Data Strategy and Business Value Creation

Introduction

Data has become a strategic resource rather than merely an operational by-product. Organisations use data to understand customers, improve processes, manage risk, develop products, automate decisions and create digital ecosystems. However, possessing large volumes of data does not automatically create value. Value emerges when data is relevant, trustworthy, accessible, governed appropriately and connected to organisational capabilities and strategic decisions.

A data strategy is a coordinated set of choices concerning how an organisation acquires, manages, analyses, protects, shares and exploits data in support of business objectives. It connects corporate strategy with data assets, technology, people, processes, governance and measurable outcomes. A strong strategy therefore asks not only “What data do we have?” but also “Which strategic problems should data help us solve, and how will the resulting value be realised?”

This chapter examines data strategy through three complementary perspectives: the value-chain perspective, the resource-based view of the firm and the capability perspective. It then applies these perspectives to data products, monetisation, competitive advantage and platform business models.

Learning Objectives

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

1. Explain the role of data strategy in corporate and competitive strategy.
2. Analyse data-related activities using value-chain, resource-based and capability perspectives.
3. Design a data strategy aligned with business objectives.
4. Distinguish data assets, information products and data products.
5. Evaluate direct and indirect data monetisation opportunities.
6. Assess whether data can create sustainable competitive advantage.
7. Explain the economics and governance requirements of platform business models.
8. Recommend practical governance, measurement and implementation approaches.

1. From Corporate Strategy to Data Strategy

Corporate strategy defines the organisation’s choices about markets, customers, products, geographic scope and sources of advantage. Data strategy translates these choices into decisions about information. For example, a retailer pursuing customer intimacy may prioritise customer identity resolution, recommendation data and consent management. A manufacturer competing through operational excellence may prioritise sensor data, predictive maintenance and process analytics.

Data strategy should therefore begin with business outcomes rather than technology procurement. A useful alignment chain is:

Strategic objective → business decision or process → required data → analytical capability → action → measurable value.

For example:

Reduce delivery failures → improve route and exception decisions → integrate order, traffic and location data → develop prediction and optimisation capability → reroute deliveries → reduce cost and improve customer satisfaction.

A data strategy normally addresses six questions:

A common failure is to define a “data strategy” as a technology architecture alone. Data warehouses, data lakes, cloud services and artificial intelligence may be important enablers, but they are not strategic outcomes. Strategy requires prioritisation and choices, including what not to collect, build or share.

2. Data as a Strategic Resource

Data differs from many traditional resources. It can often be copied without being depleted, but its value may decline through age, duplication, inaccuracy, privacy restrictions or changing market conditions. Data also becomes more valuable when combined with other data and capabilities. A customer record, for instance, may have limited value in isolation but become highly useful when linked to transactions, service interactions and consent preferences.

Important characteristics include:

• Replicability: Digital data can usually be copied at low marginal cost.
• Non-rivalry: One department’s use does not necessarily prevent another’s use.
• Context dependence: Meaning and usefulness depend on business context.
• Network effects: Data generated by users can improve a service, attracting more users and generating further data.
• Accumulation: Historical data can support forecasting, benchmarking and model training.
• Risk sensitivity: Poor data can produce incorrect decisions, discrimination, security incidents or regulatory breaches.
• Perishability: Data may lose relevance as circumstances change.

Data should not be treated as automatically valuable simply because it is large. A practical data-value assessment considers relevance, quality, timeliness, uniqueness, accessibility, legality, scarcity and the organisation’s ability to use it.

3. The Value-Chain Perspective

The value-chain perspective examines how activities contribute to value creation. Originally developed for analysing business activities, it can be adapted to data by tracing the movement from data generation to business action.

A data value chain commonly contains the following stages:

1. Generation: Data is produced through transactions, sensors, applications, employees, customers, partners and public sources.
2. Acquisition: Data is collected or obtained through systems, contracts, APIs, surveys or partnerships.
3. Storage and integration: Data is stored, catalogued, linked and made interoperable.
4. Preparation and quality management: Errors, duplicates, missing values and inconsistent definitions are addressed.
5. Analysis and modelling: Descriptive, diagnostic, predictive or prescriptive methods are applied.
6. Productisation: Insights are embedded in reports, applications, APIs, workflows or automated services.
7. Decision and action: Employees, customers or systems use the output.
8. Feedback and learning: Outcomes are measured and fed back into processes, data and models.

