V15: Network Effects — how to analyze it

The network effect = the strongest moat. Every new user makes the service more valuable to everyone else. Competitors cannot enter even if they have better technology — they lack the users. This is why Facebook defeated Google+.

This is V15 — Network effects in the AKM1 model: one of the 20 variables that together determine whether a company is an institutional quality stock or a company of stories. In this article you get the variable explained, how to calculate it yourself, and how three of history's greatest investors would have interpreted it.

Why the variable exists

Network effects have existed as long as human networks. The Roman road system (1st century BC) — more cities connected to the road network increased the value for all cities. The telephone network (Alexander Graham Bell 1876) — Metcalfe's law in its purest form. TV networks (1930s) — more viewers attracted more advertisers; more advertisers funded more content. Payment networks: Visa (1958) and Mastercard (1966) — more card-accepting stores attracted more cardholders; more cardholders attracted more stores. The internet (1990s) — the most users attracted the most services; the most services attracted the most users.

The power of the network effect — the ultimate moat

What is a network effect? A network effect arises when the value of a product or service increases for each user as more people use it. Classic example: the telephone. A telephone is worthless if you are the only user. Two telephones = a little value. 100 telephones = useful. 1 million telephones = indispensable. Every new telephone user increases the value for all existing users — a network effect. This is fundamentally different from traditional products. A chocolate bar is worth the same no matter how many others eat chocolate bars. A telephone becomes more valuable with every new user. The difference is enormous economically. Traditional companies grow linearly (each new customer = 1 unit of value). Network companies grow exponentially (each new customer increases the value for all existing ones). Metcalfe's law: the value of a network ∝ n² (the square of the number of users). 10 users = 100 units of value. 100 users = 10 000 units of value (a 100x increase for 10x users).

Valuing network companies in practice

Finding network data in the annual report Network companies often report more operational metrics than traditional companies. Three places to look. First: the CEO statement and investor presentation — network companies report user statistics here. Spotify reports MAU (Monthly Active Users), premium subscribers, ARPU; Blocket (Schibsted) reports GMV (Gross Merchandise Value), transactions, users; Klarna reports active consumers, active merchants, GMV. Second: the note on 'Segments' or 'Geographic information' — breaks down users by region. Third: the note on 'Performance metrics' or 'Operating metrics' — a dedicated section for network KPIs. Swedish examples: Spotify has very transparent reporting of user statistics (quarterly); Blocket (Schibsted-owned) reports less in detail but provides GMV and transaction volume; Klarna reports active consumers and merchants.

Network traps — illusions and erosion

Big ≠ network effect A common mistake is confusing a 'big company' with a 'network company'. A company can have 10 million customers without a network effect. ICA has 5 million loyal customers in Sweden but no network effect — if one customer leaves ICA, the value for other ICA customers does not decrease. Volvo has millions of drivers but no network effect — your Volvo does not become more valuable because more people buy Volvos. A network effect requires users to INTERACT with each other through the platform. The test: 'If one user leaves, does the value for existing users decrease?' If yes → network effect. If no → just size, not a network. Blocket: if one seller leaves, the offering for buyers shrinks → yes, a network effect.

Three perspectives on network effects

Peter Lynch: Lynch argued that network effects are the strongest moat. He mentioned Visa, American Express and H&R Block as examples. 'As more people use the network, it becomes more valuable to everyone — an unbeatable spiral.' Lynch warned, however, that networks can collapse (MySpace → Facebook).

Benjamin Graham: Graham (who wrote before the internet) saw network effects as 'economies of scale' — big-company advantages. He argued that scale gives lower costs per unit, which is a form of the moat. Graham warned that scale advantages can be eroded by technology. He preferred measurable scale advantages (factories) over digital networks.

AK1's interpretation: AKM1 weight 6%. Network effects are assessed in three ways: (1) direct networks — more users = more value (Meta, telephone networks), (2) indirect networks — two-sided platforms (Uber, Airbnb), (3) data networks — more data = a better product (Google, Netflix recommendations). AKM1 warns of the 'network illusion' — companies that appear to have a network but where switching costs are low.

Go deeper

From network recognition to moat architecture

Want to practice with worked examples, chapter by chapter? The course Network effects (V15) contains 6 chapters, the Lynch and Graham perspectives, and how AK1 uses the variable in the wave matrix. See also the complete guide to Swedish stock analysis for how all 20 variables fit together.

FAQ

What are network effects?

Network effects arise when the value of a product increases for existing users every time someone new joins — the telephone is the classic example: a single phone is useless, a million are indispensable. According to Metcalfe's law, the value of a network grows roughly as the square of the number of users, which explains why network companies can grow exponentially.

How do you test whether a company has a genuine network effect?

Använd lämna-testet: om en användare lämnar plattformen, minskar då värdet för de övriga? ICA kan ha 5 miljoner stamkunder utan nätverkseffekt — lämnar en kund påverkas inte de andra. På Blocket minskar utbudet för köparna om en säljare försvinner, alltså en äkta nätverkseffekt. Stort är inte samma sak som nätverk.

How does AKM1 weigh network effects in equity analysis?

The variable carries a 6 percent weight and distinguishes between direct networks, indirect two-sided platforms and data networks. The model warns against network illusions — companies that look like networks but whose switching costs are so low that users could leave tomorrow. Practice further in the course Network Effects (V15).

This is educational financial analysis, not investment advice.