Programmatic Advertising and Ad Technology

Introduction and Context

Programmatic advertising is the automated buying and selling of digital advertising inventory through software, data, and standardized transaction protocols. It is not synonymous with simply buying advertisements cheaply or replacing human judgment with algorithms. Rather, it is an infrastructure for making decisions about whether, where, when, and at what price an impression should be purchased.

The central managerial problem is the allocation of scarce attention. An advertiser seeks audiences likely to respond, publishers seek to monetize available inventory, and intermediaries coordinate these objectives under constraints involving price, privacy, quality, speed, and contractual obligations. A programmatic transaction may be completed in milliseconds, yet it reflects decisions about brand strategy, audience definition, bidding, measurement, fraud prevention, and legal compliance.

Programmatic systems are particularly important because digital media supply is fragmented across websites, mobile applications, connected television, digital audio, retail media, and emerging environments. Automation provides scale and decision speed, but it also creates opacity. A sophisticated media planner therefore needs to understand both the technical architecture and the governance mechanisms that make automated buying accountable.

Learning Objectives

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

1. Explain the roles of ad servers, demand-side platforms, supply-side platforms, exchanges, and data platforms.
2. Describe real-time bidding, private marketplaces, header bidding, and programmatic guaranteed transactions.
3. Distinguish first-party, second-party, and third-party data and evaluate identity solutions.
4. Apply contextual targeting, supply-path optimization, and quality controls.
5. Evaluate the economics, measurement, privacy, and governance of programmatic deals.
6. Design a defensible programmatic buying framework for a complex campaign.

1. The Programmatic Advertising Ecosystem

A simplified transaction begins when a user loads a page or opens an application. The publisher identifies an available advertising impression and sends a bid request through a supply-side platform or exchange. Eligible demand-side platforms evaluate the request against an advertiser’s campaign rules, estimate the impression’s value, and return a bid. The auction or deal mechanism selects an outcome, after which the creative is served and measurement events are recorded.

The principal participants are as follows.

Advertiser and agency

The advertiser supplies the marketing objective, budget, audience strategy, creative assets, brand requirements, and business constraints. An agency may plan, negotiate, execute, and optimize the campaign. The advertiser remains accountable for lawful processing, truthful claims, and appropriate use of consumer data, even when execution is outsourced.

Ad server

An ad server is a system that stores, selects, delivers, and measures advertising creative. It may operate on the advertiser side, the publisher side, or both.

An advertiser ad server, often called a third-party ad server in traditional terminology, centralizes campaign trafficking, rotation, frequency rules, tracking, and reporting across publishers. A publisher ad server manages the publisher’s inventory, direct-sold campaigns, priority rules, and yield decisions. Modern platform boundaries can overlap, but the functional distinction remains useful.

Typical ad-server functions include:

• creative hosting and version control;
• campaign and line-item prioritization;
• frequency and recency controls;
• impression, click, and conversion logging;
• creative quality and policy checks; and
• reconciliation of delivered volume against contractual commitments.

An ad server is not necessarily a buying platform. It may execute a decision made elsewhere.

Demand-side platform (DSP)

A DSP enables buyers to purchase impressions across multiple sources. It receives bid requests, applies campaign eligibility rules, evaluates audience and contextual signals, predicts outcomes, submits bids, and reports performance. A DSP may provide access to open auctions, private marketplaces, curated supply, connected television, audio, or retail media.

The DSP’s decision is constrained by budget, bid strategy, pacing, frequency, geography, device, environment, brand suitability, and measurement requirements. Its optimization model might seek conversions, completed views, reach, or another objective. Machine learning can estimate response probability, but the quality of the result depends on the quality and legality of the input signals.

Supply-side platform (SSP)

An SSP helps publishers package and sell inventory. It connects publisher ad servers to exchanges and buyers, supports auction and deal rules, enforces publisher controls, and reports revenue and delivery. The SSP may optimize among direct and indirect demand sources.

An SSP is not merely a reverse DSP. It represents publisher interests, including floor prices, buyer permissions, inventory classification, latency constraints, and demand competition.

Ad exchange

An exchange is a marketplace that facilitates automated transactions between demand and supply. It may conduct auctions, transmit bid requests, apply eligibility rules, and return a winning bid. In practice, exchanges and SSPs may be integrated or marketed as distinct products, so organizational labels should not be treated as universal technical definitions.

