Network effects happen when a product or service becomes more valuable as more people use it. More videos on YouTube attract more viewers, which attracts more creators. The six main types are the industry standard, media and content, marketplaces, platforms, data, and becoming the verb in your category. They are the hardest moat for a competitor to copy.
Key takeaways
- A network effect means every new user makes the product better for every other user. That loop is the deepest moat a subscription business can dig.
- The six types: industry standard, media/content/messaging, marketplaces, platforms, data, and being the verb in a category.
- Solve the chicken-or-egg problem on the supply side first: give suppliers value before you have an audience.
- Disintermediation kills marketplaces. Add value after the match (payments, booking, guarantees) and charge for new connections, not repeat ones.
- In 2026, AI has made generic content cheap, so unique data, community and shareable outputs matter more than ever.
In 2004, a fresh Stanford graduate named Pete Flint arrived in Silicon Valley and needed somewhere to live.
He assumed someone had already built a decent online tool for searching houses. He went looking.
He found nothing.
A little later he was sitting at the back of a real estate conference in San Francisco, and it hit him. This industry was still spending billions of dollars on offline advertising. What if agents could showcase their listings online, and buyers could search them the way he'd wanted to?
That idea became Trulia, which Pete founded with Sami Inkinen in 2005. And it came with a problem.
A pair of them, actually.
How do you attract listings to a site that has no audience? And how do you attract an audience to a site that has no listings?
Pete's team started with the agents. Most brokers didn't even have a data feed to share, so Trulia built a search engine dedicated to real estate and let agents index their listings in it. The agents got traffic. More listings brought more buyers. More buyers brought more inquiries. More inquiries convinced more agents to list.
That is a network effect in action. Eventually Trulia sold to Zillow in a deal valued at $3.5 billion.
And it all started with a search engine and some legwork with local agents.
What's changed since the book
Trulia is now one of Zillow Group's brands rather than a company of its own. The network it built is the asset that survived. Zillow reported 2025 revenue of $2.6 billion, up 16%, a positive full-year net income of $23 million, and 221 million average monthly unique users in Q4 2025. It makes most of its money from the professional side (agents and landlords paying for leads), not from consumers. The full update is in the Zillow and Trulia case study.
What are network effects, in plain English?
A network effect is when a product or service becomes more valuable as more people use it.
YouTube is the easy example. It wouldn't be worth much if nobody uploaded videos. But millions of people do. More videos attract more viewers. More viewers give creators a reason to keep making videos. More videos bring more viewers, and...
You get the picture.
People often ask about direct vs indirect network effects, so let me split them out:
- Direct (same-side) network effects: every new user makes the product better for users like them. A messaging app is more useful when more of your friends are on it.
- Indirect (cross-side) network effects: new users on one side make the product better for the other side. More sellers attract more buyers, and more buyers attract more sellers. Trulia, eBay and Uber all run on this.
Most of the six types below are indirect. That matters, because indirect networks have two sides to win, and that's where the chicken-or-egg problem comes from.
Why I'd pick network effects over every other moat factor
In Chapter Five of the book I walk through the factors that protect a subscription business from copycats. If you put me on the spot and asked me to choose just one, I'd choose network effects.
Here's why.
The other moat factors make your moat wider and deeper. Network effects build a fortress behind it. Even if a competitor gets across the moat, they smash into a product that gets better every time one of your customers uses it. To beat you, they don't just need a better product. They need your users.
When I wrote the book, seven of the world's top ten companies (Apple, Microsoft, Google, Amazon, Tencent, Alibaba and Facebook) relied on network effects to some extent. You don't need to be one of them to borrow the idea.
If you want to see where network effects sit next to brand, switching costs and the rest, the MOATS framework puts them all on one page.
The six major types of network effects
I could spend a whole book on the many types of network effects. The venture firm NFX has mapped far more than six on its blog if you want to go deeper. But to keep things simple, here are the six I think are the most powerful and the most likely to apply to the widest range of businesses.
Network effect #1: The industry standard
The industry standard is inescapable. It's embedded in every product and workflow in its niche.
The Real Estate Institute of New South Wales (REINSW) is a great example. Years ago, every one of the thousands of real estate agencies in Sydney had its own forms and agreements. There was no standard.
Then REINSW came along.
It became the advocacy group for agents and created standardized forms and documents that many agents still use. It isn't famous outside Australia. It doesn't need to be. In its niche it's the standard, which makes it nearly impossible for a similar organization to replace.
There's one ingredient that makes this network effect take off: expertise.
