Managerial Strategy, Transformation and the Future of Programmatic

Introduction

Programmatic advertising has evolved from an automated method of purchasing display impressions into an organisational capability spanning data, media, technology, measurement, commerce and customer experience. The managerial question is no longer simply how to bid efficiently. It is how an advertiser should design its operating model, allocate control between internal teams and external partners, govern data and algorithms, and create value across increasingly connected channels.

A modern programmatic ecosystem may include demand-side platforms (DSPs), supply-side platforms (SSPs), ad exchanges, identity services, data clean rooms, retail-media networks, connected television (CTV), digital out-of-home (DOOH), commerce platforms and measurement providers. Each additional component can create reach and insight, but also introduces cost, duplication, opacity, privacy exposure and operational complexity.

Strategic competence therefore requires a portfolio perspective. A firm must decide which capabilities are sources of competitive advantage, which are routine utilities, and which should be shared with specialist partners. It must also evaluate whether apparent performance improvements result from genuine incremental value or from attribution practices, duplicated reach and fee extraction.

Learning Objectives

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

1. Evaluate build-versus-buy and in-housing decisions using strategic, economic and governance criteria.
2. Design a vendor-selection and remuneration framework for programmatic partners.
3. Compare operating models, including agency-led, hybrid and advertiser-led structures.
4. Analyse the roles of omnichannel media, CTV, retail media, DOOH and commerce media.
5. Explain the managerial significance of privacy-enhancing technologies and data clean rooms.
6. Assess sustainability, market consolidation and the likely effects of agentic advertising.
7. Construct a transformation roadmap with suitable controls, metrics and accountability.

1. Programmatic Capability as a Strategic Asset

A capability is strategically important when it contributes to valuable outcomes, is difficult to imitate, and can be integrated with other organisational resources. Programmatic buying itself is increasingly standardised; access to a DSP is rarely a sustainable differentiator. Differentiation is more likely to arise from the quality of first-party data, experimentation, creative systems, measurement design, business integration and organisational learning.

A useful distinction is between commoditised capabilities and differentiating capabilities. Bid execution, standard trafficking and routine reporting may be obtained efficiently from a vendor. By contrast, customer-value modelling, privacy governance, brand-safety policy, incrementality testing and the translation of commercial objectives into media decisions may deserve internal ownership.

The resource-based view of strategy suggests four questions:

• Value: Does the capability improve revenue, margin, customer experience or risk control?
• Rarity: Is it uncommon among competitors?
• Imitability: Can competitors reproduce it easily?
• Organisation: Is the firm organised to exploit it?

The final question is often neglected. A company may purchase excellent technology yet fail to benefit because marketing, data science, procurement, legal, finance and information security operate independently.

2. Build Versus Buy and the Logic of In-Housing

2.1 The decision is not binary

Build-versus-buy decisions are better understood as a spectrum:

• Buy: acquire a standard service from an external provider.
• Configure: adapt a vendor platform to internal processes.
• Partner: share responsibility with a specialist under defined governance.
• Build: create proprietary technology, models or workflows.
• Acquire: purchase a firm or capability when speed and control justify the investment.

Most mature advertisers use a hybrid model. They may own strategy, data governance, experimentation and commercial accountability, while buying infrastructure, auction access and specialist operations.

2.2 Total cost of ownership

The relevant comparison is not a vendor invoice against the salary of an internal employee. It is total cost of ownership (TCO), including technology, personnel, integration, security, training, compliance, opportunity cost and transition risk.

Here, each C represents a cost over the same evaluation period and in the same currency. The formula is an accounting framework rather than a prediction model. Costs should be discounted when comparing multi-year alternatives, and assumptions about volume, inflation, platform fees and staff utilisation must be explicit.

A vendor option may have a lower visible cost but higher switching costs or weaker transparency. An internal option may require substantial fixed investment yet become economical at scale. A simple break-even analysis can be useful. If internalisation requires fixed annual cost F and variable cost v per impression, while outsourcing costs p per impression, the internal model is cheaper when:

where Q is annual impressions. Thus the break-even volume is:

This model assumes constant unit costs and excludes quality differences, risk and capital costs. It is therefore a screening tool, not a complete investment case.