Data platforms

A customer data platform (CDP) generally organizes first-party customer data for activation, service, and analytics. A data management platform (DMP) traditionally organized audience segments, especially cookie-based segments, for advertising activation. The boundary has become less stable as privacy restrictions, identity changes, clean rooms, and media platforms have developed.

A CDP typically emphasizes known customer records and consented first-party events. A DMP historically emphasized pseudonymous audience categories and short-lived activation data. Neither platform automatically creates lawful, accurate, or valuable audiences; governance, data provenance, freshness, and match quality are decisive.

2. How a Programmatic Impression Is Transacted

A typical open-auction sequence is:

1. A publisher makes an impression available.
2. The publisher’s technology or SSP sends a bid request containing permitted signals, such as placement, format, device, approximate location, content category, and an identifier or addressability signal where lawful.
3. Multiple DSPs evaluate the request against campaign rules.
4. Each eligible DSP submits a bid and, commonly, a creative reference.
5. The marketplace selects a winner according to its auction and deal rules.
6. The creative is returned, rendered, and measured.
7. Billing, verification, and reporting systems reconcile the event.

Latency is a material constraint. A bid decision must be made within a short time window. Consequently, audience models and campaign rules are preprocessed, while only the final eligibility and value decision occurs in the auction.

Auctions and price formation

An auction determines which eligible demand source wins and what it pays. Auction mechanics differ by platform and contract. In a simplified first-price auction, the winner pays its submitted bid, subject to currency, fees, floors, and platform rules. In a second-price-like mechanism, the winner pays an amount related to the next highest eligible bid plus an increment, although many modern markets have moved toward first-price or hybrid designs.

The buyer should not assume that a nominal bid equals the publisher’s net revenue. The transaction may include DSP fees, SSP fees, data charges, verification costs, currency conversion, and other deductions. Transparent fee and log-level reporting is therefore central to economic evaluation.

A useful planning metric is cost per thousand impressions:

Cost and impressions must refer to the same reporting scope and period. CPM is a buying-price metric, not a measure of effectiveness. A low CPM can be inefficient if viewability, attention, reach quality, or conversion value is poor.

For example, a campaign spending $12,000 for 4,000,000 served impressions has:

CPM = ($12,000 ÷ 4,000,000) × 1,000 = $3.00

If only 3,000,000 impressions are viewable, the viewable-impression cost is:

(12,000 ÷ 3,000,000) × 1,000 = $4.00

The second figure is not a replacement for CPM; it answers a different question.

3. Buying Methods and Market Structures

Open auction

Open auction inventory is available to many eligible buyers. It provides scale and flexible optimization but may expose buyers to variable quality, duplicated supply paths, counterfeit inventory, inappropriate content, and limited transparency. Controls such as allowlists, blocklists, sellers.json, SupplyChain object data, and independent verification reduce—but do not eliminate—these risks.

Private marketplace

A private marketplace (PMP) is a controlled auction or invitation-based arrangement in which selected buyers receive access to specified inventory. A deal identifier communicates terms such as publisher, placement, format, geography, audience, floor price, and dates.

PMPs can improve quality and predictability, but the label alone does not guarantee premium inventory. Buyers should verify actual delivery, fee structure, audience composition, viewability, fraud rates, and contractual rights.

Preferred deal

A preferred deal gives a buyer an opportunity to purchase inventory at an agreed fixed price before the publisher offers the impression to broader auction demand. The buyer usually has no obligation to purchase every impression. This arrangement offers access and price predictability but may not guarantee volume.

Programmatic guaranteed

Programmatic guaranteed combines automated execution with a committed delivery obligation. The buyer and seller agree on volume, price, dates, audience or placement, and creative requirements. It resembles direct buying operationally, while trafficking and reporting occur through programmatic infrastructure. It is useful when reach, premium context, or contractual certainty matters more than auction flexibility.

Direct and indirect supply

Direct supply usually represents a publisher or media owner’s authorized relationship. Indirect supply may pass through several intermediaries. More intermediaries can increase reach but also create duplicated auctions, fees, latency, and opportunities for misrepresentation. Supply-path optimization addresses this issue.

4. Header Bidding and Yield Management

In a traditional waterfall, a publisher ad server calls demand sources sequentially according to historical price or priority. A source rejected at a high level may never see an impression even if it would have paid more. Header bidding changes this by allowing multiple demand sources to submit bids concurrently, often before the publisher ad server makes its final decision.

