and how confident should decision-makers be?

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Explain the relationship between IT governance, value creation, performance, and benefits realization.

2. Distinguish inputs, activities, outputs, outcomes, impacts, leading indicators, and lagging indicators.

3. Design and critique value frameworks, OKRs, balanced scorecards, KPIs, and executive dashboards.

4. Apply maturity assessment, benchmarking, and causal analytics to technology performance problems.

5. Evaluate measurement risks, data-quality issues, unintended consequences, and ethical concerns.

6. Develop an evidence-based performance governance approach for a complex organization.

Value as a Governance Construct

Value is not synonymous with financial return. In different organizational settings, value may include profit, cost avoidance, service accessibility, safety, trust, compliance, innovation capacity, employee capability, or social benefit. A technology investment can create strategic value indirectly. For example, an identity platform may not produce revenue by itself, but it can reduce fraud, improve user access, support regulatory compliance, and enable new digital services.

A useful value chain separates five levels:

• Inputs: financial resources, people, data, infrastructure, and executive attention.
• Activities: development, implementation, training, process redesign, and support.
• Outputs: deployed systems, completed transactions, trained users, or available reports.
• Outcomes: changed behaviour, faster service, lower error rates, stronger controls, or increased adoption.
• Impacts: sustained strategic, financial, social, or institutional consequences.

Confusing these levels is a major source of weak governance. Delivering a system is an output; realizing shorter processing time is an outcome. A high system-availability percentage does not prove that the system is valuable if users avoid it or if the underlying process remains ineffective.

Value governance should also recognize multiple stakeholders. A measure that signals value to the finance function may not capture value for customers, regulators, employees, or communities. Stakeholder value mapping makes these interests visible and helps prevent a narrow financial interpretation of performance.

Theories and Frameworks for Governing Value

Benefits Realization Management

Benefits realization management treats benefits as planned improvements that require ownership, enabling conditions, measurement, and sustained adoption. A benefits dependency network links investment objectives to business changes, enabling capabilities, measurable benefits, and strategic goals. The network helps decision-makers identify assumptions. For instance, a mobile application may be expected to reduce call-centre demand, but that benefit depends on usability, customer awareness, process integration, and sufficient transaction functionality.

Benefits should have a named owner, a baseline, a target, a realization date, and a method of verification. Benefits may be financial or non-financial, direct or indirect, tangible or intangible. Avoided costs should not automatically be reported as cash savings unless spending actually decreases or capacity is redeployed.

Balanced Scorecard Logic

The balanced scorecard translates strategy into a set of linked objectives and measures. Its traditional perspectives are financial, customer or stakeholder, internal process, and learning and growth. In an IT governance context, these can be adapted without losing the underlying logic. For example, learning and growth may include engineering capability and data literacy; internal process may include release reliability and incident recovery; customer may include user trust and service accessibility; financial may include realized benefits and cost efficiency.

The scorecard is most useful when objectives are connected through explicit strategic hypotheses. Improved staff capability may enable better architectural practices, which may improve service reliability, which may increase user trust and reduce avoidable operating costs. The scorecard should not become a catalogue of unrelated metrics.

Objectives and Key Results

OKRs express an important objective and a small number of measurable key results. The objective is directional and meaningful; key results specify observable evidence of progress. “Improve digital service trust” is an objective. “Increase successful first-time completion from 72 to 88 percent” and “reduce severe privacy incidents from four to zero per quarter” are possible key results.

OKRs encourage focus, transparency, and periodic review. They are not equivalent to task lists. A key result should measure a changed condition, not merely the completion of an activity. Nor should every key result be treated as a rigid performance-contract target. In innovative work, ambitious stretch objectives can be valuable, but governance must distinguish learning from failure and prevent teams from lowering ambition to protect evaluations.

Performance Prism and Stakeholder-Centred Views

Stakeholder-centred frameworks ask five related questions: who are the stakeholders, what do they want, what does the organization want from them, what processes create value, and what capabilities sustain those processes? This perspective is particularly relevant to public services, universities, healthcare, and regulated industries, where legitimacy and trust are part of performance. It also counters the tendency to optimize internal IT metrics while ignoring the experience of service recipients.

