Affiliate, Influencer and Performance Partnerships

Introduction and Context

Performance partnerships are arrangements in which an organization collaborates with external individuals, publishers, creators, agencies, technology platforms, or complementary firms to generate measurable business outcomes. Unlike purely awareness-oriented sponsorships, performance partnerships link at least part of partner compensation to outcomes such as qualified leads, completed applications, subscriptions, sales, or retained customers.

Affiliate marketing traditionally uses tracked links, codes, or feeds to compensate publishers for referred transactions. Influencer marketing uses creators’ credibility, audiences, and content capabilities to affect awareness, consideration, and conversion. Strategic performance partnerships extend beyond both categories: examples include co-marketing with a software firm, referral arrangements with professional advisers, marketplace integrations, loyalty-program alliances, and lead-generation agreements with comparison websites.

The central managerial challenge is not simply recruiting many partners. It is designing a system that produces incremental, profitable, brand-safe, legally compliant, and strategically valuable demand. This requires alignment among five elements:

1. a clearly defined business outcome;
2. an economically defensible commission structure;
3. reliable attribution and tracking;
4. partner recruitment, enablement, and governance; and
5. continuous evaluation of quality, incrementality, and fraud risk.

A partner may appear productive because it receives credit for customers who would have purchased without the partnership. Conversely, a partner may influence discovery or consideration without receiving last-click credit. Graduate-level performance marketing therefore requires both operational measurement and causal reasoning.

Learning Objectives

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

• distinguish affiliate, influencer, referral, and broader performance-partnership models;
• design commission structures using contribution economics and customer quality;
• specify tracking architectures and attribution rules;
• recruit, classify, brief, and enable suitable partners;
• develop essential contractual and governance provisions;
• assess influencer effectiveness beyond vanity metrics;
• identify lead-quality problems and common forms of partner fraud;
• evaluate incrementality using experiments and quasi-experimental methods; and
• make budget, optimization, and termination decisions using managerial evidence.

1. The Partnership Value Chain

A performance partnership can be understood as a value chain rather than a single promotional placement:

Partner selection → audience exposure → interaction → referral or lead → conversion → fulfilment → retention → measurement and payment

Value can be created at any stage. A creator may stimulate demand weeks before a customer later converts through a search engine. A comparison publisher may provide high-intent traffic but little original persuasion. A professional referral partner may generate fewer leads, yet those leads may have substantially higher lifetime value.

Important partnership categories

These categories overlap. An influencer can operate as an affiliate when a unique link or code determines compensation. An affiliate can also function as a content producer. Classification should therefore be based on the value proposition and measurement method, not merely the label used by the organization.

The partner proposition

A credible partner program answers four questions:

• Why should the partner participate? Revenue, audience value, exclusive content, product access, professional reputation, or strategic learning may matter.
• Why should the audience respond? The offer must be relevant, credible, understandable, and sufficiently valuable.
• Why should the firm pay? The outcome must contribute economically after product costs, fulfilment, incentives, refunds, and partner compensation.
• Why should the relationship continue? Both parties need predictable processes, fair attribution, timely payment, and evidence of mutual value.

2. Economic Design of Commission Structures

Commission design converts a commercial objective into an incentive system. The correct rate depends on margin, customer value, conversion quality, risk, and the amount of work performed by the partner.

For a transaction-based program, contribution after partner cost can be represented as:

where:

• C = contribution after variable costs and partner compensation, in currency;
• R = recognized revenue;
• V = variable product, fulfilment, payment, and service costs;
• P = partner commission or fee; and
• F = other directly attributable variable marketing or incentive costs.

The relationship is applicable when revenue and costs are measured on a comparable transaction basis. It is not a complete profitability measure because fixed costs, overhead, and long-term customer value are excluded.

If commission is a percentage r of revenue, then P = rR. To ensure a minimum contribution C_min, the maximum commission rate is:

This calculation assumes that the percentage commission is applied to the same revenue base used in the equation. Contracts must specify whether discounts, taxes, shipping, refunds, cancelled orders, and repeat purchases are included.

Worked example: sale commission

A firm sells a subscription for $120 in recognized first-period revenue. Variable service and payment costs are $42, other attributable incentives are $8, and management requires at least $40 contribution before fixed costs. The maximum commission is:

r_max = (120 − 42 − 8 − 40) ÷ 120 = 30 ÷ 120 = 0.25

The maximum economically permissible rate is therefore 25%. A proposed 30% commission would produce only $34 contribution and fail the stated threshold.

