Session 9 – Viral Marketing, Centrality & Strength of Ties

Viral marketing

Uses customer networks to spread messages, build brand awareness, or support other marketing objectives; encourages people to pass along a message through WOM

Word of mouth (WOM)

Customer-to-customer communication about a product, service, or experience; not every recommendation becomes viral

Owned vs. Paid vs. Earned media

Owned = channels the company controls (website, blog, social accounts); Paid = channels the company pays to leverage (display, paid search, sponsorships, TV/print); Earned = customers become the channel through WOM, buzz, and viral sharing

Why is earned media especially important to viral marketing?

Customers themselves become the distribution channel and spread the message through their networks

Why is storytelling useful in viral marketing?

Stories are interesting/entertaining, grab attention, motivate sharing, and can build trust

Why can a story shared by another person be persuasive?

People may be more receptive to a personable story from a familiar source or friend than to a message coming directly from a brand

Four major advantages of viral marketing

Trust, targeting, engagement, and cost

Trust and targeting advantages of viral marketing

Trust = a familiar, credible sender can make a message more persuasive; Targeting = social contacts may know who would find a message relevant because they know one another's tastes

Engagement and cost advantages of viral marketing

Engagement = people may voluntarily share useful/interesting content; Cost = customer sharing can reduce paid distribution costs

Marketing communication cost in viral marketing

Cost involves both creation of the content/ad copy and sharing/distribution; consumers can provide a low-cost sharing mechanism, but strong creative content is still required

Three major disadvantages of viral marketing

Scale, predictability, and control

Scale and predictability problems in viral marketing

Scale = most content does not achieve massive reach; the lecture notes only 3% of YouTube videos have more than 25K views. Predictability = marketers cannot reliably know which content will go viral

Control problem in viral marketing

Once consumers begin sharing, marketers have less control; a message can be altered, misunderstood, hijacked, or interpreted differently

Small firms and viral marketing

Can benefit from an inexpensive advertising option because resources are limited, but smaller reach and loyal audiences can make successful viral diffusion harder

Large firms and viral marketing

Existing large networks and loyal customers can make viral diffusion easier and can reinforce brand identity

Why is viral marketing difficult to engineer?

A large amount of digital content goes unnoticed, and marketers cannot reliably create a story that people will choose to spread

Why does initial distribution matter in viral marketing?

Who receives the message can matter as much as the message itself; network structure and the centrality of initial recipients can affect diffusion

Centrality analysis

Identifies important structural positions in a network because not all nodes occupy the same role

What must be defined before choosing a centrality measure?

What "importance" means for the specific objective; different centrality measures capture different structural roles

Degree centrality

Counts the number of direct connections a node has; useful for identifying a popular or well-connected hub

Does high degree centrality guarantee influence?

No; having more direct connections creates more direct reach but does not necessarily mean greater influence

Adjacency metrics

Use a series of 0s and 1s to indicate where connections exist between nodes

Preferential attachment

Already powerful/popular nodes with more connections tend to attract additional connections

"Good gets better" vs. "rich get richer"

Good gets better = high-quality content organically attracts connections; Rich get richer = popularity itself attracts additional connections

Popularity effect

Users may want to associate with popular people, ideas, or items, further increasing their popularity regardless of objective measurable characteristics

How can firms leverage already-central nodes?

Pay influencers or affiliates with money or goods to share a message through their existing large/mature platforms

Betweenness centrality

Counts a node's share of shortest paths between other pairs of nodes; identifies bridges, bottlenecks, and information gatekeepers

Why is high betweenness useful for information diffusion?

A node can be important because it connects otherwise separate parts of a network, allowing information to pass between communities

Limitation of betweenness centrality

It assumes shortest-path routes, but real messages may travel through other routes

Example of high betweenness

A person connecting their home community with their Purdue community can act as a bridge because communication between the groups may pass through them

Closeness centrality

Measures how close a node is, on average, to all other nodes; high-closeness nodes can reach the whole network relatively quickly

Closeness centrality relationship

Closeness = 1 / average shortest-path distance to all other nodes

Eigenvector centrality

A node is important when it is connected to other important/high-centrality nodes; one's importance is influenced by the importance of one's connections

Degree vs. eigenvector centrality

Degree asks how many direct connections you have; eigenvector also considers how important/well-connected those connections are

Beta (β) in the eigenvector-centrality discussion

β captures the power of someone's connectivity

Summation of Aij in the centrality discussion

Represents degree centrality

Which centrality identifies someone with many direct connections?

Degree centrality

Which centrality identifies a bridge, bottleneck, or information gatekeeper?

Betweenness centrality

Which centrality identifies someone who can spread information quickly throughout the network?

Closeness centrality

Which centrality identifies structural influence or prestige from connections to important people?

Eigenvector centrality

Professor emphasis for centrality

You do NOT need to calculate the centrality measures, but you must understand which centrality type is appropriate for a given situation

Strength of ties

Connections are not all the same strength; strong and weak ties play different roles in community formation and information diffusion

Strong vs. weak ties

Strong ties = relationships among people who work, live, or play together and interact frequently; Weak ties = relationships between members of different groups that are used less frequently and require less attention/effort

How can tie strength be operationalized?

Frequency of communication/exchange, reciprocity of interaction, closeness, and type of interaction; reciprocal vs. one-way following can also provide information

Granovetter transitivity

If A and C are tied and B and C are tied, A and B are more likely to also become connected

How does tie strength affect transitivity?

The stronger C's ties to A and B, the more likely A and B are to form a tie

What drives transitivity?

Similarity breeds liking, cognitive dissonance, and spending substantial time with mutual connections

Strength of weak ties

Weak ties connect a person and their immediate social circle to other, different social circles

Why are weak ties valuable?

They provide access to new information, diverse ideas, and opportunities outside a person's immediate group and help new ideas spread between communities

Practical example of the strength of weak ties

Job opportunities can spread between different social groups through weak ties

Strong ties/transitivity vs. weak ties/bridging

Strong ties support transitivity and dense connections within communities; weak ties can bridge different communities and expose people to new information

Facebook sharing — news feed finding

People are more likely to share a link from a close friend than from an acquaintance

Facebook sharing — no-feed baseline

People are still more likely to share a close friend's link; strong ties have a higher baseline probability of sharing

Facebook sharing multiplier

Multiplier = sharing probability with feed / sharing probability without feed

What does the Facebook sharing multiplier measure?

How much seeing a link in the news feed changes sharing probability relative to the no-feed baseline

For which ties is the Facebook feed multiplier stronger?

Weak ties

Strong-tie sharing vs. weak-tie multiplier

Strong ties have the higher overall sharing probability, but the feed multiplier is larger for weak ties because feed exposure creates a larger relative increase over their lower baseline

Centrality measures — quick comparison

Degree = direct connections/popular hub; Betweenness = bridge/gatekeeper; Closeness = quickly reach whole network; Eigenvector = connections to important nodes

Viral marketing — big picture

Effective viral diffusion depends not only on creating shareable content but also on the customer network: who receives it, where those people sit in the network, and how strong their relationships are