REINSW was founded by local real estate experts who build tools for other experts. Using those tools properly requires real estate knowledge, and that need for expertise embeds them even deeper.
So here's your question: is there anything your business does that could become the standard way your industry does something? A checklist, a template, a certification, a set of definitions? Give people something so good they can't picture using anything else, back it with expertise, and you're digging the moat deeper.
Network effect #2: Media, content, messaging
Attention is valuable.
Television proved it for decades. Shows with higher ratings attract advertisers who pay more to be in them. The network earns more, the writers and actors earn more, and it all comes from attention.
This network effect has two layers:
- The content marketplace. Content is the supply, attention is the demand. More good content draws more attention.
- Messaging or advertising. Once you have the attention, advertisers come calling. Now the supply flips: for advertisers, the supply is the audience, not the content.
Think of it as a three-sided marketplace: creators, audience, advertisers.
The Urban Developer in Australia shows how a small team can use it. When I wrote the book it was a team of twelve curating the most important news from Australia's property development sector. Those stories attracted attention. The attention attracted brands that wanted to reach the readers. Then it added a subscription, TUD+, with exclusive content and webinars. That's a second revenue stream, and a recurring one.
Netflix runs the subscription version at enormous scale, and since the book came out it has added the advertising layer too. At its May 2026 upfront, Netflix said its ad-supported plan had more than 250 million monthly active viewers. Same content, two ways to get paid.
Network effect #3: Marketplaces
Marketplaces bring buyers and sellers together. More sellers attract more buyers. More buyers attract more sellers, because sellers can sniff out the profit.
The beauty of marketplaces is that you can start small by focusing on a local micro-niche. Trulia didn't travel the length of the United States signing up agents. It stuck to Silicon Valley, got listings indexed, attracted buyers and built from there.
You could do the same thing for almost any service. Picture an "Uber for physiotherapists" in your area. Approach physios who want more work, build a simple first version of the app, run a few ads to reach patients, then add a subscription for discounted monthly visits.
Realestate.com.au shows how defensible a marketplace becomes once it wins. When I wrote the book it had been Australia's leading property portal for more than two decades and had generated $1 billion of revenue in 2021. It has kept compounding. For the year to June 2026, REA Group reported revenue of A$1,793 million, up 7%, with an Australian operating EBITDA margin of 66%. There's a graveyard of businesses that tried to copy it. The only way to unseat an incumbent marketplace is to build a better experience for buyers and sellers at the same time.
Marketplaces have two weak spots.
Multi-tenanting. Nothing stops an eBay seller also listing on Amazon and Etsy. That dilutes the network effect. eBay's answer was to give sellers more value through its eBay Stores membership: sellers pay a subscription and get tools and benefits that make eBay their main shop. Win-win.
Not getting paid. This one is a cautionary tale.
Ethan Stock built Zvents into the largest events site of its kind, with 14 million monthly users. Event organizers listed their events, and new users pulled in more organizers.
The problem? Ethan wasn't getting paid.
His plan was to grow until he could ask organizers to pay for listings. But visitors expected a complete list of every local event. When he asked organizers to pay, they said, "What for?" He couldn't remove the free listings without losing the visitors, and without the visitors he had nothing to charge for. He called it "the expectation of completeness."
The lesson: a marketplace network effect is only worth having if you've found a way to get paid. Facilitate the transaction (Uber, Upwork), take a slice of the sale (eBay) or charge per inquiry (Bark). Work out the money early.
Network effect #4: Platforms
Platforms are like marketplaces with one key difference: what's on offer is only available on that platform.
Video game consoles are the classic example. Nintendo, Sony and Microsoft each sell hardware that costs hundreds of dollars, so most gamers only buy one. Each draws people in with exclusive games you can't play anywhere else, ties developers in with exclusivity deals, and adds subscriptions such as Game Pass. Communities form around each platform.
Software companies do it too. Most CRMs have an app store of third-party add-ons. Autodesk has its developer network.
And small businesses have run platforms for years without calling them that. A wedding planner has preferred dressmakers, venues and photographers. A GP has a team of specialists. When I ran my marketing agency, I was the platform for my clients' marketing. If a client had plenty of traffic but a lousy website, I'd plug in my preferred web designer and manage the job.
People can still multi-tenant on platforms, but it happens less than in marketplaces.
Network effect #5: Data
A data network effect occurs when a product gets more valuable as it gathers more data, and using the product generates more data.