2.3 What should be in-house?

In-housing is strongest when the advertiser has sufficient scale, stable talent demand, a distinctive data asset, strong governance maturity and a need for rapid cross-functional learning. It is less attractive when media volume is small, talent is scarce, markets are fragmented, or the company cannot maintain technical and regulatory capabilities.

Capabilities commonly suitable for internal ownership include:

• business and media strategy;
• customer and consent governance;
• measurement principles and test design;
• commercial negotiation and fee oversight;
• brand, suitability and risk policy;
• proprietary audience or value models;
• cross-channel planning and organisational learning.

Capabilities often bought or shared include exchange infrastructure, specialist creative production, verification services, identity resolution, managed-service operations and temporary implementation expertise.

The main risks of in-housing are not merely operational. They include talent concentration, dependence on a small number of employees, weak segregation of duties, inadequate audit capability and the possibility that internal teams reproduce vendor incentives without independent challenge.

3. Operating Models and Organisational Design

3.1 Three common models

A hybrid model should not mean that every party performs every task. It requires a rights-and-responsibilities matrix. For each activity, one party should be accountable, others responsible or consulted, and decision rights should be documented. Areas requiring explicit ownership include audience creation, bid rules, supply-path selection, creative approval, budget movement, incident response and measurement.

3.2 The centre-of-excellence model

A programmatic centre of excellence can provide standards, training, measurement templates, privacy controls, vendor governance and experimentation support to regional or brand teams. It should not become a bureaucratic approval layer. Its value lies in reusable capabilities and evidence-based learning.

A mature operating model separates four functions:

1. Value setting: commercial objectives, customer strategy and budget priorities.
2. Decisioning: audience, bidding, frequency, channel and supply choices.
3. Execution: trafficking, optimisation, creative delivery and troubleshooting.
4. Assurance: audit, privacy, brand safety, fraud, financial control and incrementality measurement.

Separating assurance from execution reduces conflicts of interest. The party paid to deliver media should not be the only party determining whether delivery created value.

4. Vendor Selection and Remuneration

4.1 Selection criteria

A vendor should be evaluated across more than reach and platform functionality. A weighted scorecard may include:

• business and channel fit;
• transparency of inventory, fees and auction mechanics;
• data use, consent management and retention controls;
• interoperability and portability of data;
• identity and addressability capabilities;
• measurement quality and access to logs;
• brand safety, fraud and suitability controls;
• security, resilience and incident response;
• service quality and talent continuity;
• sustainability and supply-path efficiency;
• commercial flexibility and termination rights.

Scores should be supported by evidence such as audit reports, technical documentation, test results, references and contract commitments. A polished demonstration is not evidence of operational performance.

4.2 Remuneration design

Remuneration shapes behaviour. Percentage-of-media-spend fees can encourage spending growth rather than efficiency. Fixed retainers improve predictability but may weaken incentives to improve outcomes. Performance fees can reward results but become problematic when outcomes are noisy, attribution is disputed or the agency controls the measurement method.

A sound remuneration model typically combines:

• a transparent fixed fee for defined services;
• pass-through treatment for third-party costs;
• clearly disclosed platform and technology charges;
• service-level commitments;
• incentives tied to controllable, audited outcomes;
• safeguards against unsafe inventory and non-incremental conversions.

Performance compensation should use a pre-agreed measurement protocol. If a vendor is rewarded for return on advertising spend (ROAS), the parties must define revenue, costs, attribution windows, cancellations, margins and treatment of existing customers. Revenue-based ROAS is not equivalent to profit-based return.

4.3 Contractual governance

Contracts should address audit rights, data ownership, subcontracting, model explainability, security, breach notification, business continuity, conflicts of interest, inventory reselling, deletion on termination and assistance with transition. The advertiser should have access to sufficient logs to reconcile spend and evaluate supply quality without requiring unrestricted access to personal data.

5. Omnichannel and Emerging Media Environments

5.1 Omnichannel planning

Omnichannel programmatic planning coordinates messages and budgets across display, video, audio, social, search, CTV, DOOH and commerce environments. The objective is not to force all channels into one buying tool. It is to coordinate reach, frequency, sequencing, creative and measurement while respecting each channel’s commercial and technical constraints.

A useful planning sequence is:

1. Define the customer and business outcome.
2. Map addressable and non-addressable inventory.
3. Establish channel roles in the journey.
4. Set reach, frequency and outcome hypotheses.
5. Allocate budgets using tests rather than historical habit alone.
6. Measure both channel outcomes and incremental contribution.