Customer lifetime value and payback

For subscription or repeat-purchase businesses, first-order contribution may understate value. A simplified contribution-based customer lifetime value is:

where m_t is contribution margin per active customer period, s_t is the probability that the customer remains active in period t, d is the discount rate per period, T is the evaluation horizon, and A is acquisition cost, including partner compensation. This model assumes that retention probabilities and contribution estimates are sufficiently stable for the chosen horizon. It should not be treated as a precise forecast when retention is highly uncertain or cohorts are immature.

A simpler operational metric is payback period: the number of periods required for cumulative contribution to recover acquisition cost. Partner programs should establish a maximum acceptable payback because high lifetime value does not automatically justify unlimited upfront expenditure.

Commission structures

Common structures include:

• Cost per sale: payment for a validated transaction.
• Cost per qualified lead: payment when a lead satisfies defined criteria and passes validation.
• Cost per action: payment for a specified application, registration, booking, or trial.
• Revenue share: payment based on a percentage of recognized revenue.
• Tiered commission: higher rates after volume or quality thresholds are achieved.
• Hybrid compensation: fixed content or placement fee plus variable commission.
• Bonus and malus mechanisms: additional payment for quality, retention, or strategic outcomes, and reductions for invalid, refunded, or non-compliant activity.

Tiered commissions should not reward volume at the expense of quality. For example, a partner generating 1,000 leads with a 2% acceptance rate may be less valuable than one generating 200 leads with a 30% acceptance rate. A quality-adjusted schedule can use accepted applications, funded accounts, or retained customers rather than raw lead counts.

3. Tracking, Attribution, and Data Architecture

Tracking should be designed before recruitment. A minimally reliable system records the partner identifier, campaign, creative or placement, timestamp, landing page, consent status where required, conversion event, order or lead identifier, value, status, and subsequent validation outcome.

Tracking methods

• Tagged URLs: Parameters identify partner, campaign, placement, and creative. They are simple but can be lost through redirects, app transitions, or copied links.
• First-party cookies: Useful for browser-based journeys, but affected by consent choices, browser restrictions, deletion, and device switching.
• Promo or referral codes: Helpful for offline, creator, and podcast contexts, though codes can be shared or used after exposure from another channel.
• Server-to-server or postback tracking: Sends validated events between systems and is generally more robust than relying only on browser events.
• Platform reporting: Convenient but may use platform-specific definitions, windows, and modeled conversions.
• CRM matching: Connects leads to downstream qualification, revenue, retention, and refund outcomes.

No tracking method is complete. A robust architecture reconciles partner-platform events, web analytics, order systems, and CRM records. It also documents event definitions and deduplication rules.

Attribution versus incrementality

Attribution allocates observed credit among touchpoints. Incrementality asks whether the outcome would have occurred without the partner exposure. These are related but distinct questions.

A last-click model assigns all conversion credit to the final measurable referrer. It is simple and operationally useful but may overvalue coupon sites, branded search interception, or retargeting. A linear model divides credit equally among recorded touchpoints. Position-based or algorithmic models may reflect interaction patterns, but their output remains dependent on observed data and assumptions. None proves causal impact.

A partner program should therefore report at least three layers:

1. Delivery: clicks, views, reach, and content output.
2. Attributed performance: leads, sales, revenue, commission, and return on ad spend.
3. Business quality and causality: accepted leads, refunds, retention, new customers, control-group lift, and incremental profit.

For partner-attributed return on ad spend:

ROAS is a ratio, not profit. It is meaningful only when the revenue definition, attribution window, returns treatment, and cost base are clearly specified.

4. Recruitment, Segmentation, and Enablement

Recruitment should begin with a partner hypothesis rather than a list of popular personalities. The firm should identify the audience, need state, purchase context, trust relationship, geographic scope, and partner capability required.

Potential partners can be scored against weighted criteria. A simple evaluation score is:

where x_i is the normalized score for criterion i, w_i is its weight, and the weights sum to 1. The method assumes that criteria are sufficiently defined and that scores are comparable. It supports prioritization; it does not replace due diligence.