Value is not created equally at every stage. A sophisticated data platform may create little value if its outputs are not adopted by decision makers. Conversely, a modest dataset can be highly valuable when it directly improves an important decision.

Applying the Value Chain

Managers can analyse each activity by asking:

• What input is required?
• Who owns the activity?
• What quality standard applies?
• What cost, delay or risk is introduced?
• What output is produced?
• Who uses the output and what decision does it improve?

This analysis often reveals bottlenecks. A bank may have advanced fraud models but unreliable customer identifiers. A hospital may collect extensive clinical data but lack interoperability. A retailer may have accurate sales data but fail to make inventory recommendations available at the point of replenishment.

The value-chain perspective is especially useful for identifying investment priorities. It prevents an organisation from concentrating only on analytics while neglecting data capture, governance, integration, adoption and feedback.

4. The Resource-Based View and Data Advantage

The resource-based view argues that sustained competitive advantage depends on resources and capabilities that are valuable, rare, difficult to imitate and difficult to substitute. This logic is often summarised as the VRIN framework:

Some data is widely available and therefore provides little lasting advantage. Public demographic information, for example, may be useful but not rare. In contrast, a firm may possess unique longitudinal data generated through trusted customer relationships, proprietary operations or a specialised network.

Yet data alone is rarely sufficient for sustained advantage. Competitors may purchase similar data, hire analysts or reproduce an algorithm. More defensible advantage usually results from combinations of:

• proprietary or privileged data;
• high-quality data governance;
• specialised analytical models;
• domain knowledge;
• embedded workflows;
• trusted customer relationships;
• rapid experimentation and learning;
• organisational ability to act on insights.

The resource-based view also highlights path dependence. A firm that has accumulated high-quality data over many years, while maintaining customer trust and consistent definitions, may be difficult to imitate even if competitors can observe its products.

The VRIO Extension

The VRIO framework adds the question of organisation: can the firm organise itself to exploit the resource? A company may possess valuable data but fail to benefit because ownership is unclear, systems are disconnected or incentives discourage sharing. Thus, data advantage requires governance, skills, leadership and operating models in addition to data holdings.

5. Capability Perspectives

A capability is the organisation’s ability to deploy resources through repeatable processes. Data capabilities include more than technical competence. They combine people, processes, technology, governance and managerial routines.

Core data capabilities include:

• Data governance: Decision rights, policies, standards, accountability and issue resolution.
• Data architecture: Structures and technologies that support storage, integration, access and processing.
• Data engineering: Reliable pipelines, transformation, orchestration and monitoring.
• Data literacy: The ability of employees to interpret and use data appropriately.
• Analytical capability: Statistical, optimisation, machine-learning and visualisation competence.
• Product management: The ability to convert data and insights into usable products.
• Data ethics and privacy: Processes for lawful, fair, secure and responsible use.
• Change and adoption: The ability to alter workflows, incentives and behaviour.

Dynamic capabilities are especially important in volatile environments. They involve sensing opportunities, seizing them through investments and reconfiguring resources as conditions change. For example, an organisation may detect a new fraud pattern, develop a model, integrate it into authorisation workflows and revise it as criminals adapt.

A mature data strategy treats capability development as an investment portfolio. Some capabilities are foundational, such as identity management and data quality. Others are differentiating, such as real-time personalisation or advanced optimisation. Priorities should reflect business needs rather than fashionable technologies.

6. Designing an Aligned Data Strategy

A practical strategy-development process can be organised into seven stages.

Stage 1: Clarify strategic intent

Translate corporate goals into measurable outcomes. Examples include increasing retention, reducing working capital, improving safety, entering a new market or meeting regulatory obligations.

Stage 2: Identify high-value decisions and use cases

List decisions that are frequent, consequential, improvable and actionable. A use case should identify its user, decision, required data, intervention and expected outcome. “Use artificial intelligence in marketing” is too vague; “rank next-best offers for consenting customers to improve conversion without increasing complaint rates” is more actionable.

Stage 3: Assess the current state

Evaluate data availability, quality, architecture, skills, governance, culture, security and adoption. Maturity assessments should identify constraints, not simply assign labels.

Stage 4: Define the target operating model

Specify responsibilities for data owners, stewards, custodians, product managers, analysts and decision makers. Decide how central teams and business units will collaborate. Common models include centralised, federated and hub-and-spoke arrangements.