Cross-channel frequency is difficult because identifiers, login systems, household graphs and measurement windows differ. A reported frequency of four may not mean that the same person received four exposures. Consequently, omnichannel claims require careful qualification.

5.2 Connected television

CTV combines television-style content with digital delivery and, in some environments, addressable buying. Its strengths include premium audiovisual attention, household reach and opportunities for outcome measurement. Its limitations include fragmented supply, inconsistent definitions of completion, co-viewing uncertainty, limited deterministic identity and possible duplication between publisher and intermediary reporting.

CTV strategy should distinguish:

• programmatic guaranteed, where inventory and price are committed in advance;
• private marketplace buying, where selected buyers access restricted supply;
• open-market buying, where auction access is broader and quality variation is greater.

CTV requires controls for frequency, content suitability, fraud, device duplication and the difference between an ad being delivered and being meaningfully viewed.

5.3 Retail media and commerce media

Retail media uses retailer-owned or retailer-controlled inventory, data and transaction environments. Commerce media is broader: it includes media and audience opportunities connected to shopping intent, marketplaces, payment ecosystems, product content and conversion.

Retail media can connect exposure to sales more directly than many other channels, but this apparent advantage can be misleading. Retailers may measure only their own ecosystem, favour last-touch attribution and charge for access to data generated by the advertiser’s customers. Strategic questions include data portability, incremental sales, margin, category effects and whether media is influencing demand or merely capturing existing intent.

5.4 Digital out-of-home

DOOH enables automated buying of screens in public environments. It can support geographic, temporal and contextual targeting, but individual-level attribution is usually inferential rather than deterministic. Suitable measures may include qualified reach, location-based visitation studies, brand-lift research and controlled geographic experiments. DOOH should not be evaluated solely by applying online click metrics to an offline medium.

6. Privacy-Enhancing Technologies and Clean Rooms

Privacy-enhancing technologies (PETs) seek to make useful analysis possible while reducing exposure of identifiable data. Relevant approaches include aggregation, pseudonymisation, tokenisation, encryption, differential privacy, secure multi-party computation, federated learning and clean-room environments.

A data clean room is a controlled environment in which parties compare or analyse datasets under restrictions on access, output and identity exposure. Typical uses include audience overlap, measurement, suppression, conversion analysis and media planning. A clean room is not automatically privacy-preserving merely because it has that label. Governance must address purpose limitation, consent, minimum audience thresholds, output inspection, re-identification risk, retention and permitted users.

PETs involve trade-offs. Stronger aggregation can reduce analytical precision. Differential privacy introduces statistical noise; its protection depends on the privacy parameters and the full set of queries, not on a marketing label. Encryption protects data in particular states but does not eliminate the need for access control or lawful purpose.

The managerial principle is privacy by design: collect less, retain for shorter periods, restrict access, document purpose, test re-identification risk and provide meaningful accountability. Consent banners alone do not constitute a complete privacy strategy.

7. Sustainability and Supply-Path Management

Programmatic advertising consumes energy through data centres, auctions, data transfer, creative delivery and duplicated intermediaries. Sustainability strategy should therefore address both media emissions and broader environmental effects.

Supply-path optimisation can reduce waste by selecting efficient routes to inventory, removing unnecessary resellers, limiting excessive bid requests and using direct publisher relationships where quality is equivalent. However, the cheapest or shortest path is not automatically the best path. It must also satisfy reach, diversity, quality, privacy, fraud prevention and publisher-value objectives.

A credible sustainability programme should specify system boundaries and measurement methods. Claims must distinguish estimated operational emissions from independently verified reductions. Reducing impressions may reduce energy use but may also reduce campaign reach; the strategic question is emissions per effective outcome, not emissions in isolation.

8. Consolidation, Interoperability and Market Power

The ecosystem has experienced continuing consolidation among platforms, data providers, agencies, measurement firms, commerce companies and media owners. Consolidation can produce integration efficiencies and improve investment in technology. It can also increase dependency, reduce contestability and make it difficult for advertisers to compare performance.