Designing Measures That Matter

A KPI is a defined indicator used to monitor performance against a meaningful objective. A metric is any quantified measure; not every metric deserves KPI status. A good KPI has a clear purpose, operational definition, owner, data source, frequency, target or reference range, and escalation rule. Its interpretation should also specify the population, time period, exclusions, and calculation method.

Measures differ by temporal role. Leading indicators provide early signals of future performance, such as phishing-test participation, backlog age, test coverage, or adoption of a control. Lagging indicators record outcomes after they occur, such as customer retention, realized savings, service interruptions, or regulatory findings. A mature system uses both. Leading indicators are actionable but can be weakly related to outcomes; lagging indicators are consequential but may arrive too late for intervention.

Measures also differ by level. Strategic indicators show contribution to organizational objectives. Tactical indicators guide portfolios, products, and capabilities. Operational indicators manage daily service performance. A common governance failure is to aggregate operational activity measures into an executive dashboard without explaining strategic significance.

A KPI design test can ask:

1. What decision will this measure inform?

2. What objective or stakeholder need does it represent?

3. Is the definition stable and reproducible?

4. Is the data complete, timely, and sufficiently accurate?

5. Can the accountable team influence it without manipulating it?

6. What trade-offs or harms could the target create?

7. What action follows if performance is above, below, or ambiguously near target?

Targets require care. A target can be absolute, relative to a baseline, benchmarked externally, or expressed as a control limit. A single target may conceal volatility and distributional differences. Mean resolution time, for example, may improve while a small group of vulnerable users experiences much longer delays. Distributions, percentiles, failure severity, and equity dimensions may therefore be necessary.

Dashboards as Decision Systems

A dashboard is not merely a visual display. It is an interface between evidence and governance decisions. Effective dashboards provide a hierarchy: a small number of strategic signals at the top, diagnostic views beneath them, and links to underlying data and commentary. Traffic-light colours can aid rapid scanning, but they should not replace explanation. A red indicator should identify magnitude, cause, owner, and proposed response.

Executive dashboards should distinguish actual performance from forecast performance, baseline from target, and correlation from attribution. They should display trends rather than isolated values and make data freshness visible. Drill-down capability allows leaders to move from an enterprise result to a portfolio, service, location, demographic group, or process stage.

Dashboard governance includes data ownership, metric definitions, access controls, audit trails, version management, and review of obsolete indicators. Automated dashboards can spread errors rapidly when source systems change or data pipelines silently fail. A visually sophisticated dashboard with an unvalidated data model is a governance risk rather than an analytical achievement.

Maturity Assessment and Benchmarking

Maturity models describe progressive capability, often moving from ad hoc practice to defined, measured, managed, and optimized practice. They can assess governance domains such as benefits management, service measurement, portfolio prioritization, data quality, architecture, or cybersecurity. Maturity assessment is valuable when it identifies capability gaps and informs a practical roadmap.

However, maturity levels are not universal laws. A highly mature process may be inappropriate where uncertainty is high or the cost of formalization exceeds its benefit. Assessments can also become compliance rituals in which organizations seek a higher score rather than better outcomes. Evidence should include artefacts, observed behaviour, stakeholder experience, and results—not only policy documents.

Benchmarking compares performance or capability with a reference group. Internal benchmarking may compare business units; external benchmarking may compare peer organizations; functional benchmarking may compare similar processes across sectors. Comparability is essential. Differences in service complexity, regulation, geography, architecture, risk appetite, accounting treatment, and customer population can make a numerical comparison misleading.

Benchmarking should therefore ask what is being normalized, whether definitions match, and whether the comparison identifies a transferable practice. A lower cost per transaction may reflect superior efficiency, lower service quality, or exclusion of complex users. Benchmarking is a source of questions, not automatic proof of superiority.

Causal Analytics and Strategic Learning

Descriptive analytics answers what happened. Diagnostic analytics explores why it happened. Predictive analytics estimates what may happen. Prescriptive analytics evaluates possible actions. Strategic governance requires movement across these forms while maintaining appropriate caution about causal claims.

A correlation between a new digital channel and improved customer satisfaction does not establish that the channel caused the improvement. Other explanations may include seasonal demand, policy changes, user self-selection, or unrelated service improvements. Causal analysis begins with a theory of change and identifies the intervention, outcome, affected population, comparison condition, timing, and possible confounders.