Stage 5: Prioritise initiatives

Prioritisation should consider expected value, feasibility, risk, strategic fit, dependency and time to benefit. A simple scoring model may be useful, but scores should support managerial judgement rather than replace it.

Stage 6: Build a roadmap

Sequence foundational work and visible use cases. Early projects should demonstrate value while establishing reusable capabilities. A roadmap may include identity resolution, metadata, data quality controls, privacy processes, analytical products and platform improvements.

Stage 7: Measure and adapt

Track adoption and outcomes, not merely technology delivery. Review assumptions as markets, regulations and business priorities change.

7. Data Products and Information Products

An information product is a structured output that supports a defined user or decision. A data product is a more complete product-oriented capability: it has identifiable users, a purpose, an owner, defined quality expectations, documentation, access controls, a delivery interface and a mechanism for feedback and improvement.

Examples include:

• a real-time inventory availability service;
• a customer lifetime-value dataset;
• a fraud-risk API;
• a supplier-performance dashboard;
• a route-optimisation application;
• a curated feature set for approved machine-learning models;
• a benchmark sold to external customers.

A data product should be evaluated across several dimensions:

Product thinking changes the focus from delivering a dataset once to managing a service over its life cycle. It also encourages explicit service-level expectations, documentation and accountability.

8. Data Monetisation

Data monetisation means converting data-related assets or capabilities into economic or strategic value. It has two broad forms.

Direct monetisation

The organisation sells or licenses data, reports, benchmarks, APIs, analytical services or data-enabled software. Examples include market intelligence subscriptions, industry benchmarks and risk-scoring services.

Indirect monetisation

The organisation uses data to improve its existing business. Benefits may include higher conversion, lower operating cost, reduced fraud, better retention, improved asset utilisation or faster innovation. Indirect monetisation is often more valuable and less risky because data remains connected to the organisation’s core proposition.

A basic investment appraisal compares expected benefits with costs. Net present value can be represented as:

Here, CF_t is the net cash flow in period t, r is the discount rate per period, n is the number of periods and I₀ is the initial investment. The calculation assumes that cash flows and the discount rate are estimated consistently and that timing is represented correctly. A positive NPV indicates that the estimated discounted benefits exceed the initial investment, subject to uncertainty in the assumptions.

For a simple annual project, return on investment may be expressed as:

ROI is easy to communicate but does not account for the timing of cash flows, risk or benefits that are difficult to monetise. Therefore, financial measures should be supplemented with operational, customer, risk and strategic indicators.

Monetisation screening questions

Before commercialising data, ask:

1. Is there a clearly defined customer problem?
2. Is the data legally collectable, usable and shareable?
3. Is the data sufficiently accurate and current?
4. Is the offering differentiated from alternatives?
5. Can the organisation protect confidentiality and prevent misuse?
6. Are customers likely to trust the proposition?
7. Can delivery costs and support obligations be managed?

Selling personal or sensitive data without meaningful consent can destroy trust and create legal exposure. Monetisation must therefore be assessed together with privacy, security, competition and ethical considerations.

9. Competitive Advantage and Strategic Defensibility

Data can support several generic sources of advantage:

• Cost leadership: Better forecasting, automation and resource utilisation.
• Differentiation: Personalised service, superior reliability or unique insight.
• Focus: Specialised knowledge of a customer segment or industry niche.
• Speed: Faster sensing, experimentation and response.
• Risk management: Earlier identification of fraud, failure or compliance issues.

The strategic question is not whether data is valuable in the abstract, but whether the organisation can create more value than competitors and retain that advantage. Defensibility may arise from proprietary data, exclusive access, switching costs, embedded workflows, cumulative learning, network effects or trust.

However, data advantages can be temporary. Open data, cloud services, foundation models and commercial datasets may reduce barriers to imitation. Regulation may restrict data use, while customers may withdraw consent. Organisations should therefore avoid assuming that a large dataset guarantees a permanent moat.

10. Platform Business Models

A platform connects two or more participant groups and enables interactions, transactions or exchanges. Examples include marketplaces connecting buyers and sellers, payment platforms connecting merchants and customers, and developer platforms connecting software creators with users.

Platforms differ from traditional pipelines. A pipeline generally creates value through a sequence controlled by one organisation: inputs are transformed into outputs and delivered to customers. A platform creates value by enabling interactions among participants, often using data, standards, search, matching, identity, payment and reputation mechanisms.