Managers should monitor:

• conflicts where one firm acts across multiple sides of a transaction;
• opaque fee structures and undisclosed rebates;
• restrictions on data portability;
• self-preferencing in measurement or inventory access;
• interoperability failures;
• concentration of identity, attention or retail data.

The strategic response is not to reject all integrated platforms. It is to preserve auditability, maintain alternative routes where economically justified, require contractually defined data access and avoid allowing a single provider to define objectives, execute media and certify its own success without independent validation.

9. Agentic Advertising and the Future Operating Model

Agentic advertising refers to systems that can interpret goals, plan actions, use tools, execute buying or creative changes, observe results and revise decisions with limited human intervention. Such systems may coordinate budget allocation, audience selection, creative variants, supply-path choices and experimentation.

The promise is faster optimisation across complex environments. The risks include objective mis-specification, hidden proxy optimisation, feedback loops, discrimination, unauthorised spending, prompt or data manipulation, model drift and weak explainability.

A safe agentic architecture should include:

• a clearly bounded objective and budget;
• permission levels and human approval thresholds;
• immutable logs of recommendations and actions;
• policy constraints for privacy, suitability and frequency;
• independent monitoring and rollback capability;
• adversarial testing and red-team exercises;
• separation between experimentation and production deployment;
• accountability assigned to a named executive function.

Automation does not remove managerial responsibility. If an agent increases conversions by targeting vulnerable consumers or exploiting a measurement artefact, the organisation remains accountable for the design and deployment of the system.

10. Transformation Roadmap and Performance Management

Transformation should proceed through stages rather than through a technology purchase.

Stage 1: Diagnose

Map current spend, suppliers, data flows, fees, decision rights, skills, measurement limitations and regulatory exposure. Reconcile invoices, platform reports and independent verification.

Stage 2: Define the target state

Specify which capabilities will be owned, shared or bought. Establish principles for privacy, transparency, sustainability, experimentation and channel coordination.

Stage 3: Pilot

Select a bounded market, brand or channel. Test the proposed operating model against a baseline. Document both performance and implementation burden.

Stage 4: Institutionalise

Create training, standard operating procedures, contracts, controls, measurement governance and escalation paths. Build mechanisms for transferring learning across teams.

Stage 5: Reassess

Review the model at regular intervals because technology, regulation, supplier economics and organisational capabilities change.

A balanced scorecard can combine financial, customer, operational, learning and risk indicators. Examples include incremental profit, qualified reach, attention or completion quality, supply-path cost, test velocity, data incidents, audit findings, employee capability and publisher diversity. No single metric captures programmatic value.

Illustrative Case: Northstar Home Systems

Background

Northstar Home Systems, a fictional multinational manufacturer, purchased programmatic media through regional agencies. Its display activity was efficient by reported ROAS, but its CTV, retail media and DOOH programmes were managed separately. Procurement identified unexplained technology charges, while marketing suspected that retargeting received credit for customers who would have purchased anyway.

Problem

The company had three strategic problems: duplicated audiences across channels, limited transparency into supply and fees, and no common method for estimating incrementality. Senior management proposed bringing all buying in-house immediately.

Analysis

An internal review showed that Northstar’s distinctive assets were its customer-value models, product-margin data, consent governance and experimental design. Exchange access, global trafficking infrastructure and some specialist CTV operations were not distinctive and would be expensive to recreate. The proposed all-or-nothing in-house solution would therefore create unnecessary cost and concentration risk.

The team designed a hybrid model. Northstar would own strategy, data policy, measurement, supplier governance and budget rules. Agencies would operate selected markets under transparent fees and service levels. A single independent measurement partner would support experiments, while retail partners would participate in controlled clean-room analyses.

Approach and solution

Northstar first reconciled invoices and established a supply-path policy. It then ran geographic holdout tests for selected product categories, comparing exposed and control regions while controlling for seasonality and distribution changes. CTV and DOOH were evaluated using reach, completion, brand-lift and visitation measures rather than clicks alone. Retail media was assessed using incremental margin and new-customer outcomes.

The company introduced a fixed service fee with disclosed pass-through costs, quarterly audit rights and a bonus linked to independently assessed incremental contribution. It retained human approval for major budget movements and prohibited the use of sensitive data categories for targeting.

