Network effects strategy

Network effects strategy is the plan for building a product that grows more useful to each user as more people join it. It is not virality, and it is not a referral scheme; both move signups rather than the value each user gets.

The distinction decides where your money goes — a referral budget buys one more user, while a network strategy buys the thing that makes the next user worth more to everyone already there.

Most networks fail at the start, not at scale. An empty network has nothing to offer, so the first job is picking a group small and dense enough that the product is useful before anybody else arrives.

Key takeaways

• A network effect means each new user raises the value of the product for existing users.
• Direct effects come from your own side; indirect effects come from the other side.
• Density inside one niche beats thin coverage across many, every time.
• Network effects are defensible, not permanent, and low switching costs erode them.

How it works

A network effects strategy picks the effect you are building, then designs for density before reach. You choose a starting group, make the product useful at tiny scale, and only then open the gates to everyone else.

The clearest legal description comes from the Supreme Court’s 2018 opinion in Ohio v. American Express. It defines a two sided platform as one serving two different groups who depend on the platform to intermediate between them.

Such platforms, the opinion says, show indirect network effects: the value to one group depends on how many of another participate. Price changes on one side therefore ripple straight to the other.

Card networks are the textbook case. The opinion calls them a special type of two sided platform known as a transaction platform, which cannot make a sale to one side without making one to the other.

The scale of networked activity is measured in its own right. The United States Bureau of Economic Analysis publishes digital economy statistics covering 2005 to 2021, built to improve measures of high technology goods, services and data as an asset.

Pricing is where the strategy becomes concrete. One side usually subsidises the other, and getting that direction wrong is fatal, because the subsidised side is the one that will not turn up unprompted.

Liquidity is the honest test. Can a typical request on your network find a match today, and does the answer still hold on a quiet Tuesday in February as well as in peak season?

Marketplaces live on the indirect kind — a BPO marketplace is worth little to a buyer with ten providers listed and a great deal with a thousand, which is why listings come before advertising.

Community formats build the direct kind cheaply. Community led growth treats members as the supply side, and community forums turn answered questions into an asset that keeps paying long after the thread goes quiet.

Watch for the multihoming problem early — when users keep an account with you and a rival at the same time, the effect still exists but stops being a moat, and price becomes the deciding factor again.

Measure density, not downloads. A digital adoption rate that is high inside one team or one city beats a low one spread thinly, and sustained density is what later turns into hypergrowth.

Examples

Three shapes recur across markets: a payments network priced on both sides, a marketplace that seeds one side first, and a product whose data improves with every session. Each one earns its position in a different way.

Payment cards show the two sided balance in public. Issuers court cardholders with rewards while merchants pay the acceptance fee, and pricing on each side is set with the other side firmly in mind.

Marketplaces usually buy one side and let the other follow. Outsourcing directories and freelance platforms both built provider supply first, because a buyer who arrives at an empty listing does not come back.

Workplace software rides the social kind. Teams adopt the tool their contractors and clients already use, which is why pilots spread sideways through a company rather than down from a procurement decision.

Data effects are quieter and harder to copy — mapping and fraud scoring services improve with every trip or transaction they see, so a late entrant launches a product that is simply worse on day one.

Related terms

• Transformational growth: the step change a working network effect can produce once density holds.
• Growth marketer: runs acquisition experiments, including plenty that carry no network effect at all.
• BPO marketplace: a two sided outsourcing network where buyers and providers must both show up.
• Community led growth: using members rather than paid channels as the engine of supply.
• Hypergrowth: the pace that can follow a network effect, and sometimes arrives without one.

FAQ

What is the difference between network effects and virality?

Virality is how fast new users arrive. A network effect is whether each arrival makes the product better for everyone else, and a product can be viral with none at all.

What is an indirect network effect?

It is value that comes from the other side of a platform. The Supreme Court described it as the value to one group depending on how many members of another group take part.

How do you start a network with no users?

Pick the smallest group that is useful to itself, and serve it completely before widening out. Density inside one niche beats thin coverage everywhere, because a half empty network reads as abandoned.

Do network effects protect a business forever?

No, they are an advantage to defend rather than a settled position. Users signing up to two rival services, low switching costs and a competitor with a better starting niche all erode them.

Can network building be outsourced?

Parts of it travel well, including moderation, onboarding and support for the supply side. Keep the pricing and policy choices with your own team.

See how a two sided outsourcing network works in practice at Outsource Accelerator.