Where feasible, randomized experiments provide strong evidence, although they may be impractical or unethical for critical services. Quasi-experimental approaches include difference-in-differences, interrupted time-series analysis, regression discontinuity, matched comparison groups, and controlled pilots. These methods require assumptions that should be reported rather than hidden behind statistical terminology.

For example, difference-in-differences compares changes over time between an adopting group and a comparable non-adopting group. Its credibility depends partly on whether the groups had similar trends before intervention. Causal diagrams can help analysts distinguish confounders, mediators, and colliders. Qualitative evidence remains important for understanding mechanisms, implementation context, and effects not captured in administrative data.

Analytics should support learning rather than merely justify a decision already made. Governance forums should ask what evidence would change the decision, how uncertainty is represented, and whether observed benefits persist after initial implementation.

An Illustrative Case: Governing the Value of a University Digital Student Platform

Background

The fictional Northbridge University approved a three-year programme to replace fragmented student portals with an integrated digital platform. The strategic case predicted improved student experience, reduced administrative workload, better retention, and stronger compliance with data-protection requirements.

Situation and Problem

After the first year, the programme reported that the platform had been delivered on schedule and that 93 percent of students had logged in. Yet complaints about enrolment remained high, help-desk contacts had increased, and senior leaders disagreed about whether the programme was creating value. The programme dashboard emphasized delivery milestones, uptime, and login counts.

Analysis

The governance review separated outputs from outcomes. Deployment and login were outputs or adoption signals, not proof of improved service. Interviews and process data showed that students could access information but still had to use separate systems to complete key transactions. The login rate also concealed unequal adoption among international students and students using mobile devices.

The review created a benefits dependency network. The principal benefit—faster and more reliable enrolment—depended on process redesign, integrated identity, accurate course data, staff training, and a usable mobile interface. Baseline measures were established: median enrolment completion time, failed transaction rate, repeat contacts, accessibility defects, and satisfaction by student group.

An adapted scorecard linked four perspectives. The stakeholder perspective included successful first-time completion and trust. The internal-process perspective included data-error rates and time to resolve enrolment defects. Learning and growth included staff capability in service design and analytics. The financial perspective included avoidable administrative effort rather than unsupported claims of cash savings.

The university introduced quarterly OKRs. One objective was to make enrolment dependable. Key results included reducing the 90th-percentile completion time by 35 percent, reducing failed submissions by half, and closing critical accessibility defects within agreed periods. An analytics team used an interrupted time-series design for the redesigned process and compared results across faculties implementing the change at different times. Qualitative interviews were used to explain why improvements varied.

Approach and Outcome

After two review cycles, completion time and failed submissions improved substantially in faculties that had completed process redesign. The university did not attribute all improvement to the platform; evidence suggested that redesign and staff support were necessary complements. Overall satisfaction rose modestly, but the gap between mobile and desktop users remained. The governance committee therefore funded mobile optimization rather than declaring the programme complete.

The revised dashboard displayed outcome trends, confidence notes, data-quality warnings, and distributional effects. Benefits owners remained accountable after technical delivery. Some anticipated workload savings were reclassified as capacity released for advising rather than financial savings.

Learning Points

The case demonstrates five principles. First, delivery does not equal value realization. Second, measures should follow a theory of change. Third, adoption must be examined across stakeholder groups. Fourth, causal claims require comparison and explanation. Fifth, governance should preserve uncertainty and redirect investment when evidence reveals unmet conditions.

Risks, Limitations, and Ethical Considerations

Performance systems can produce gaming. Teams may close incidents prematurely, avoid complex users, defer risky work, or redefine categories to meet targets. Balanced measures, independent validation, and attention to unintended effects reduce—but do not eliminate—these risks.

Data quality problems include missingness, inconsistent definitions, duplication, delayed records, measurement drift, and changes in source systems. A KPI should carry a confidence or quality annotation where appropriate. False precision is especially dangerous when leaders treat a number as objective simply because it has decimal places.

Attribution is another limitation. Technology outcomes are usually produced by interacting capabilities, policies, people, and external conditions. Governance should avoid assigning credit or blame beyond what evidence supports. Forecast models can also reproduce historical inequities. If performance is reported by demographic or geographic group, privacy protection, legitimate purpose, proportionality, and avoidance of stigmatizing interpretation are essential.

