Network effects
Metcalfe's law · demand-side economies of scale
When a product gets more valuable to each user as more people use it. The growth reinforces itself, which is why these businesses are winner-take-most and why the early period is so much harder than the later one.
In practice
Most businesses do not have them and claim to. A service that gets better as you scale is experiencing economies of scale; a network effect requires that users make it more valuable to each other.
The common mistake
Confusing them with scale economies or with brand. Scale lowers your costs, brand lowers acquisition cost, and neither makes the product more valuable to an existing user when someone else signs up.
A telephone has no value at all if nobody else has one. Two telephones are worth something. A million make each one nearly indispensable.
That is a network effect: the value of the product to each user rises as the number of users rises. It is a property of the product rather than of the company, and it is rarer than the phrase's popularity suggests.
The test
Ask one question: when a new customer signs up, does the product become more valuable to the existing customers?
If yes, that is a network effect. If the answer is that the company gets more efficient, or better known, or able to invest more, those are real advantages and they are economies of scale and brand. The distinction matters because the growth dynamics are entirely different and only one of them is self-reinforcing on the demand side.
Most businesses that claim network effects have scale economies. The confusion is expensive when it drives strategy, because chasing growth at any cost is rational under a genuine network effect and ruinous without one.
The kinds
Direct. More users of the same type make the product better for each other. Telephones, messaging, any marketplace where participants transact with each other directly.
Indirect. More users on one side attract more on another, which makes the first side better off. Buyers attract sellers, which attracts buyers. Most marketplaces work this way and it is why they are hard to start.
Data. Usage improves the product for everyone. Each search makes search better. This is the weakest claim in the family and is often really a scale economy wearing different clothes — the question is whether the improvement is visible to a user rather than merely to the operator.
Local. The effect applies only within a group or geography. A booking platform is worth nothing in a city with no restaurants on it however many it has elsewhere. This is what lets a smaller competitor win a market against a larger one, and it is the most commercially relevant of the four for anyone not building a global platform.
The cold start, which is the whole difficulty
A network effect works against you at the beginning with the same force it works for you later. Nobody joins because nobody has joined.
The standard ways through are worth knowing because each is a real strategy rather than a trick.
Be useful with one user. Give the product standalone value before any network exists, then let the network make it better. Anything a single person would use alone can grow into a network; a product that requires ten people to be worth anything must find ten simultaneously.
Start narrow. Own a small market completely — one city, one profession, one campus — where local density is achievable. A dense small network beats a sparse large one, always.
Subsidize the harder side. In a two-sided market one side is scarcer. Pay for it, in money or attention, until the other side makes it self-sustaining.
For a service business
Most service businesses have none of this, and knowing so is useful rather than discouraging, because it means the strategy that works is a different one.
What service businesses actually have is switching costs and referral loops. A referral loop resembles a network effect and is not one: more clients produce more referrals, which grows the business, but an existing client is not better off because you took on another. That is a growth mechanism and not a moat, and the difference shows when a competitor appears.
The genuine version in services is narrow and real: a firm with many clients in one industry accumulates comparative knowledge that makes it more valuable to each of them. That is a true network effect, it is why specialists outperform generalists, and it is available to anyone willing to give up the rest of the market.
Concept web
Open the full webQuestions
What are network effects?
When a product becomes more valuable to each user as more people use it. The test is whether a new signup makes the product better for existing customers, rather than merely making the company more efficient or better known.
What is the difference between network effects and economies of scale?
Scale economies lower your costs as you grow. Network effects raise the value of the product to users as others join. Only the second is self-reinforcing on the demand side, and confusing them leads to growth spending that cannot pay off.
Do service businesses have network effects?
Rarely. Referral loops resemble them but do not make existing clients better off. The genuine version is specialization: a firm with many clients in one industry accumulates comparative knowledge that makes it more valuable to each of them.