Chapter 4 Exercises: The Business Model of Engagement

These exercises are organized into five tiers of increasing cognitive demand, from recall through synthesis and evaluation. Complete the lower tiers before attempting the higher ones; the later questions assume you have internalized the foundational material.

Tier 1: Recall

These questions test basic factual comprehension. Answers should be retrievable directly from the chapter text.

1.1 What does CPM stand for, and what does it measure?

1.2 List the five steps of a Real-Time Bidding auction in order, from the moment a user loads a page to the moment an ad appears.

1.3 What were the three variables that EdgeRank used to rank Facebook News Feed content in 2011?

1.4 Who are the three founders introduced in Velocity Media, and what are their roles?

1.5 In what year was the Facebook emotional contagion study conducted, and in what year were its results published? What was the approximate number of users whose feeds were manipulated?

Tier 2: Comprehension

These questions require you to explain concepts in your own words, make inferences from the text, and demonstrate understanding of causal relationships.

2.1 Explain in your own words why engagement metrics became the optimization target for advertising-supported platforms. What is the causal chain connecting ad revenue to engagement maximization? Write out the chain as a numbered sequence of steps.

2.2 The chapter describes a "data flywheel" that creates barriers to entry for new competitors. Explain how this flywheel works. What inputs feed it, and what outputs does it produce? Why does it make it difficult for a new platform to compete with established players on CPM rates?

2.3 The chapter distinguishes between "surveillance advertising" and "contextual advertising." Describe each approach and explain how they differ in their privacy implications, their CPM rates, and their engagement-optimization incentives.

2.4 What is Goodhart's Law, and how does it apply to the use of engagement metrics as proxies for user satisfaction? Use a specific example from the chapter to illustrate.

2.5 The chapter identifies several "compounding layers" of the incentive trap: measurement, competition, capital, talent, and regulatory vacuum. Describe each layer in two or three sentences. How do they interact to create a system that is resistant to change from within?

2.6 Explain the significance of the Like button as a data collection instrument. How did its introduction in 2009 change the kind of data Facebook had access to, and how did this change the behavior of EdgeRank's successor algorithms?

2.7 The chapter states: "A user who is mildly happy and mildly bored scrolls indefinitely. A user who is deeply satisfied closes the app and goes about their life." Explain what this means for the relationship between user happiness and advertising-supported platform revenue. Why might a platform's advertising model actually create incentives against deep user satisfaction?

2.8 Why does the subscription model create better incentive alignment between platform interests and user interests than the advertising model? What are the specific limitations that prevent subscription models from replacing advertising models for social media platforms?

Tier 3: Application

These questions ask you to apply chapter concepts to new scenarios, perform calculations, or analyze cases using the frameworks introduced in the text.

3.1 CPM Calculation. A platform has 50 million Daily Active Users. On average, each user sees 12 ads per day. The average CPM rate across all users and ad types is $6.50.

(a) How many total ad impressions does the platform serve daily? (b) What is the platform's daily advertising revenue? (c) What is the platform's annual advertising revenue? (d) If the platform's engineering team implements a new recommendation algorithm that increases average session time by 15%, and this produces a proportional increase in daily ad impressions, what is the new annual revenue? What is the dollar value of that 15% engagement increase?

3.2 CPM Calculation. An advertiser has a budget of $50,000 and wants to run a campaign targeting high-income professionals (ages 35–54, household income $150,000+). The relevant platform charges a CPM of $28 for this audience segment.

(a) How many impressions can the advertiser purchase with their full budget? (b) If the platform's click-through rate for this audience segment is 0.8%, how many clicks does the advertiser receive? (c) If 3.5% of clicks result in a conversion (purchase), how many conversions does the campaign generate? (d) If each conversion is worth $400 to the advertiser, what is the campaign's return on ad spend (ROAS)?

3.3 Apply the seven-step causal chain from the chapter (platforms need revenue → engagement maximization) to Velocity Media's situation at the time of its Series A. For each step in the chain, identify the specific Velocity Media equivalent: what revenue source, what inventory constraint, what engagement metric, what measurement proxy, what divergence between proxy and wellbeing, what content ecosystem effect, and what self-reinforcing dynamic?

3.4 Imagine you are Dr. Aisha Johnson, six months into your role at Velocity Media. You've confirmed that the recommendation algorithm systematically surfaces emotionally provocative content disproportionately relative to its educational value. Write a one-page internal memo to Sarah Chen and Marcus Webb making the business case — not just the ethical case — for rebalancing the algorithm. What arguments would you make? What data would you want to present? What risks would you need to acknowledge?

3.5 A social media platform is considering switching from a purely advertising-supported model to a hybrid model (free tier with ads, premium tier without). Using the chapter's analysis of Spotify's hybrid model, describe the specific changes you would expect to see in: (a) the platform's incentive to optimize engagement for premium users vs. free users; (b) the types of manipulative design patterns that would become more or less commercially necessary; (c) the platform's internal organizational structure and how it would likely evolve.

3.6 The chapter claims that "regulation works best when it targets the specific financial mechanism that creates the bad incentive." Design a hypothetical regulation specifically targeting the connection between engagement optimization and advertising revenue. Your regulation should: (a) identify the precise mechanism being targeted; (b) describe what behavior it would prohibit or require; (c) anticipate at least two ways platforms might attempt to circumvent it; (d) describe how the regulation might be enforced.

3.7 Compare the CPM structures for three different advertising contexts: (1) a high-intent search ad, (2) a behaviorally targeted social media feed ad, (3) a contextual display ad on a news site. For each, explain what drives the CPM rate, who the typical advertiser categories are, and what this implies about the value that different types of user attention hold in the market.

