Network Effects as Felt Experience
A network feels worthless until the specific people you care about are on it, then it feels essential.
network effect: The product gets more valuable to each user as more people use it. mean a product's value rises with the number of users, but people do not experience an aggregate count. They experience whether their own contacts are present. The felt curve is therefore flat and frustrating for a long stretch, then steepens abruptly once a personal threshold is crossed.
Felt value against people you actually know
How it shows up in software
Messaging apps, marketplaces, social products, and collaboration tools all live on this curve. The dangerous stretch is early, where honest usage feels empty and the team is tempted to fake activity. Products that survive it usually give standalone value before the network exists, or they seed a small dense group rather than spreading thinly.
Using it well
- Make the product useful for one person alone, so the first session has value before anyone else arrives.
- Seed a narrow, dense group, such as a single campus, team, or city, rather than spreading the same effort across many.
- Measure density of the user's own graph, such as how many of their contacts are active, rather than total accounts.
- Show the user who they know is already here, and let them import or invite the specific people who make it work.
Where it turns manipulative
- Faking activity in an empty network, with bot accounts or staff posing as users, is fraud, and several dating and social products have been sanctioned for it.
- Uploading a user's contact list without informed consent to manufacture a network, then messaging those contacts as though the user invited them, misuses both the data and the relationship.
- Locking exported data so that leaving means abandoning the network raises switching costs by hostage-taking rather than by value.
Where you have seen it
WhatsApp
Builds the contact list from the phone address book, so a new user sees which of their existing contacts are reachable on day one.
Figma
Works as a single-player design tool before any collaborator joins, which carries a user through the empty stretch.
Facebook
Launched campus by campus, keeping each new network dense enough to be useful before opening the next.
What the research says
- Katz and Shapiro, 1985Well evidenced
Formalised direct and indirect network externalities, showing that expectations about future adoption shape present adoption and can produce multiple equilibria.
Theoretical economics. It explains the dynamics rather than measuring any particular product.
- Tucker, 2008Mixed evidence
Studied a video messaging rollout inside a large bank and found adoption was driven by the presence of a small number of influential contacts rather than by total adopters.
One firm, one technology, with identification resting on the internal hierarchy. It supports the local-network claim without settling its generality.
- Ugander, Backstrom, Marlow and Kleinberg, 2012Well evidenced
Facebook recruitment data showed probability of joining depended on the number of distinct connected components among a person's contacts on the service, not on the raw count of contacts.
Very large observational dataset. Correlational, but the structural finding is hard to explain away.
Grades are a judgement about the evidence, not about how useful the idea is. Plenty of contested effects are still worth knowing, as long as you do not cite them as settled.