Network effects

A direct network effect occurs when the value to users in one group increases as more users from that same group participate. A telephone network is a classic example. An indirect or cross-side network effect occurs when participation by one group increases value for another group, such as more sellers attracting buyers on a marketplace.

Platforms must solve the “chicken-and-egg” problem of attracting initial participants. Strategies include subsidies, exclusive partnerships, narrow market entry, internal supply, strong curation and solving a high-value problem for one side first.

Data can strengthen a platform by improving matching, recommendations, fraud detection, pricing and trust. But data concentration can also create privacy, competition and fairness concerns. Platform governance must define participation rules, access, ranking principles, dispute resolution, data rights and accountability.

11. Governance, Ethics and Regulation

Data strategy is inseparable from responsible governance. Governance establishes who may make decisions about data, under what conditions and with which controls.

Important principles include:

• Purpose limitation: Collect and use data for specified legitimate purposes.
• Data minimisation: Avoid collecting data that is unnecessary for the purpose.
• Transparency: Explain relevant uses, decisions and limitations.
• Accuracy: Provide processes for correction and validation.
• Security: Protect confidentiality, integrity and availability.
• Accountability: Assign owners and maintain evidence of compliance.
• Fairness: Test for unjustified discrimination and unequal impacts.
• Human oversight: Ensure appropriate review of consequential automated decisions.

Regulatory obligations vary by jurisdiction and sector. Privacy laws may provide rights concerning access, correction, deletion, portability, consent and automated decision-making. Sector-specific requirements may apply to health, finance, employment, education and critical infrastructure. Contracts may also restrict secondary use or onward sharing.

Ethical governance goes beyond legal compliance. A use may be technically lawful yet harmful, manipulative or inconsistent with reasonable expectations. Impact assessments, stakeholder consultation, model documentation, access controls, retention schedules and incident-response procedures are practical safeguards.

12. Artificial Intelligence and Contemporary Data Strategy

Artificial intelligence can increase the value of data through classification, prediction, recommendation, optimisation and generation. Nevertheless, AI performance depends on data relevance, representativeness, labelling, quality and monitoring. Generative AI introduces additional concerns, including confidentiality leakage, fabricated outputs, intellectual-property disputes and unclear accountability.

An AI-enabled data strategy should specify:

• approved use cases and prohibited uses;
• data provenance and permission to use training data;
• validation and testing requirements;
• human review and escalation;
• monitoring for drift, bias and performance decline;
• security against prompt injection and data extraction;
• documentation of models, limitations and users;
• procedures for correcting harmful outputs.

AI should be treated as one capability within a broader data operating model. Without reliable data, sound processes and accountable decision rights, AI may accelerate poor decisions rather than create value.

13. Illustrative Case: Meridian Grocers’ Data Strategy

Background

Meridian Grocers is a fictional regional supermarket with physical stores, an online channel and a loyalty programme. It faces rising delivery costs, food waste and intense competition from digital retailers. The company has large quantities of transaction, inventory, promotion and customer data, but these are stored in separate systems.

Problem

Management initially proposes building a central data lake and purchasing an AI recommendation tool. However, the business case is unclear. Store managers distrust central reports, product identifiers differ across systems and customers have not been given clear choices about personalisation.

Analysis

Using the value-chain perspective, Meridian finds that data is generated at checkout, online ordering, warehouse operations and supplier interactions. Integration and product identification are major bottlenecks. Analysis also shows that the highest-value decisions concern replenishment, markdown timing, delivery-slot allocation and personalised promotions.

Using the resource-based view, transaction data alone is not rare because competitors possess similar information. Meridian’s potential advantage lies in combining local demand patterns, supplier relationships, store-level operational knowledge and a trusted loyalty relationship. Using the capability perspective, the company lacks common definitions, data ownership, model-monitoring routines and frontline adoption.

Approach and solution

Meridian adopts a federated operating model. Business units own data definitions and use cases, while a central team provides shared engineering, security and governance services. It establishes product ownership for three data products:

1. A store-level replenishment service.
2. A waste and markdown recommendation service.
3. A consent-aware customer offer service.

The first release focuses on replenishment because it has measurable operational value and limited privacy risk. Data-quality rules are introduced for product identifiers, stock levels and delivery dates. Recommendations are displayed in the existing store workflow rather than in a separate analytics portal. Managers can override recommendations, and overrides are recorded for learning.