3.8 Using the five layers of the incentive trap (measurement, competition, capital, talent, regulatory vacuum), evaluate whether a hypothetical publicly traded social media company could successfully "choose" to prioritize user wellbeing over engagement maximization. For each layer, describe what obstacle the company would face and what, if anything, could overcome it.

Tier 4: Analysis

These questions require deeper analytical engagement: identifying assumptions, evaluating evidence, examining tensions and contradictions, and developing critical perspectives.

4.1 The chapter frames the incentive trap as a structural problem rather than a problem of individual moral failure. Critically evaluate this framing. In what ways is it accurate and helpful? In what ways might it let individual decision-makers off the hook too easily? Are there specific roles or decision points where individual moral responsibility is more significant than the structural analysis acknowledges?

4.2 The emotional contagion study (2012/2014) revealed that Facebook could deliberately alter users' emotional states through feed manipulation. The chapter notes that if Facebook could make users happier, the advertising model gave it no incentive to do so. Analyze the ethical implications of this capability-incentive gap. What specific obligations, if any, does possessing this capability create? How does the asymmetry of knowledge (Facebook knows it can do this; users don't) affect the ethical analysis?

4.3 The chapter introduces Velocity Media as a fictional example but uses it to illustrate "the structural pressures facing real platforms." Identify three respects in which Velocity Media's situation might be meaningfully different from the situations faced by large established platforms like Facebook or YouTube. How might these differences affect the applicability of the structural analysis?

4.4 The chapter's description of the real-time bidding auction emphasizes what happens in 100 milliseconds without user awareness. Analyze the informed consent implications of this system. What would "meaningful consent" to participation in this system actually require? Is meaningful consent achievable within the current architecture? What would need to change?

4.5 The chapter claims that advertising CPM rates create "stronger financial incentives to engage wealthy users than poor ones." Analyze the distributional justice implications of this claim. Who bears the greatest cost of engagement-maximizing design choices? Who captures most of the benefit? Does this distribution raise concerns beyond individual user harm?

4.6 Wikipedia is presented as evidence that "at massive scale, a platform can be organized around values rather than around an advertiser-serving engagement metric." Critically evaluate this claim. In what sense is Wikipedia a platform? How is its use case different from social media platforms? What specific structural features allow it to resist engagement-optimization pressure, and are those features replicable in social media contexts?

4.7 The chapter traces a causal chain from the business model to harmful outcomes. But correlation is not causation: advertising-supported platforms also produce enormous social value (free information access, global connection, creator income). Construct the strongest possible counter-argument to the chapter's central thesis. What would a defender of the advertising model say, and how would you respond?

Tier 5: Synthesis and Evaluation

These questions require integrating material across the chapter and beyond, generating original analysis, and defending evaluative judgments.

5.1 Design Challenge. Sarah Chen has asked you, as a consultant, to redesign Velocity Media's business model before its Series B funding round. Your constraint: the redesigned model must be commercially viable (able to attract investment and achieve profitability within three years) while materially reducing the engagement-maximization incentive. Propose a specific model, describe its revenue mechanics, estimate rough CPM/revenue equivalents, explain how it changes the incentive structure, and identify its main risks. Be specific — "subscription model" is not an answer; you need to describe exactly what users pay for, how much, and why they would.

5.2 The chapter argues that the business model is "the constraint that shapes everything." Evaluate this claim against an alternative hypothesis: that the constraint is not the business model itself but the competitive market structure that prevents any single platform from unilaterally changing its model. If the market structure hypothesis is correct, what does it imply about the effectiveness of regulation that targets individual platforms versus regulation that establishes industry-wide rules?

5.3 The Facebook News Feed Arc begins with EdgeRank in 2011 and ends (in this chapter) with the 2014 emotional contagion study. Using only the information provided in this chapter, construct a timeline of how the engagement-optimization logic evolved and intensified at Facebook from 2006 to 2014. At each stage, identify the specific design decision, the business rationale behind it, and the unintended consequence that followed.

5.4 Compare the three business model alternatives described in the chapter (subscription, cooperative, contextual advertising) along four dimensions: (a) alignment of platform incentives with user wellbeing; (b) commercial viability at social media scale; (c) accessibility to low-income users; (d) governance accountability. Which alternative performs best overall, and which dimension do you consider most important? Defend your choices.

5.5 Cross-Chapter Synthesis. Drawing on what you have learned in Chapters 1–4, develop a theory of how the business model of engagement connects to the concept of "algorithmic addiction" introduced in the book's title. What specifically does the advertising model need users to be addicted to? How does it create those addictions? Is "addiction" the right frame, or does it obscure as much as it reveals?

5.6 Policy Evaluation. Three different regulatory proposals have been put forward in response to the attention economy:

• Proposal A: Prohibit behavioral targeting entirely; require all digital advertising to be contextual.
• Proposal B: Require platforms to offer a paid, ad-free tier at a regulated maximum price, ensuring low-income users have access.
• Proposal C: Require platforms to measure and publicly report user wellbeing metrics alongside engagement metrics, with third-party auditing.

Evaluate each proposal against the chapter's framework. What specific incentive structure would each proposal change? What unintended consequences might each produce? Which proposal, in your view, most effectively targets the root cause of the problem?

5.7 The chapter opens with YouTube's 2006 founding decision as an illustration of how a seemingly administrative choice ("how do we pay for bandwidth?") becomes a determinative constraint on everything that follows. Write a 500-word essay arguing either (a) that this determinism is overstated — platforms have more agency to change their models than the chapter acknowledges — or (b) that the determinism is understated — the chapter focuses on the business model but misses other founding decisions (network effects, technical architecture, regulatory environment) that are equally or more constraining. Use specific examples to support your argument.