The conceptual sequence is:

User request
     ↓
Publisher page or app identifies impression
     ↓
Multiple SSPs and exchanges receive an opportunity
     ↓
Bids return within a latency limit
     ↓
Publisher ad server compares eligible demand
     ↓
Creative is selected, rendered, and measured

Header bidding can improve competition and publisher yield, but it introduces complexity. Multiple SSPs may represent the same inventory, increasing duplication. Poorly designed wrappers can create latency, timeout bias, user-experience problems, and inconsistent consent handling. Server-side header bidding may reduce browser-side latency but can weaken some identity and auction signals.

For buyers, header bidding means that an apparent auction opportunity may not represent unique supply. The same impression can appear through several paths. Bid duplication can inflate reach estimates and obscure incremental value.

5. Identity, Addressability, and Data Governance

Programmatic targeting historically relied heavily on third-party cookies and mobile advertising identifiers. Browser restrictions, operating-system controls, regulation, consent requirements, and consumer expectations have reduced the reliability of these signals. The strategic response is not to find a single universal replacement but to use a portfolio of lawful, purpose-appropriate signals.

Data categories

Identity systems map interactions to a recognized person, household, device, account, or probabilistic group. Deterministic matching uses a known relationship, such as a logged-in account. Probabilistic matching infers a relationship from patterns such as device, browser, or network signals. Probabilistic methods are uncertain and should not be represented as individual certainty.

A responsible identity strategy specifies the entity being recognized, the purpose, retention period, consent basis, security controls, opt-out mechanism, and acceptable activation environments. Identity resolution should not be treated as permission. The ability to match data does not establish a lawful basis to use it.

Privacy-enhancing approaches

Important approaches include consented first-party activation, clean rooms, cohort or contextual methods, on-device processing, data minimization, and aggregated measurement. A clean room permits controlled analysis or matching between parties without freely exchanging raw records. Its usefulness depends on governance, minimum audience thresholds, query controls, security, and whether outputs can be used for the stated purpose.

Contextual targeting

Contextual targeting selects advertising opportunities based on the content or environment rather than an inferred individual profile. Modern contextual systems can use page text, semantic taxonomies, entities, sentiment, video transcripts, app metadata, and content adjacency. Contextual relevance can be especially valuable when addressability is limited.

Contextual targeting still requires governance. A keyword may occur in news about a harmful event rather than in content promoting it. Classification models can misinterpret sarcasm, multilingual content, or user-generated material. Buyers should combine contextual inclusion with suitability exclusions, human review for sensitive categories, and post-delivery verification.

6. Supply-Path Optimization

Supply-path optimization (SPO) is the systematic evaluation and selection of routes from buyer to publisher. The objective is not always to choose the shortest path. A longer route may offer unique inventory, better identity, stronger fraud controls, or a valuable contractual relationship. SPO seeks the best economic and strategic route for a defined objective.

A buyer may evaluate:

• publisher authorization and seller transparency;
• total and incremental fees;
• win rate and auction duplication;
• viewability, invalid traffic, and brand suitability;
• latency and timeout performance;
• deal execution and reporting quality;
• unique reach and frequency contribution; and
• data leakage and contractual controls.

A simple effective-cost comparison can be expressed as:

“Qualified impressions” must be defined in advance—for example, impressions meeting viewability, fraud, geography, and brand-safety requirements. The formula is not a universal effectiveness measure; it is a cost-efficiency measure conditional on the chosen quality definition.

SPO should be conducted with log-level or sufficiently detailed supply data. A path that appears inexpensive on a media invoice may be expensive after duplicate impressions, verification, and intermediary charges are considered.

7. Bidding, Optimization, and Measurement

A DSP may optimize toward impressions, reach, clicks, video completion, conversions, revenue, or predicted lifetime value. Optimization requires a clear objective and an appropriate measurement design.

For a conversion campaign, a simplified expected-value logic is:

The expression assumes that the conversion probability is estimated for the specific impression opportunity and that the value definition is consistent. It excludes costs unless the decision rule explicitly compares expected value with total cost. In practice, models must account for attribution uncertainty, delayed conversions, frequency, incrementality, and budget constraints.

A bid strategy can fail even when its prediction model is accurate. If the conversion tag is missing, the attribution window is unsuitable, the algorithm receives biased training data, or the campaign optimizes to cheap but nonincremental conversions, observed performance may be misleading.

Important measurement distinctions include:

• Served impression: an ad response was delivered according to the platform’s event definition.
• Viewable impression: the impression met a specified viewability standard.
• Completed view: a video reached its defined completion point.
• Attributed conversion: a conversion was assigned to an exposure under a chosen attribution rule.
• Incremental conversion: a conversion caused by advertising beyond what would have occurred without it.