Waze looks like a simple directions app. Underneath, it collects live traffic conditions from every driver. Nobody cares about yesterday's traffic. They care about right now. More drivers mean better data, which means more drivers.
Yelp's reviews work the same way. So does IMDb. When I wrote the book, IMDb had more than eight million titles and 83 million registered users contributing data. Could a rival copy it? Maybe, with bottomless pockets and endless time. Why would anyone bother?
Data on its own isn't valuable. Data that users want, and contribute to themselves, is.
Network effect #6: Being the verb in a category
Americans don't ask for a tissue. They ask for a Kleenex. You don't photocopy something, you Xerox it. You Google it. You Photoshop it.
This network effect happens when a product is so widely adopted that replacing it becomes very hard. It helps if your name can work as a verb. ("ABC Quality Photocopiers" was never going to become one.)
One update here. DocuSign became the verb for e-signature, and it's a key story in land and expand. But by fiscal 2026 its growth had slowed to 8% as e-signature commoditized, and it had to build a second product to grow again. Being the verb protects you from copycats. It doesn't protect you from your category becoming cheap.
The six network effects at a glance
| Type | How the loop works | Book example | A small-business version | Main risk |
|---|---|---|---|---|
| Industry standard | Everyone uses it because everyone else does | REINSW forms | Your checklist or template becomes the default in your niche | Being replaced by a free standard |
| Media, content, messaging | Content draws attention; attention draws advertisers or subscribers | The Urban Developer, Netflix | A niche newsletter with sponsors and a paid tier | Content becomes cheap (see AI below) |
| Marketplaces | More sellers draw buyers; more buyers draw sellers | Trulia, eBay, realestate.com.au | A local directory of tradies or physios | Disintermediation, multi-tenanting, not getting paid |
| Platforms | Exclusive add-ons draw users; users draw add-on makers | Xbox, PlayStation, CRM app stores | Your preferred-partner network | Partners going direct |
| Data | Usage creates data; data improves the product | Waze, Yelp, IMDb | Benchmarks built from your clients' (anonymized) results | Privacy rules, and data nobody wants |
| Being the verb | Habit and language lock in the default | Google, Xerox, DocuSign | Being "the" provider people name in your town | Commoditization of the category |
The chicken or the egg? How to start a network with no audience
When you look at Trulia, you see the classic chicken-or-egg problem. No listings, no audience. No audience, no listings.
The good news is it isn't an unanswerable question. One side is usually harder to win than the other, and that tells you where to start. Most of the time, it's the supply side.
Here are the moves I've seen work:
- Give suppliers value before you have an audience. Trulia gave agents traffic from a search engine built for them.
- Subsidize the hard side. Uber has at times paid drivers to join, because more drivers attract more riders.
- Pay both sides to join. PayPal used to give you $5 for referring a new user, and another $5 to the new user. It doesn't need to now. Its network effects (and its link to eBay) made it an industry standard.
- Pre-fill the supply side with public information. Say you're building a marketplace that connects accountants with clients. You could pre-fill each accountant's listing from their website, send them the inquiries that come in as free leads, and then offer a paid subscription for more leads. That's almost exactly what urban.com.au did with new apartments and homes. It created free listings, sent inquiries to developers as leads, and sold premium packages once developers saw the value.
One clever strategy a hairdresser marketplace used to grow rapidly
Another clever play is to build some software.
StyleSeat did that in hairdressing. It gave stylists free software to run their businesses (scheduling and touch-free payments) and asked only for a card processing fee.
The value attracted the hairdressers. The hairdressers attracted the clients. As the clients grew, more hairdressers came on board.
The egg hatched. The chicken was born.
What StyleSeat did once the network existed
StyleSeat didn't stay free. By 2026 its own pricing pages show a $35 monthly subscription for stylists (with a promotional rate and a free trial), card processing at about 2.5% plus 30 cents in person, and a "New Client Connection" fee of 30% of the first appointment, capped at $50. That fee is charged once, only for brand-new clients the marketplace sends, and never again for the same client. StyleSeat says more than 250 million appointments have been booked through it. Read the StyleSeat case study for the full breakdown.
That last detail is the smartest pricing decision in this chapter. It brings us to the biggest danger for any network.
Which network effect could you build?
Score your business against all six network effects in about five minutes with the free Network Effects Scorecard.
How do you stop people going around your network?
The fancy word is disintermediation.
A supplier and a customer meet through your platform. That's what you want. Then they get along, swap numbers and work together directly.
They cut out the middleman. You.
