A growing user base is evidence of growth. A network effect requires a further link: participation by additional users must improve the value experienced by other users.
For an investor, the distinction matters because growth can be purchased while a durable advantage may remain unproven. The diligence task is to identify the mechanism and test whether it becomes stronger as the relevant network expands.
Name who benefits from whom
Write the claim as a sentence: “When more of X participate, Y receives a better Z.” A local marketplace might claim that more available suppliers reduce buyers' waiting times. A professional community might claim that more relevant participants improve the quality of answers.
In a two-sided market, the sides can benefit differently and may require different pricing. Rochet and Tirole's work models the importance of attracting participation on both sides and choosing the price structure between them. It does not imply that every platform becomes defensible simply by adding users. Rochet and Tirole, working paper.
Once the mechanism is explicit, choose an outcome that measures user benefit: successful matches, time to completion, useful responses or relevant product availability. Total registrations is usually too remote from that claim.
Find the boundary of the network
An additional supplier in Manchester may do little for a buyer in Bristol. Another accountant may not improve a marketplace for electricians. Activity outside the relevant location, category or time window can inflate the headline network without improving the experience.
Analyse the smallest unit in which participants interact. For a local service marketplace, that might be a city, service category and day. For enterprise software, it might be the team or organisation using the product together.
Then examine whether growth within that unit changes outcomes. A company expanding into many thin markets can look larger while delivering no stronger experience in any one of them.
Compare outcomes, then challenge the explanation
Suppose an illustrative marketplace reports the following:
| Measure | Less dense market | Denser market |
|---|---|---|
| Active relevant suppliers | 20 | 60 |
| Requests successfully filled | 55% | 80% |
| Customers returning within 90 days | 25% | 40% |
The pattern is consistent with a beneficial network effect. It does not establish causation. The denser market might have older customers, higher marketing subsidies, better service management or more attractive demand.
Request comparable cohorts and account for differences in pricing, tenure and acquisition channel. Where feasible, use a carefully designed experiment or phased rollout to test a specific mechanism. Record remaining uncertainty rather than presenting an observational comparison as proof.
Remove the subsidy explanation
A discounted service can attract buyers even if the network has not improved. Payments to suppliers can temporarily create availability that disappears when incentives end.
Track the user outcome alongside the cost of producing it. Does the fill rate remain attractive when incentives reduce? Do suppliers still earn enough to participate? Is repeat use sustained at prices that support the business?
This does not mean all subsidies are wasteful. They may help establish an initially thin market. The investment case should distinguish the cost of building a useful network from a continuing need to pay both sides to stay.
Look for strength that survives concentration and competition
Calculate how much activity depends on the largest participants. If three suppliers fulfil most requests, losing one may matter more than adding hundreds of inactive accounts.
Also investigate whether users participate across competing networks. A supplier can value access to buyers while listing on several services. The network may improve the product without giving the company much pricing power.
Watch for negative effects too: congestion, low-quality listings, spam or competition among suppliers can offset the benefit of additional participation. More activity is useful only while its net effect on the relevant user remains positive.
State the investment conclusion narrowly
A strong finding sounds like this: “Within established city categories, higher relevant supplier availability is associated with faster fulfilment and repeat use; the pattern persists after the introductory discount ends, although causal evidence remains incomplete.”
That is more useful than “network effects: yes”. It identifies where the advantage operates, what supports it and what still needs testing. Our separate network-effect grading framework helps classify the mechanism before assigning confidence to the claim.
FAQ
What is meant by network effects?
A network effect exists when participation by additional users changes the value experienced by others. For a positive effect, specify who benefits and how. More registrations alone do not demonstrate a stronger product; the relevant participants must improve an outcome that users value.
What is the difference between direct and indirect network effects?
Direct effects operate among users on the same side, such as colleagues joining a shared communication tool. Indirect effects work across groups, such as relevant suppliers improving a marketplace for buyers. The benefit can differ by direction, location and category, so test each mechanism separately.
Is virality a network effect?
Virality concerns how a product spreads. A network effect concerns what participation does to its value. A product can acquire users through referrals without becoming more useful as they join. After accounting for acquisition activity, ask whether existing users receive a better experience.
What is a negative network effect?
Additional participation can reduce value through congestion, spam or excessive competition on one side. A shrinking network can also lose usefulness and encourage further departures. Examine the net effect on each group instead of assuming that more activity is always better.