Regulatory and professional obligations may include data protection, records management, accessibility, auditability, cybersecurity, sector-specific reporting, and explainability. Dashboard access must follow least-privilege principles, and sensitive performance data should not be exposed merely for convenience. Automated decision systems require human oversight, documentation of model limitations, monitoring for drift, and channels for contesting harmful outcomes.

Emerging Directions and Best Practice

Organizations increasingly combine observability data, financial information, user research, process mining, and strategic planning data. Process mining can reveal bottlenecks between system events; natural-language analysis can identify recurring service concerns; machine learning can forecast demand or incident risk. These applications are valuable only when their data-generating processes and decision consequences are understood.

A contemporary value-governance practice should:

• begin with strategic outcomes and stakeholder needs rather than available data;
• maintain a small, coherent set of decision-relevant measures;
• connect benefits, objectives, KPIs, owners, and governance forums;
• combine leading, lagging, qualitative, quantitative, and distributional evidence;
• publish definitions, assumptions, baselines, and data-quality limitations;
• review causal logic and unintended consequences;
• treat dashboards as controlled decision products;
• reassess measures when strategy, technology, or context changes; and
• use evidence to learn and adapt, not merely to defend prior investments.

The mature organization is not the one with the greatest number of metrics. It is the one that can explain how evidence informs a decision, how value is expected to arise, what uncertainty remains, and who is accountable for acting.

Chapter Summary

Value governance connects IT investments and capabilities to stakeholder outcomes and strategic impact. Effective systems distinguish inputs, activities, outputs, outcomes, and impacts; assign benefits ownership; and make strategic assumptions visible. Benefits realization management, balanced scorecards, OKRs, KPI hierarchies, dashboards, maturity models, and benchmarking provide complementary tools rather than interchangeable solutions.

High-quality performance governance depends on valid definitions, credible baselines, appropriate targets, reliable data, and attention to trade-offs. Leading and lagging indicators should be combined, while dashboards should support decisions rather than merely display information. Causal analytics strengthens claims about contribution, but assumptions and uncertainty must remain visible. Ethical governance requires protection against gaming, unfair treatment, privacy loss, unsupported attribution, and automated discrimination. Ultimately, performance measurement is a learning and accountability system through which organizations decide whether to continue, adapt, stop, or redirect technology investments.

Key Terms

Value governance: The structures and practices used to direct, evaluate, and sustain value from organizational resources and investments.

Benefits realization management: The coordinated identification, ownership, measurement, and delivery of intended benefits.

Value chain: A sequence linking inputs and activities to outputs, outcomes, and impacts.

Output: A directly produced deliverable or product of an activity.

Outcome: A change in behaviour, condition, capability, or performance resulting from outputs.

Impact: A broader and often longer-term strategic, social, or financial consequence.

Objective and Key Results (OKR): A management approach pairing a directional objective with measurable results.

Balanced scorecard: A strategy-performance framework linking objectives and measures across multiple perspectives.

Key performance indicator (KPI): A strategically relevant measure used to monitor and govern performance.

Leading indicator: A measure that provides an early signal of future performance.

Lagging indicator: A measure that records an outcome after it has occurred.

Dashboard: A governed visual interface that presents evidence for monitoring and decision-making.

Maturity model: A framework describing progressive levels of organizational capability.

Benchmarking: Systematic comparison with an internal, external, or functional reference group.

Theory of change: An explicit explanation of how activities and enabling conditions are expected to produce outcomes.

Causal analytics: Methods used to estimate whether and how an intervention produces an effect.

Attribution: The reasoned assignment of an observed result to a particular intervention or contributor.

Data quality: The degree to which data is accurate, complete, timely, consistent, valid, and fit for purpose.

Measurement gaming: Behaviour that manipulates activities, definitions, or reporting to improve a measured result without improving the underlying objective.

Review Questions

Short Answer Questions

1. Distinguish an output from an outcome and provide an IT-related example of each.

2. What is the principal purpose of benefits realization management?

3. Explain the difference between a leading indicator and a lagging indicator.

4. Why should dashboard data freshness and metric definitions be governed?

5. What is meant by measurement gaming?

Long Answer Questions

1. Explain how a balanced scorecard can connect IT capabilities to organizational strategy.

2. Discuss the design characteristics of a strategically useful KPI system.

3. Compare OKRs and benefits realization management as mechanisms of value governance.

4. Explain the appropriate uses and limitations of maturity assessment and benchmarking.

5. Discuss how causal analytics can improve decisions about technology investments.

Analytical / Application Questions

1. An organization reports high cloud availability but declining customer satisfaction. Design an analytical approach that investigates whether the availability KPI represents service value.