For customer offers, Meridian introduces clear consent controls and excludes sensitive inferences. Model performance is monitored by store, customer segment and product category. The company evaluates value using availability, waste, margin, customer satisfaction and complaint measures rather than model accuracy alone.

Outcome

The initiative creates a reusable identity and governance foundation while delivering a visible operational improvement. Store adoption increases because the product is embedded in existing routines. Meridian postpones external data sales until internal data quality, privacy controls and product management are mature.

Learning points

The case demonstrates that data strategy should begin with decisions and outcomes, not a technology label. It also shows that competitive advantage is produced by combinations of data, relationships, capabilities and adoption. Finally, responsible data use is not a separate compliance exercise; it shapes product design and customer trust.

14. Challenges and Limitations

Data strategies commonly encounter the following difficulties:

• Unclear ownership: Nobody is accountable for definitions, quality or access decisions.
• Fragmented architecture: Systems use incompatible identifiers and formats.
• Low data quality: Missing, duplicated or outdated records undermine analysis.
• Short-termism: Organisations fund visible technology but not foundational work.
• Low adoption: Employees do not trust or understand analytical outputs.
• Overstated business cases: Benefits are estimated without credible baselines.
• Privacy and security risk: Data is collected or shared beyond reasonable expectations.
• Model drift: Relationships change as customers, markets and behaviours evolve.
• Vendor dependence: Proprietary tools may create switching costs and reduce flexibility.
• Cultural resistance: Data initiatives may be perceived as surveillance or loss of professional autonomy.

These limitations imply that strategy must be iterative. Pilot projects should test assumptions, establish baselines and generate evidence. Governance should be proportionate to risk, and the organisation should be willing to retire data products that no longer create value.

15. Best-Practice Checklist

A robust data strategy should:

1. Begin with corporate objectives and important decisions.
2. Define measurable outcomes and credible baselines.
3. Treat data products as services with owners and users.
4. Combine value-chain, resource and capability analysis.
5. Invest in quality, metadata, identity and interoperability.
6. Clarify decision rights through practical governance.
7. Embed analytics in workflows rather than relying only on dashboards.
8. Build privacy, security and ethics into design.
9. Monitor adoption, business outcomes, model performance and unintended effects.
10. Develop capabilities progressively through a sequenced roadmap.

Chapter Summary

Data strategy aligns data resources and capabilities with corporate objectives. Its central concern is not data volume but the creation of measurable value through better decisions, processes, products and relationships. The value-chain perspective traces data from generation to action and exposes bottlenecks in quality, integration and adoption. The resource-based view asks whether data and related assets are valuable, rare, difficult to imitate and supported by the organisation. The capability perspective focuses on repeatable abilities in governance, engineering, analytics, product management, literacy and responsible use.

Data products convert data into usable, managed services. Monetisation may be direct, through licensing and data-enabled offerings, or indirect, through improved performance in the core business. Competitive advantage depends on combinations of data, knowledge, relationships, technology and organisational capability, and may be temporary when data is easily imitated. Platform models create value by enabling interactions among participant groups and may benefit from network effects, while requiring careful rules for trust, access, privacy and competition.

Effective data strategy is iterative and outcome-oriented. It requires governance, ethical judgement, security, adoption and continuous measurement. Artificial intelligence can increase data value, but only when supported by reliable data, appropriate controls and accountable decision processes.

Key Terms

Data strategy: A coordinated plan for using data to support organisational objectives.

Data asset: Data that may provide operational, analytical or strategic value.

Data value chain: The sequence from data generation and acquisition to analysis, action and learning.

Resource-based view: A theory explaining advantage through valuable organisational resources and capabilities.

VRIN: Valuable, rare, inimitable and non-substitutable characteristics of strategic resources.

VRIO: A framework adding the question of whether the organisation is structured to exploit a resource.

Data capability: An organisational ability to manage and use data effectively.

Dynamic capability: The ability to sense, seize and reconfigure resources as conditions change.

Data governance: Decision rights, standards, policies and accountability for data.

Data product: A managed data-enabled service with users, ownership, quality expectations and feedback mechanisms.

Data monetisation: Conversion of data-related assets or capabilities into economic or strategic value.

Direct monetisation: Revenue generated by selling or licensing data or data-enabled services.

Indirect monetisation: Value created by improving an organisation’s existing operations or offerings.

Competitive advantage: The ability to create greater value than competitors or achieve superior strategic outcomes.

Platform business model: A model that enables interactions among two or more participant groups.