Useful criteria include audience relevance, engagement quality, content credibility, historical conversion, compliance record, brand fit, geographic reach, technical capability, exclusivity constraints, and estimated incremental reach. Audience size should normally receive less weight than relevance and quality.

Recruitment process

1. Define the commercial and audience objective.
2. Develop a partner profile and exclusion criteria.
3. Discover candidates through networks, platforms, social listening, industry communities, and existing customer referrals.
4. Audit audience authenticity, content history, disclosures, conflicts, and reputation.
5. Conduct a commercial and operational fit assessment.
6. Pilot with limited spend, clear learning objectives, and a defined review date.
7. Scale only after quality and compliance evidence are established.

Enablement includes product education, approved claims, brand assets, disclosure guidance, landing pages, tracking links, creative examples, FAQs, escalation routes, and payment schedules. Overly restrictive briefs can suppress authenticity; overly vague briefs can create legal, reputational, and measurement risk.

5. Contracts and Governance

A performance partnership agreement should translate strategic intent into enforceable operating rules. Important provisions include:

• parties, scope, territory, term, and exclusivity;
• approved promotional methods and prohibited traffic sources;
• commission basis, payment timing, validation, reversal, refunds, and taxes;
• attribution window, tracking hierarchy, deduplication, and reporting access;
• intellectual-property ownership and content usage rights;
• claims substantiation, disclosure, privacy, and data-processing obligations;
• brand-safety standards and approval procedures;
• confidentiality, cybersecurity, and incident notification;
• audit rights and record-retention requirements;
• suspension, clawback, termination, and post-termination obligations; and
• dispute resolution and applicable law.

Contracts should avoid vague terms such as “qualified lead” unless qualification is operationally defined. A qualified lead may require valid contact details, consent, serviceability, non-duplication, target geography, minimum need, and confirmation by a sales or verification process. The definition should also specify the time allowed for rejection and the evidence required.

Influencer agreements require additional precision concerning number and format of posts, timing, platform, usage rights, whitelisting or paid amplification, exclusivity, editing, disclosure, prohibited claims, and conduct-related termination. Payment for reach alone may be appropriate for an awareness objective, but it should not be reported as performance compensation unless a measurable outcome is also defined.

6. Measuring Influencer Effectiveness

Influencer effectiveness depends on the campaign objective. A reach campaign may emphasize unique reach and completed views; a consideration campaign may examine qualified traffic, content saves, branded search changes, or assisted conversions; a conversion campaign should emphasize validated sales, contribution, new-customer rate, and incremental lift.

Useful measures include:

• Engagement rate: interactions divided by a defined audience denominator. The denominator must be stated because followers, impressions, and reach produce different results.
• View-through rate: completed or threshold views divided by eligible video starts or impressions, depending on the platform definition.
• Click-through rate: clicks divided by impressions or link views, with the denominator explicitly reported.
• Cost per qualified lead: total campaign cost divided by validated qualified leads.
• Cost per incremental customer: campaign cost divided by additional customers attributable to the intervention relative to a valid comparison group.
• Creator-attributed contribution: contribution margin after creator fees, commissions, incentives, refunds, and other direct costs.

High engagement can reflect controversy, giveaways, or audience behavior that does not translate into commercial value. Follower counts can include inactive, purchased, duplicated, or geographically irrelevant accounts. Therefore, firms should inspect audience geography, age or eligibility where lawful and relevant, comment quality, repeated engagement patterns, traffic sources, conversion cohorts, and historical content consistency.

7. Lead Generation, Quality, and Fraud Risk

Lead programs create a measurement problem because a submitted form is not equivalent to commercial value. A lead funnel may proceed from submission to validation, contact, qualification, opportunity, sale, activation, and retention. Each stage should be tracked.

A useful diagnostic is the stage conversion rate:

where N_a is the number entering stage a and N_b is the number reaching stage b. The formula assumes that both counts refer to the same cohort and observation period.

For example, 2,000 submitted leads produce 1,200 validated leads, 300 sales opportunities, and 90 customers. Validation is 60%, opportunity conversion from validated leads is 25%, and customer conversion from opportunities is 30%. The overall submission-to-customer rate is 4.5% because 90 ÷ 2,000 = 0.045. The partner should not be evaluated solely on the 2,000 submissions.