Outcome

After two quarters, Northstar reported improved fee visibility, fewer duplicated supply routes and faster learning cycles. Some channels showed lower attributed ROAS after last-touch credit was removed, but controlled tests indicated that several retained activities generated incremental sales. The organisation did not eliminate agencies; it changed their role from opaque media resellers to accountable operating partners.

Learning points

1. In-housing should follow capability analysis, not fashion.
2. Reported efficiency is not equivalent to incremental business value.
3. Clean rooms and sophisticated measurement require governance, not merely technology.
4. Remuneration should reward outcomes that are measurable and within the partner’s influence.
5. Transformation is an operating-model programme involving people, contracts and controls.

Challenges and Limitations

Programmatic strategy remains constrained by fragmented identifiers, inconsistent measurement, probabilistic attribution and changing regulation. Privacy protection can reduce addressability and make historical benchmarks unreliable. Clean rooms may reproduce power imbalances if a dominant platform controls both data and analytical rules. Sustainability estimates often depend on incomplete supply-chain information. AI systems can optimise toward measurable proxies while damaging less visible outcomes such as trust, inclusion or long-term brand equity.

Managers should therefore treat models and dashboards as decision aids rather than objective reality. Triangulation—combining platform data, independent verification, experiments, qualitative research and financial evidence—is more defensible than relying on a single source.

Best-Practice Principles

• Start with business value and customer trust, not platform features.
• Assign clear decision rights and independent assurance.
• Compare total cost and risk, not headline fees.
• Own distinctive capabilities; buy commoditised infrastructure selectively.
• Require transparent contracts, logs, fee disclosure and exit rights.
• Evaluate every channel according to its role and appropriate evidence.
• Use clean rooms and PETs with explicit privacy governance.
• Measure incrementality where decisions depend on causal claims.
• Include sustainability and supply quality in procurement.
• Keep human accountability for automated and agentic decisions.

Chapter Summary

Programmatic advertising is now an organisational and strategic system rather than a narrow buying technique. Build-versus-buy decisions should compare total cost, strategic distinctiveness, risk, scalability and governance. In-housing is most effective when firms retain strategy, data stewardship, measurement and accountability while selectively using external infrastructure and expertise.

Operating-model design requires explicit decision rights, independent assurance and remuneration that does not reward opaque spending or non-incremental outcomes. Omnichannel, CTV, retail media, commerce media and DOOH expand opportunity but require channel-appropriate measurement and careful treatment of identity, frequency and attribution.

Privacy-enhancing technologies and clean rooms can support collaboration, but they do not replace lawful purpose, data minimisation and accountability. Sustainability requires attention to supply paths and effective outcomes. Consolidation creates efficiencies alongside dependency and conflicts of interest. Agentic advertising may improve speed and coordination, but bounded permissions, monitoring, audit trails and human responsibility are essential.

The durable managerial advantage lies in integrating technology, people, contracts, experimentation and governance into a learning system that creates measurable business value while protecting consumers and market integrity.

Key Terms

Agentic advertising: Automated advertising systems that plan, act, observe results and adapt toward defined objectives.

Build-versus-buy: The decision whether to develop a capability internally or obtain it from an external provider.

Commerce media: Media and audience activity connected to shopping, marketplaces, transactions and product discovery.

Connected television (CTV): Television-like content delivered through internet-connected devices or services.

Centre of excellence: A specialised internal group that provides standards, expertise, training and governance.

Data clean room: A controlled environment for analysing data across parties while restricting identity exposure and outputs.

Digital out-of-home (DOOH): Digitally delivered advertising on public-space screens.

Incrementality: The additional outcome caused by advertising beyond what would have occurred without it.

In-housing: Performing selected advertising capabilities within the advertiser organisation.

Omnichannel planning: Coordinating media channels around audiences, journeys, frequency and outcomes.

Privacy-enhancing technology: A technical method that enables useful processing while reducing privacy exposure.

Programmatic guaranteed: Automated buying in which price and inventory are committed in advance.

Retail media: Advertising using a retailer’s media inventory, data or transaction environment.

Supply-path optimisation: Selecting efficient, transparent and suitable routes to advertising inventory.

Total cost of ownership: The complete cost of acquiring, operating, governing and replacing a capability.

Transparency: The ability to inspect fees, data use, inventory, decision rules and outcomes.

Vendor governance: Processes for selecting, monitoring, auditing and managing external providers.

Zero-party data: Information a consumer intentionally and proactively provides to a brand.