2. Construct a value chain, two OKRs, and a balanced-scorecard logic for a cybersecurity awareness programme. Identify possible unintended consequences.

3. A digital transformation programme reports savings based on reduced manual effort, but staffing costs have not changed. Evaluate the claim and recommend a benefits-verification method.

22. Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes an IT output from an IT outcome?

A. An output is a long-term strategic effect, while an outcome is a project cost.

B. An output is a delivered product or capability, while an outcome is a resulting change in performance or behaviour.

C. An output is always financial, while an outcome is always non-financial.

D. An output is a stakeholder opinion, while an outcome is a technical specification.

Correct Answer: B. An output is a delivered product or capability, while an outcome is a resulting change in performance or behaviour.

Question 2

A university measures the percentage of students who log into a new portal but wants to know whether the portal improves enrolment. What is the main analytical problem?

A. Login percentage is necessarily a lagging indicator.

B. The portal cannot be measured with quantitative data.

C. Adoption is being treated as evidence of an outcome without measuring the intended benefit.

D. Enrolment outcomes should be excluded from IT governance.

Correct Answer: C. Adoption is being treated as evidence of an outcome without measuring the intended benefit.

Question 3

Which feature most strongly indicates that an OKR is well designed?

A. It lists all project activities in chronological order.

B. It uses only financial measures.

C. It links a meaningful directional objective to observable changes in performance.

D. It guarantees that every target will be achieved.

Correct Answer: C. It links a meaningful directional objective to observable changes in performance.

Question 4

Why should a balanced scorecard use linked objectives rather than a collection of unrelated metrics?

A. Linked objectives express the organization’s theory of how capabilities produce outcomes.

B. Unrelated metrics are always more accurate.

C. Linked objectives eliminate the need for stakeholder measures.

D. A scorecard can contain only four metrics.

Correct Answer: A. Linked objectives express the organization’s theory of how capabilities produce outcomes.

Question 5

A team reduces its reported incident-resolution time by closing complex incidents and reopening them later. This is an example of:

A. Causal inference

B. Measurement gaming

C. External benchmarking

D. Data normalization

Correct Answer: B. Measurement gaming

Question 6

Which statement about benchmarking is most defensible?

A. A lower numerical cost always demonstrates superior performance.

B. Benchmarking is valid even when definitions and service complexity differ.

C. Benchmarking should examine comparability and generate questions rather than automatic conclusions.

D. External benchmarks are always more useful than internal benchmarks.

Correct Answer: C. Benchmarking should examine comparability and generate questions rather than automatic conclusions.

Question 7

An organization compares performance changes in units that adopted a new system with changes in similar units that did not. Which causal approach does this most closely resemble?

A. Difference-in-differences

B. Randomized controlled trial

C. Simple descriptive reporting

D. Maturity scoring

Correct Answer: A. Difference-in-differences

Question 8

What should an executive dashboard do when a KPI turns red?

A. Remove the KPI immediately.

B. Identify the magnitude, likely cause, accountable owner, and possible response.

C. Replace the KPI with a more favourable measure.

D. Treat the result as proof of technology failure.

Correct Answer: B. Identify the magnitude, likely cause, accountable owner, and possible response.

Question 9

Why might a claimed reduction in manual effort not qualify as realized financial savings?

A. Effort can never be measured.

B. Financial savings require that expenditure actually decreases or capacity is demonstrably redeployed.

C. Non-financial benefits are invalid.

D. Only external auditors can measure effort.

Correct Answer: B. Financial savings require that expenditure actually decreases or capacity is demonstrably redeployed.

Question 10

Which practice best addresses ethical risks in performance analytics?

A. Maximizing the number of data fields collected regardless of purpose.

B. Hiding uncertainty to make dashboards easier to interpret.

C. Monitoring distributional effects, protecting privacy, and documenting limitations and assumptions.

D. Using automated decisions without human review.

Correct Answer: C. Monitoring distributional effects, protecting privacy, and documenting limitations and assumptions.