Network effect: A change in user value resulting from changes in participation by users or participant groups.

Data quality: The degree to which data is fit for its intended purpose.

Data product owner: The person accountable for a data product’s purpose, performance and development.

Purpose limitation: The principle that data should be collected and used for specified legitimate purposes.

Model drift: Decline in model performance caused by changes in data, behaviour or the operating environment.

Review Questions

Short Answer Questions

1. What is the difference between a corporate strategy and a data strategy?
2. List the main stages of a data value chain.
3. Explain the meaning of VRIN in the resource-based view.
4. Distinguish direct from indirect data monetisation.
5. What is a data product, and how does it differ from a raw dataset?

Long Answer Questions

1. Explain how the value-chain, resource-based and capability perspectives complement one another in data strategy.
2. Describe a systematic process for developing a data strategy aligned with corporate objectives.
3. Discuss the conditions under which data can produce sustainable competitive advantage.
4. Explain the characteristics, opportunities and governance challenges of platform business models.
5. Evaluate the ethical, privacy and regulatory issues that should be considered before monetising data.

Analytical / Application Questions

1. A logistics company has accurate historical delivery data but high delivery failure rates. Use the data value-chain perspective to identify likely bottlenecks and recommend priorities.
2. A healthcare organisation wants to sell an external analytics service based on patient data. Evaluate the opportunity using resource, capability, monetisation and governance perspectives.
3. Design a two-year data-strategy roadmap for a medium-sized retailer seeking to improve retention and reduce inventory waste. State the business objectives, data products, capabilities, governance controls and outcome measures.

Multiple Choice Questions with Answers

Question 1

Which sequence best represents the central logic of an aligned data strategy?

A. Technology purchase → data collection → staffing → corporate strategy

B. Strategic objective → decision → required data → capability → action → value

C. Data volume → storage expansion → dashboard creation → value

D. Regulation → software selection → data deletion → revenue

Correct Answer: B. Strategic objective → decision → required data → capability → action → value

Question 2

Which activity belongs primarily to the data value-chain stage of productisation?

A. Generating a transaction record

B. Correcting duplicate customer identifiers

C. Embedding a fraud-risk score in an authorisation API

D. Collecting data from a sensor

Correct Answer: C. Embedding a fraud-risk score in an authorisation API

Question 3

Under the resource-based view, which characteristic is NOT part of the VRIN framework?

A. Valuable

B. Rare

C. Inimitable

D. Automatically profitable

Correct Answer: D. Automatically profitable

Question 4

Why may proprietary data fail to create competitive advantage?

A. Data cannot be copied

B. The organisation may lack capabilities to use it effectively

C. Data is always legally unusable

D. Data has no relationship to business decisions

Correct Answer: B. The organisation may lack capabilities to use it effectively

Question 5

Which is an example of indirect data monetisation?

A. Licensing an industry benchmark to external firms

B. Selling access to a customer-data API

C. Using demand forecasts to reduce inventory waste

D. Charging a subscription for a market database

Correct Answer: C. Using demand forecasts to reduce inventory waste

Question 6

What most clearly distinguishes a data product from a raw dataset?

A. A data product must contain only numerical data

B. A data product has users, ownership, quality expectations and a delivery mechanism

C. A raw dataset is always more valuable

D. A data product cannot be updated

Correct Answer: B. A data product has users, ownership, quality expectations and a delivery mechanism

Question 7

In a platform business model, an indirect network effect occurs when:

A. Users copy data at no cost

B. More users on one participant side increase value for another side

C. A platform removes all governance rules

D. A firm reduces the number of participant groups

Correct Answer: B. More users on one participant side increase value for another side

Question 8

Which principle requires collecting and using data only for specified legitimate purposes?

A. Purpose limitation

B. Network neutrality

C. Economies of scale

D. Model drift

Correct Answer: A. Purpose limitation

Question 9

A data initiative has an initial investment of $100,000 and expected first-year net benefits of $130,000. Using the simple ROI formula in the chapter, what is the ROI?

A. 13%

B. 23%

C. 30%

D. 130%

Correct Answer: C. 30%

Question 10

Which practice is most appropriate for a responsible AI-enabled data strategy?

A. Deploying a model without monitoring after launch

B. Using all available personal data regardless of consent

C. Monitoring performance, bias, drift and human-review requirements

D. Replacing governance with model accuracy alone

Correct Answer: C. Monitoring performance, bias, drift and human-review requirements