5. Networks and Viral Campaigns

Question 1

Diffusion

Answer 1

Diffusion is the process through which a set of actors acquire a given property over time

• Contracting a contagious disease
• Adopting a new product or service

Studying diffusion in a network context allows us to answer questions like:

• Under which conditions will an outbreak (set of adoptions) spread to a nontrivial portion of the population
• What percent of the population will eventually become infected?

Question 2

Diffusion of Innovations (Rogers)

Answer 2

Rogers defined 5 categories of adopters

• Innovators
• Early adopters
• Early majority
• Late majority
• Laggards

S-shaped adoption curve:

• Different types of people adopt at different times
• No network information

Question 3

SIR and SIS Model

Answer 3

SIR Model

• Susceptible: Nodes can be infected by their infected neighbors
• Infected: Nodes can stop being infected
• Removed: Nodes do not get infected and do not infect others
  
  • Immune
  • Dead

The SIS model is a variation of the SIR model in which nodes become susceptible again after infection.

Question 4

Threshold Models

Answer 4

Threshold models assume that individuals’ behavior depends on the number of other individuals already engaging in that behavior.

• For each individual, there is a point after which they shift their behavior
• Thresholds are specific for each individual
• A smooth S-shaped curve is usually obtained

Question 5

Strategies to foster product adoption

Answer 5

Strategies to increase adoption in a network setting include:

• Seeding central (important) people/users
• Engineering viral features in the product
• Incentivizing user with referrals

Question 6

Seeding

Answer 6

When considering seeding, we need to answer the following questions:

• What are the interventions we are considering
• Which customers should we target?
• How will targeted consumers react to our intervention?
• How to measure the impact of our intervention?

Question 7

The Influentials Hypothesis

Answer 7

Influentials: A minority of individuals who influence an exceptional number of their peers. They are important to the formation of public opinion

However:

• Large cascades of influence are driven not by influentials but by a critical mass of easily influenced individuals
• The influentials hypothesis requires more careful specification and testing

Question 8

Network Externalities

Answer 8

Externalities refer to situations where the payoffs to one individual are affected by the actions of others, where those actions do not directly involve the individual in question.

Question 9

Relevance of peer influence

Answer 9

• Peer effects are stronger in less popular products
• Almost all blockbusters (products adopted by millions of users) do not exhibit substantial peer influence at any stage of their lifecycle
• There is a discrepancy between observed local peer effects and the lack of peer influence in large adoption cascades that are characteristics of successful products

Question 10

Appeal vs Reach

Answer 10

• Influence-driven products can achieve a higher reach than preference-driven for moderate levels of appeal
• In cases where appeal is high, adoption for products are likely to occur even in the absence of influence

Question 11

Network Effects

Answer 11

Network effects play an important role in explaining why referral policies are more likely to be present in some platforms than others:

• Platforms that exhibit no network effects may use referrals to accelerate and expand their growth
• Platforms that exhibit network effects are less likely to request as users have direct benefits from inviting their friends to the platform

Question 12

Local and global network effects

Answer 12

• For some platforms, network effects are not important at a local scale but are important globally.
• In such cases, there are no direct incentives for inviting friends to the platform, leading to an under-supply of users.
• Platforms try to correct this distortion by creating incentives for users to invite their friends.
• Such requirements can backfire
  
  • Members may refuse to invite new members
  • Abandon the platform

Question 13

How do referral policies contribute to the growth of a platform?

Answer 13

• Harsher policies lead to more invites, sign-ups, approvals and payments, but to less activity.
• Individuals invited by members that got harsher policies pay more but are not more likely to invite new members.

Question 14

Target the Ego or Target the Alter:

Benefits or Pro-Active Churn Management

Answer 14

Proactive churn management improves customer retention and value:

• Churn reduced by 11% when compared to reactive churn management
• Expected lifetime valuation increased by 1.8%

Socially-based proactive churn management can be used to provide additional value on churn management.

• Additional increase of 3.6% in lifetime valuation for households managed together with their contacts
• Additional reduction of 8% on churn rate