Attribution is not the same as causation. Randomized holdouts, geo experiments, conversion lift studies, and carefully designed quasi-experiments can provide stronger evidence of incrementality than last-touch attribution.

8. Programmatic Deal Governance

Programmatic deal governance is the set of commercial, technical, legal, and operational controls governing a transaction. It should be designed before activation, not after a discrepancy occurs.

Deal specification

A deal record should identify:

1. buyer, seller, platform, and authorized intermediaries;
2. inventory type, placement, format, environment, and geography;
3. dates, volume, price, currency, floors, and billing basis;
4. audience or contextual rules and their data source;
5. creative specifications and approval procedures;
6. frequency, pacing, and exclusion rules;
7. measurement provider and reconciliation standard;
8. fraud, viewability, suitability, and invalid-traffic thresholds;
9. data-use, retention, security, and deletion obligations; and
10. remedies for underdelivery, noncompliance, or disputed measurement.

Technical governance

Implementation should validate consent signals, ads.txt or app-ads.txt authorization, sellers.json records, SupplyChain information, creative landing pages, malware scans, click trackers, and macros. A buyer should test whether a deal ID actually enforces the promised inventory and whether the DSP can report delivery by publisher, placement, device, and supply path.

Financial governance

Reconciliation should compare platform logs, publisher records, independent verification, and invoices. Discrepancies can arise from time zones, counting methodologies, invalid traffic, late events, currency conversion, auction fees, or creative failures. Contracts should state which source controls in a dispute.

Human governance

Automation requires accountable decision rights. A campaign should have named owners for strategy, privacy, brand suitability, creative approval, measurement, incident response, and budget authorization. A human should be able to pause a campaign quickly when a severe safety, legal, or reputational issue emerges.

9. Ethical, Regulatory, and Social Issues

Programmatic advertising raises questions beyond efficiency.

Privacy and autonomy. Extensive profiling can create surveillance concerns, especially where individuals do not understand the collection or use of data. Consent should be specific, informed, revocable, and technically respected. Sensitive categories require heightened caution and may be prohibited under law or policy.

Discrimination. Targeting and optimization can exclude or disadvantage groups through explicit rules or model behavior. Audiences, bids, delivery, and outcomes should be tested for disparate effects where relevant. Removing a protected attribute does not eliminate proxy discrimination.

Children and vulnerable audiences. Age-sensitive environments require stronger restrictions on targeting, data collection, creative, and measurement.

Brand safety and suitability. Brand safety concerns illegal or clearly harmful environments; suitability concerns whether an otherwise lawful environment fits a particular brand. These concepts should not be collapsed into one universal blacklist.

Environmental impact. Each bid request, data transfer, and rendering event consumes energy. Reducing duplicated supply paths, unnecessary bid requests, heavy creative, and low-value impressions can improve both efficiency and environmental performance.

Transparency and market power. Concentrated platforms may control access, measurement, identity, and fees simultaneously. Advertisers should seek independent measurement, contractual audit rights, interoperable data where possible, and clear disclosure of incentives.

10. An Applied Case: The Northstar Outdoor Campaign

The following is a fictional illustrative case.

Background

Northstar Outdoor, a national retailer, launches a six-week campaign for winter equipment. It wants incremental store visits and online sales among adults interested in outdoor recreation. The initial plan uses open-auction display, third-party audience segments, and a single DSP.

Problem

After two weeks, the dashboard reports a low CPM and many attributed conversions. However, sales lift is unclear. The advertiser discovers that 28 percent of served impressions fail its viewability threshold, supply reports show the same publishers through several SSPs, and the audience provider cannot clearly document data provenance. The campaign also reaches news content that the brand considers unsuitable.

Analysis

The media team separates four questions that had been conflated: delivery, quality, attribution, and incrementality. It audits the supply chain, compares platform and independent measurement, examines conversion paths, and creates a geographic holdout for a portion of the target market. It also reviews consent and audience contracts with legal and privacy specialists.

Approach and solution

Northstar reduces the open-auction allowlist, selects two authorized supply paths, and negotiates a PMP with outdoor publishers. It supplements consented first-party customer audiences with contextual categories such as hiking, skiing, and equipment reviews. Sensitive-news exclusions are applied with semantic verification rather than a simple keyword list.