Review Questions

Short Answer Questions

1. Why is in-housing not equivalent to purchasing a DSP directly?
2. Identify four components of total cost of ownership in a build-versus-buy decision.
3. What is the principal purpose of a data clean room?
4. Why should DOOH not be evaluated solely through click-through rate?
5. State two governance controls needed for agentic advertising.

Long Answer Questions

1. Critically evaluate the advantages and risks of agency-led, advertiser-led and hybrid operating models.
2. Explain how remuneration structures can influence vendor behaviour and propose a suitable model for a large advertiser.
3. Analyse the strategic opportunities and measurement challenges of retail media and commerce media.
4. Discuss the relationship between privacy-enhancing technologies, clean rooms and privacy-by-design governance.
5. Develop a transformation roadmap for an advertiser seeking greater programmatic transparency and internal capability.

Analytical / Application Questions

1. An advertiser faces an internal fixed cost of 2,000,000 currency units, an internal variable cost of 0.40 per thousand impressions and an outsourced price of 0.90 per thousand impressions. Using the break-even framework, calculate the annual impression volume at which internal execution becomes cheaper. State the assumptions and limitations of the calculation.
2. Design a vendor scorecard with at least six criteria for selecting a CTV and retail-media partner. Assign weights and explain how you would validate the vendor’s claims.
3. A campaign reports excellent last-touch ROAS but shows no improvement in sales in matched control regions. Diagnose at least three possible explanations and recommend a measurement response.

Multiple Choice Questions with Answers

Question 1

Which capability is most likely to represent a distinctive strategic asset for a sophisticated advertiser?

A. Access to standard exchange inventory

B. Routine trafficking of campaign tags

C. Proprietary customer-value modelling integrated with incrementality testing

D. Generic platform interface training

Correct Answer: C. Proprietary customer-value modelling integrated with incrementality testing

Question 2

In the break-even expression Q* = F ÷ (p − v), what does F represent?

A. Annual impression volume

B. Internal fixed cost

C. Outsourced unit price

D. Internal variable cost

Correct Answer: B. Internal fixed cost

Question 3

Which operating model generally combines advertiser control with external specialist expertise?

A. Pure agency-led model

B. Hybrid model

C. Unmanaged open-market model

D. Fully decentralised model

Correct Answer: B. Hybrid model

Question 4

Why can a percentage-of-spend remuneration model create a conflict of interest?

A. It prevents all media buying

B. It may reward higher spending rather than better incremental outcomes

C. It eliminates platform fees

D. It guarantees independent measurement

Correct Answer: B. It may reward higher spending rather than better incremental outcomes

Question 5

What is a central governance requirement for a data clean room?

A. Unlimited export of person-level records

B. Removal of all contractual restrictions

C. Controls on permitted outputs and re-identification risk

D. Use of data without a defined purpose

Correct Answer: C. Controls on permitted outputs and re-identification risk

Question 6

Which statement best describes programmatic guaranteed buying?

A. Inventory and price are committed in advance

B. All inventory is purchased anonymously in an open auction

C. No publisher relationship is required

D. The buyer pays only after a click

Correct Answer: A. Inventory and price are committed in advance

Question 7

Why is last-touch ROAS potentially misleading?

A. It cannot report revenue

B. It may credit advertising for conversions that would have occurred anyway

C. It always excludes retargeting

D. It measures only environmental impact

Correct Answer: B. It may credit advertising for conversions that would have occurred anyway

Question 8

Which practice most directly supports supply-path optimisation?

A. Adding every available intermediary

B. Selecting transparent and efficient routes to suitable inventory

C. Maximising bid requests regardless of value

D. Removing all independent verification

Correct Answer: B. Selecting transparent and efficient routes to suitable inventory

Question 9

Which control is particularly important for agentic advertising?

A. Unlimited autonomous budget movement

B. No logging of system actions

C. Human approval thresholds and rollback capability

D. Permanent exemption from privacy policy

Correct Answer: C. Human approval thresholds and rollback capability

Question 10

What is the main strategic danger of allowing one provider to execute media and certify its own success?

A. Improved independent assurance

B. Reduced need for contracts

C. Conflict of interest and weak auditability

D. Guaranteed incrementality

Correct Answer: C. Conflict of interest and weak auditability