Common fraud and abuse patterns include:

• duplicate or fabricated leads;
• incentivized submissions without genuine intent;
• click injection, cookie stuffing, or forced redirects;
• adware and browser extensions that overwrite referrals;
• trademark bidding where prohibited;
• unauthorized coupon or brand-code distribution;
• bot traffic and automated form completion;
• fake engagement or purchased followers; and
• self-referrals, collusion, or transaction cancellation after commission payment.

Controls include unique identifiers, duplicate checks, CAPTCHA or behavioral screening where proportionate, velocity rules, IP and device anomaly review, source-level monitoring, delayed commission validation, refund clawbacks, affiliate approval, code restrictions, and manual investigation of outliers. Controls must be risk-based: excessive friction can reduce legitimate conversion and unfairly penalize partners.

8. Incrementality and Causal Evaluation

Incrementality is the additional outcome caused by the partnership. The fundamental comparison is:

where Y_T is the outcome rate or contribution for the treatment group exposed to the partnership and Y_C is the corresponding outcome for a valid control or counterfactual group. If using rates, both terms must use comparable denominators.

A randomized holdout experiment may assign eligible audiences, regions, or partner opportunities to treatment and control. For example, if the treatment group has a purchase rate of 6.0% and the control group has 4.5%, absolute lift is 1.5 percentage points and relative lift is (0.060 − 0.045) ÷ 0.045 = 33.3%. The incremental customers are estimated by multiplying the absolute lift by the number of eligible treatment units, subject to sound randomization and adequate sample size.

Where randomization is impractical, firms may use geo experiments, matched markets, difference-in-differences, or propensity-score methods. These approaches require assumptions such as comparable trends, limited interference, stable measurement, and adequate control of confounding. They produce estimates, not certainty.

Incremental profit should incorporate the cost of the intervention:

A partner with strong attributed ROAS may have weak incremental profit if it mainly captures existing demand. Conversely, a creator with limited direct-code sales may generate positive incremental profit through discovery if a properly designed study demonstrates lift.

9. Illustrative Case: Northstar Home Insurance

This is a fictional illustrative case. Northstar Home Insurance wants to acquire younger homeowners through affiliates and personal-finance creators. Its existing program pays 12% of first-year premium for every tracked policy. The program produces many applications, but policy activation is low and refund rates are high.

Background and problem

The last-click dashboard reports $1.8 million in attributed premium and $216,000 in commissions. However, 18% of applications are duplicates or invalid, 22% of activated policies cancel within 60 days, and coupon-style partners receive substantial credit for branded searches. Management cannot determine whether creators generate new demand or merely redirect customers already intending to purchase.

Analysis

Northstar maps the funnel from click to quote, application, validation, activation, and 60-day retention. It discovers that partner quality varies substantially. Personal-finance educators generate fewer applications but higher validation and retention. Cashback partners generate volume but weak incremental evidence. The company also finds that the commission base does not clearly exclude cancelled policies.

Approach and solution

Northstar redesigns the program in five ways:

1. It changes the primary payout from raw application to activated policy that remains active beyond a defined validation period.
2. It introduces a smaller initial payment and a retention bonus, with clawback for cancellation or fraud.
3. It separates creator education partnerships from coupon and cashback affiliates in reporting.
4. It creates unique first-party links and codes, reconciles events with the CRM, and establishes one deduplication hierarchy.
5. It conducts a randomized regional holdout for selected creators, comparing eligible exposed regions with matched control regions while monitoring parallel trends.

The company also adds contract provisions covering insurance-claim language, disclosure, prohibited targeting, data handling, audit access, and termination for misleading content.

Outcome

After one quarter, raw applications decline, but validated applications and retained policies rise. Management observes that the creator segment has lower attributed volume than cashback partners but higher contribution per activated policy. The holdout analysis estimates positive incremental lift for two creators and no reliable lift for one large partner. Northstar increases investment in the first two, renegotiates the third’s compensation around validated outcomes, and reduces payments to partners whose activity cannot be verified.

Learning points

The case demonstrates that volume is not the same as value, last-click credit is not causality, and a commission plan can shape partner behavior. It also shows why contracts, CRM integration, quality controls, and experimentation must be designed as one operating system.

10. Managerial Decisions and Best Practices

Managers should treat a partner program as a portfolio. Different partners may serve discovery, education, conversion, retention, or market access. A single universal commission rate can distort the portfolio because partners contribute at different stages and incur different risks.