The team adopts frequency controls across addressable environments where technically possible, uses an identity-limited reach model for non-addressable inventory, and evaluates the campaign using qualified CPM and incremental sales rather than attributed conversions alone. Creative variants are tested, but the test design keeps audience and placement conditions comparable.

Outcome

The revised plan produces fewer served impressions and a higher nominal CPM. It improves viewable delivery, reduces duplicated supply, and provides clearer publisher-level reporting. The holdout indicates that incremental sales are positive but lower than the original attributed-conversion report suggested. Northstar therefore maintains the PMP and contextual strategy while reducing spend on low-quality open-auction paths.

Learning points

First, cheap inventory is not necessarily efficient inventory. Second, identity and audience scale must be evaluated alongside permission and accuracy. Third, attribution dashboards can overstate causal impact. Fourth, supply-path decisions should integrate quality, fees, reach, and strategic value. Finally, governance is a performance capability, not merely a compliance burden.

11. Contemporary Developments and Best Practices

The field is moving toward hybrid addressability: consented first-party data, publisher authentication, contextual intelligence, clean rooms, modeled measurement, and aggregated reporting. Connected television and retail media are expanding programmatic buying while introducing new concerns about deduplicated reach, household identity, walled-garden transparency, and inconsistent standards.

Generative AI can assist with creative variants, semantic classification, anomaly detection, and log analysis. It should not be treated as inherently objective. Training-data bias, hallucinated classifications, disclosure requirements, copyright questions, and inappropriate automated decisions require human review and documented controls.

A graduate-level operating model should follow these practices:

1. Begin with a business objective and causal measurement plan, not a platform feature.
2. Define quality standards before bidding begins.
3. Use the minimum data necessary for the stated purpose.
4. Maintain clear separation between permission, identity, targeting, and measurement.
5. Compare supply paths using qualified outcomes and total cost.
6. Require auditable deal terms and independent verification for material spend.
7. Test frequency, reach, incrementality, and model drift.
8. Establish pause procedures and named accountability.
9. Review algorithmic outcomes for exclusion and unintended bias.
10. Reassess the programmatic architecture as browsers, regulation, and platform policies change.

Chapter Summary

Programmatic advertising is a software-mediated market connecting advertisers, agencies, publishers, ad servers, DSPs, SSPs, exchanges, data platforms, and verification providers. A DSP evaluates bid opportunities, while an SSP helps publishers package and sell inventory. Real-time bidding enables rapid impression-level decisions, but auctions differ in price mechanics and transparency.

Open auctions provide scale but require rigorous quality controls. PMPs, preferred deals, and programmatic guaranteed transactions provide increasing degrees of access, price, and delivery control. Header bidding increases publisher competition but can introduce latency and duplicated supply. Identity systems support addressability but do not themselves create consent or accuracy. Contextual targeting offers a privacy-resilient approach, although classification and suitability remain governance challenges.

Supply-path optimization evaluates authorization, fees, duplication, quality, reach, and reporting rather than simply choosing the shortest route. Measurement must distinguish served, viewable, attributed, and incremental outcomes. Effective governance integrates deal specifications, technical validation, financial reconciliation, privacy, brand suitability, security, and human accountability. The strategic role of the marketing manager is therefore to connect automated execution with transparent objectives, defensible data practices, and credible evidence of value.

Key Terms

Ad server: Technology that stores, selects, delivers, and measures advertising creative.

Ad exchange: A marketplace that facilitates automated transactions between advertising demand and supply.

Contextual targeting: Selecting inventory according to content or environment rather than an inferred individual profile.

Customer data platform (CDP): A system that organizes customer data, commonly emphasizing first-party records and activation.

Demand-side platform (DSP): Technology used by advertisers or agencies to evaluate and purchase digital impressions.

First-party data: Data collected directly by an organization through its own interactions.

Header bidding: A process allowing multiple demand sources to compete before a publisher ad server makes its final decision.

Identity resolution: Linking interactions to a person, household, device, account, or modeled group.

Incrementality: The additional outcome caused by advertising beyond the counterfactual outcome without advertising.

Open auction: A marketplace in which broadly eligible buyers compete for available inventory.

Private marketplace (PMP): An invitation-based auction or controlled marketplace for selected buyers.

Programmatic guaranteed: An automated transaction with agreed price and committed delivery terms.

Real-time bidding (RTB): Automated impression-level bidding conducted within a short transaction window.

Supply-chain object: Technical information describing the intermediaries involved in a programmatic transaction.

Supply-path optimization (SPO): Evaluation and selection of media routes between buyer and publisher.

