Theory 01
Network Effects: Value Depends on Other Users
Origin: Iansiti & Lakhani (2020); IEB Session 2 theme
A network effect exists when the value of a product or service for one user increases as more users, complementors, or data-producing participants join the same system. This is a demand-side economy of scale: the product becomes more valuable because adoption itself changes the value proposition.
This differs from ordinary scale economies. A traditional factory becomes cheaper per unit when it produces more. A digital network becomes more useful when more people, data, or complements participate. That is why digital markets often become strategically fragile: once one firm pulls ahead, the advantage can reinforce itself.
Direct Network Effects
More users on the same side increase value for each other. Messaging, payment networks, and collaboration tools become useful because others are present.
Indirect Network Effects
More users on one side attract complementors on another side, which then makes the platform more valuable for the original users.
Data Network Effects
More usage generates more data, which improves algorithms, which improves the product, which attracts more usage.
Switching Costs
Users may stay because their contacts, history, workflows, integrations, or data are already embedded in the network.
Theory 02
Positive Feedback and Market Tipping
Origin: Iansiti & Lakhani (2020); Shapiro & Varian-style information economics
Network markets are shaped by positive feedback: success increases the probability of further success. More users generate more value, more complements, more data, and more legitimacy. This can lead to market tipping, where one or a few firms capture a disproportionate share of the market.
Critical Mass
A network product may be unattractive before enough users join. Once it passes a threshold, adoption becomes self-reinforcing. This explains why firms may subsidize early users, offer free tiers, spend heavily on acquisition, or tolerate losses during the launch phase.
| Stage | Strategic Problem | Typical Response |
|---|---|---|
| Pre-critical mass | The product lacks value because too few others use it. | Subsidies, free access, seeding complements, partnerships, compatibility. |
| Growth loop | More users improve value, but rivals may copy features. | Build data advantage, strengthen complements, reduce friction, improve matching. |
| Lock-in | Users depend on the installed base and ecosystem. | Switching costs, standards, APIs, identity, reputation, workflow integration. |
| Tipping | A leader can become difficult to displace. | Regulatory attention, niche entry, differentiation, multi-homing strategies. |
Theory 03
Platforms and Multi-Sided Digital Markets
Origin: IEB digital markets theme; Parker, Van Alstyne & Choudary logic used in later platform sessions
A platform creates value by enabling interactions between distinct participant groups. The platform's strategic problem is not only to sell a product, but to govern participation: who can join, what they can do, how quality is controlled, and how value is captured without weakening the network.
| Platform Element | Meaning | Strategic Tension |
|---|---|---|
| Sides | Distinct groups such as users, developers, advertisers, sellers, buyers, or data providers. | One side may need subsidies to attract the other side. |
| Core interaction | The repeated value-creating exchange the platform enables. | The platform must reduce search, matching, trust, and transaction friction. |
| Governance | Rules, standards, moderation, access, pricing, and quality control. | Too open can reduce quality; too closed can reduce innovation. |
| Value capture | How the platform monetizes access, transactions, data, tools, or premium features. | Charging too early or on the wrong side can reduce network growth. |
Theory 04
Winner-Take-All and Digital Market Vulnerability
Origin: Iansiti & Lakhani (2020)
Digital markets can become winner-take-all or winner-take-most when network effects, low marginal costs, learning effects, and switching costs reinforce one another. This is why Iansiti & Lakhani describe traditional markets as newly vulnerable: a digital entrant may not need to own the same physical assets as incumbents to compete at scale.
Low Marginal Cost
Once software is built, serving additional users is often cheap compared with human-heavy firms.
Learning Effects
Usage creates data; data improves algorithms; better algorithms attract more usage.
Complementor Pull
Developers, suppliers, and partners prefer the platform with the most users and strongest growth.
Exam Framework
Strategic Responses in Network Markets
A strong exam answer should not simply say "network effects create monopoly." The better answer asks which network effect exists, how strong it is, whether users multi-home, whether switching costs are high, and whether complements are exclusive.
Managerial Levers
- Build critical mass: Use subsidies, partnerships, compatibility, or free access to overcome the cold-start problem.
- Encourage complements: APIs, developer tools, data access, or marketplace infrastructure can strengthen indirect network effects.
- Manage openness: Open systems grow faster, but closed systems may capture more value and maintain quality.
- Reduce multi-homing: Firms may use identity, data history, workflow integration, or exclusive complements to make switching less attractive.
- Differentiate against giants: Newcomers can focus on niche communities, trust, regulation, specialized data, or superior domain knowledge.
Case Application
Novo Nordisk — AI Ecosystem Dependency
Network effects matter for Novo because the OpenAI partnership is not just a vendor relationship. It connects Novo to a broader AI ecosystem where model providers, cloud infrastructure, data, developers, and enterprise users reinforce one another.
- OpenAI's possible network effects: More enterprise users generate more feedback, integrations, use cases, and developer attention.
- Novo's counterweight: Novo owns specialized biomedical data and domain expertise, which may be harder for a general AI platform to replicate.
- Dependency risk: If Novo's workflows, prompts, validation routines, and data pipelines become tightly integrated with one AI platform, switching costs rise.
- Strategic response: Novo should preserve portability, internal AI competence, and governance over proprietary clinical data.
Exam Preparation
Likely Oral Exam Questions
-
Core What is the difference between direct and indirect network effects? ▶
- Direct: More same-side users increase value, such as more people on a messaging app.
- Indirect: More users attract complements, such as more apps for a mobile operating system.
- Data network effects are related but distinct: more use improves model quality through more data.
-
Core Why do network markets sometimes tip toward one dominant firm? ▶
- Positive feedback makes success self-reinforcing.
- Low marginal costs let digital leaders scale quickly.
- Complementors and users often prefer the largest network.
- Switching costs and data advantages make late entry harder.
-
Synthesis Could OpenAI become a strategic dependency for Novo Nordisk? ▶
- Yes, if Novo's research workflows and data pipelines become deeply tied to OpenAI's ecosystem.
- The risk is not only technical lock-in but also bargaining power, knowledge leakage, and reduced strategic autonomy.
- Novo can reduce this by keeping internal AI expertise, preserving data portability, and using governance clauses.