Best practices include:

• define the outcome and economic threshold before setting the commission;
• distinguish attributed results from incremental results;
• pay for validated quality rather than easily inflated volume;
• use first-party and server-side data where lawful and technically appropriate;
• document tracking windows, exclusions, and deduplication rules;
• pilot before scaling and reserve budget for experimentation;
• review cohorts for retention, refund, and downstream value;
• monitor partners at source, placement, creative, and geographic levels;
• use contracts that make claims, disclosure, data, and payment rules explicit;
• create escalation and termination procedures before an incident occurs; and
• protect creator authenticity while maintaining required compliance controls.

Ethical, legal, and governance considerations

Partners influence consumers through trust and social relationships, creating obligations beyond ordinary media buying. Material commercial relationships should be disclosed clearly and prominently according to applicable law and platform rules. Claims must be truthful, substantiated, and appropriate to the product category. Sensitive products may require additional restrictions concerning age, financial vulnerability, health, employment, or personal data.

Organizations should avoid discriminatory targeting, opaque lead resale, dark patterns, undisclosed incentives, and excessive collection of personal information. Consent, purpose limitation, data minimization, security, and deletion requirements should be incorporated into the operating design. Automated fraud scoring and AI-generated content should be tested for false positives, bias, disclosure failures, and unauthorized claims. Human review remains important for high-impact decisions and reputationally sensitive content.

Contemporary Developments

Several developments are changing partner management. Privacy changes and declining availability of third-party identifiers increase the importance of first-party data, server-side event collection, consent-aware measurement, and modeled reporting. Creator commerce is becoming more integrated with storefronts, live shopping, short-form video, and social checkout. Partnership platforms increasingly automate recruitment, payment, fraud detection, and optimization, but automation can amplify poor definitions or biased historical decisions.

Generative AI can support partner discovery, briefing, content adaptation, and anomaly detection. It should not be used to fabricate testimonials, impersonate creators, conceal sponsorship, or make unsupported product claims. Synthetic content and automated outreach require clear ownership, review, and disclosure policies.

Chapter Summary

Affiliate, influencer, and performance partnerships connect external relationships to measurable commercial outcomes. Effective programs begin with a defined objective and an economic model that protects contribution after variable costs. Commission structures should reward validated quality, customer value, and appropriate partner effort rather than raw volume alone.

Tracking requires explicit event definitions, reliable identifiers, CRM reconciliation, and documented attribution rules. Attribution reports observed credit; incrementality estimates the additional outcome caused by the partnership. Both are useful, but they answer different questions. Influencer effectiveness should be assessed against the campaign objective and should extend beyond followers, impressions, and engagement to qualified traffic, contribution, retention, and causal lift.

Recruitment, enablement, contracts, fraud controls, privacy, disclosure, and brand safety form an integrated governance system. The most mature programs operate as portfolios, pilot before scaling, use experiments where feasible, and allocate budget according to incremental and durable value.

Key Terms

Affiliate marketing: A partnership model in which external partners are compensated for tracked actions or outcomes.

Attribution: Allocation of observed conversion credit among marketing touchpoints.

Clawback: Recovery or reversal of compensation after cancellation, refund, fraud, or invalidation.

Commission: Variable compensation paid to a partner for a defined outcome.

Conversion window: The period during which a touchpoint may receive credit for a conversion.

Deduplication: Rules for preventing multiple partners or systems from being paid for the same outcome.

Incrementality: The additional outcome caused by an intervention compared with a valid counterfactual.

Influencer marketing: Use of creators’ audiences, credibility, and content to affect consumer behavior.

Lead quality: The degree to which a lead is valid, relevant, qualified, and commercially valuable.

Lifetime value: The expected contribution from a customer over a defined future horizon.

Partner enablement: Training, assets, tools, guidance, and support that help partners perform correctly.

Postback tracking: Server-to-server transmission of conversion or event information.

Qualified lead: A lead meeting pre-defined validity, relevance, consent, and business criteria.

Referral marketing: Acquisition stimulated through recommendations from customers or trusted intermediaries.

ROAS: Revenue attributed to a channel divided by its defined cost.

Tiered commission: A schedule in which compensation changes after volume or quality thresholds.

Tracking parameter: A coded value appended to a URL or event to identify source and campaign information.

Validation period: A specified period during which an outcome must remain valid before final payment.