Supply-side platform (SSP): Technology that helps publishers manage and sell advertising inventory.

Viewability: Whether an impression meets a defined opportunity-to-see standard.

Data clean room: A controlled environment for permitted analysis or matching without unrestricted exchange of raw data.

Review Questions

Short Answer Questions

1. What are the primary functional differences between a DSP and an SSP?
2. Why is CPM insufficient as a standalone measure of programmatic effectiveness?
3. What is the purpose of a private marketplace?
4. How does contextual targeting differ from identity-based behavioral targeting?
5. Why does identity resolution not automatically establish permission to use data?

Long Answer Questions

1. Explain the end-to-end process of a real-time bidding transaction and identify the major points at which quality or transparency can fail.
2. Compare open auctions, PMPs, preferred deals, and programmatic guaranteed buying across scale, control, price, and governance.
3. Discuss how header bidding changes publisher yield management and what risks it creates for advertisers.
4. Develop a governance framework for a large programmatic campaign, including technical, financial, privacy, and human controls.
5. Critically assess the claim that the decline of third-party cookies makes programmatic advertising impossible.

Analytical / Application Questions

1. An advertiser pays $25,000 for 5,000,000 served impressions. Independent verification finds 3,500,000 qualified impressions. Calculate the served CPM and qualified-impression CPM, then explain which decision questions each metric answers.
2. Design an SPO scorecard for choosing between two SSP paths. Specify at least six criteria and explain how you would prevent a low fee from outweighing severe fraud or suitability problems.
3. A campaign reports many attributed sales but shows no difference between exposed and holdout regions. Diagnose at least three possible explanations and recommend a revised measurement approach.

Multiple Choice Questions with Answers

Question 1

Which platform primarily enables advertisers to evaluate and purchase impressions across multiple inventory sources?

A. Demand-side platform

B. Supply-side platform

C. Publisher ad server

D. Data clean room

Correct Answer: A. Demand-side platform

Question 2

A campaign spends $12,000 for 4,000,000 served impressions. What is its served CPM?

A. $0.30

B. $3.00

C. $30.00

D. $300.00

Correct Answer: B. $3.00

Question 3

Which transaction type normally combines automated execution with a committed delivery obligation?

A. Open auction

B. Preferred deal

C. Programmatic guaranteed

D. Contextual targeting

Correct Answer: C. Programmatic guaranteed

Question 4

What is a central purpose of supply-path optimization?

A. To maximize the number of intermediaries in every transaction

B. To select media routes using cost, quality, authorization, and strategic criteria

C. To replace all publisher ad servers

D. To ensure every bid uses third-party cookies

Correct Answer: B. To select media routes using cost, quality, authorization, and strategic criteria

Question 5

Which statement best describes contextual targeting?

A. It selects inventory based primarily on the content or environment

B. It requires a deterministic identity for every impression

C. It is identical to frequency capping

D. It guarantees that content classification is error-free

Correct Answer: A. It selects inventory based primarily on the content or environment

Question 6

What is a principal risk of poorly designed header-bidding implementations?

A. They eliminate all competition among demand sources

B. They can create latency and duplicate representations of the same inventory

C. They prevent publishers from using ad servers

D. They make contextual signals legally impossible

Correct Answer: B. They can create latency and duplicate representations of the same inventory

Question 7

Why should attributed conversions not automatically be interpreted as incremental conversions?

A. Attribution assigns credit under a rule, whereas incrementality concerns causal lift

B. Attributed conversions are always counted before an ad is served

C. Incrementality refers only to impressions, not outcomes

D. Attribution can be used only for programmatic guaranteed deals

Correct Answer: A. Attribution assigns credit under a rule, whereas incrementality concerns causal lift

Question 8

Which data source is generally most directly controlled by the organization that collected it?

A. Third-party data

B. First-party data

C. Modeled data from an unrelated provider

D. Open-auction bid data

Correct Answer: B. First-party data

Question 9

A buyer pays $25,000 for 5,000,000 served impressions. What is the served CPM?

A. $0.50

B. $5.00

C. $20.00

D. $125.00

Correct Answer: B. $5.00

Question 10

Which statement about identity resolution is most accurate?

A. A successful match automatically creates legal permission to activate data

B. Identity resolution links signals to an entity but does not by itself establish consent or lawful purpose

C. Identity resolution always uses deterministic identifiers

D. Identity resolution eliminates the need for measurement governance

Correct Answer: B. Identity resolution links signals to an entity but does not by itself establish consent or lawful purpose