Review Questions

Short Answer Questions

1. How does affiliate marketing differ from influencer marketing?
2. Why is a qualified lead definition important in a partner contract?
3. State two limitations of last-click attribution.
4. What is the purpose of a commission clawback?
5. Name two methods for detecting partner fraud.

Long Answer Questions

1. Explain how a firm should design a commission structure using contribution economics and customer lifetime value.
2. Discuss the tracking and data architecture required for a reliable partner program.
3. Compare attributed performance with incremental performance and explain why both should be reported.
4. Evaluate the main contractual and governance requirements for influencer partnerships.
5. Explain how a manager can assess influencer effectiveness without relying primarily on follower counts.

Analytical / Application Questions

1. A partner generates 800 leads, of which 560 are validated, 140 become opportunities, and 42 become customers. Calculate the validation rate, validated-lead-to-opportunity rate, opportunity-to-customer rate, and overall lead-to-customer rate. What additional evidence would you require before increasing the partner’s commission?
2. A creator receives a fixed fee and reports strong attributed sales through a code. Design an experiment or quasi-experiment to estimate whether the creator generated incremental demand rather than merely capturing existing demand.
3. Design a partner scorecard for a subscription business. Include economic, quality, causal, compliance, and strategic measures, and explain how the scorecard would inform recruitment, optimization, and termination decisions.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes attribution from incrementality?

A. Attribution measures only awareness, whereas incrementality measures only clicks.

B. Attribution allocates observed credit, whereas incrementality estimates the additional outcome caused by an intervention.

C. Attribution is always experimental, whereas incrementality is always descriptive.

D. Attribution applies to affiliates, whereas incrementality applies only to influencers.

Correct Answer: B. Attribution allocates observed credit, whereas incrementality estimates the additional outcome caused by an intervention.

Question 2

A product generates $120 revenue, $42 variable costs, $8 other attributable incentives, and requires $40 contribution before fixed costs. What is the maximum percentage commission on revenue?

A. 20%

B. 25%

C. 30%

D. 35%

Correct Answer: B. 25%

Question 3

Which payment design most directly rewards durable customer quality?

A. Payment for every click

B. Payment for every raw form submission

C. Payment for an activated customer retained beyond a validation period

D. Payment based only on follower count

Correct Answer: C. Payment for an activated customer retained beyond a validation period

Question 4

What is the main purpose of deduplication rules?

A. To increase the number of reported conversions

B. To prevent multiple systems or partners receiving credit for the same outcome

C. To eliminate the need for contracts

D. To guarantee that every lead becomes a customer

Correct Answer: B. To prevent multiple systems or partners receiving credit for the same outcome

Question 5

A campaign has 2,000 submitted leads and 90 customers. What is the submission-to-customer conversion rate?

A. 2.5%

B. 4.5%

C. 9.0%

D. 22.2%

Correct Answer: B. 4.5%

Question 6

Which metric is most likely to be a vanity metric when used alone to evaluate an influencer?

A. Retained customers

B. Incremental contribution

C. Follower count

D. Validated applications

Correct Answer: C. Follower count

Question 7

Which practice is a common form of affiliate abuse?

A. Using approved creative assets

B. Providing transparent disclosures

C. Cookie stuffing or forced referral overwriting

D. Reconciling partner events with CRM records

Correct Answer: C. Cookie stuffing or forced referral overwriting

Question 8

If treatment purchase rate is 6.0% and control purchase rate is 4.5%, what is the absolute lift?

A. 1.5 percentage points

B. 4.5 percentage points

C. 6.0 percentage points

D. 33.3 percentage points

Correct Answer: A. 1.5 percentage points

Question 9

Which contract provision is most important for preventing disputes about lead payments?

A. A vague statement that leads must be useful

B. A precise definition of a qualified lead and the validation process

C. A requirement that partners maximize impressions

D. A ban on all performance reporting

Correct Answer: B. A precise definition of a qualified lead and the validation process

Question 10

Why might a partner with high attributed ROAS have low incremental value?

A. It may be capturing customers who would have converted without the partner exposure.

B. High ROAS always means the partner is fraudulent.

C. Attributed revenue cannot be measured in any partner program.

D. Incrementality is identical to revenue attribution.

Correct Answer: A. It may be capturing customers who would have converted without the partner exposure.