In a subscription business that's deadly, because you rely on repeat usage. HomeJoy, a home cleaning start-up, raised $40 million and still went under. Too many customers and cleaners met through the app and then dealt directly, and it shut down on 31 July 2015 while also fighting worker-classification lawsuits. TakeLessons, which connects students with teachers, raised $19 million and battled the same problem for years. The HomeJoy case study explains why so many marketplaces now add software for this reason.
So how do you stop it?
Give them tools they only get by staying. Upwork offers payment protection, so freelancers know they'll be paid, and time tracking, so clients know the hours are real. You don't get either off-platform. Does that stop every freelancer leaving? Of course not. But it makes leaving risky for both sides.
Hide the contact details. Uber and Lyft mask drivers' phone numbers. A customer can't call a driver outside the app if they don't have the number.
Make loyalty cheaper. Tutor.com started its take at 50% of each transaction and lowered it with every repeat booking, so there was less reason to go around it.
Get the information another way. Then there's the smart (albeit sneaky) approach. Hired.com charged employers $10,000 for a hire, so some tried to contact candidates directly. Hired offered job seekers a $150 bottle of champagne when they landed a job. Most happily reported in, Hired asked who'd hired them, and if that employer was a customer, a $10,000 invoice followed.
Charge for new connections only. This is the modern answer, and StyleSeat's capped New Client Connection fee is the template. If you only charge the first time you introduce someone, the supplier has no reason to sneak repeat business off the platform.
Steal this: the leakage checklist
If you introduce customers to providers, ask yourself five questions. Do payments run through you? Does booking or rebooking run through you? Do you offer a guarantee or protection that only applies on-platform? Is your fee low (or zero) on repeat jobs? Do reviews and reputation live on your platform, so providers lose them if they leave? Every "no" is a hole in the bucket.
Gamification: how to make your product "multiplayer"
Gamification isn't strictly a network effect, but businesses with network effects use it all the time.
The idea comes from video games, where players earn achievements and trophies for things that keep them playing for hours, while paying a subscription or buying bonus content.
In a subscription business, gamification works best when people can see what others are doing. NFX calls this "creating a multiplayer product." YouTube's comments and view counts show you other users' impact. Amazon's reviews help the next buyer decide. Xbox and PlayStation let you see your friends' achievements, and if they've got one you haven't, you keep playing.
Engagement is the goal. Deeper engagement means more time in the product, and more interaction between users, which strengthens the network effect.
Here are two quick tips.
Tip 1: Reward users at each stage of their journey
Give people a little dopamine hit for doing what you want them to do.
In the book I described how Duolingo granted "Gold" status when you completed a lesson, which decayed if you didn't revisit it. That rewarded completion and encouraged repetition, which is how you learn a language.
Since then, gamification has become Duolingo's growth engine. In Q2 2026 it reported 58.7 million daily active users (up 23%) and 12.7 million paid subscribers (up 17%). Its biggest lever that quarter was a game mechanic: a June 2026 "Streak Revival" event in which 15.4 million learners revived lost streaks. The Duolingo case study shows how a streak became a retention machine.
Tip 2: Add a social element
Scoreboards work if your product has challenges you can rank people on. We did this in our e-learning academy. Every course you viewed or exam you completed earned points, with more for harder courses. Tons of students competed to be ranked number one in the world.
Communities work too. Clash of Clans lets players form clans, and friendships are hard to walk away from.
Sharing works best of all, because it turns engagement into growth. Fitness apps let users share runs and personal bests. Strava has built a business on this: clubs, kudos and leaderboards on route "segments," with a free app as the front door and a paid subscription ($11.99 a month or $79.99 a year in the US) for training tools. It cited more than 150 million registered users in April 2025. The Strava case study covers how it converts them.
When network effects are the wrong goal
I believe network effects are the most powerful strategic factor in the book. But I'll be honest about when to leave them alone.
- When your customers never interact. If one client's use of your service does nothing for another client, don't force a marketplace or community onto it. Build switching costs, brand and usage instead.
- When you can't work out how to get paid. Remember Zvents. Fourteen million users and no revenue model is a hobby, not a business.
- When the network gets worse as it grows. That's a negative network effect: congestion, spam, or so many sellers that each one struggles to win work. Marketplaces need curation, quality rules and sometimes caps.
- When you can only win one side with subsidies forever. Subsidies are a starter motor, not an engine.
And yes, network effects are a barrier to entry, but not a permanent one. The only way to dislodge a network is a dramatically better experience for both sides at once. That's rare, but AI is making it less rare in some categories.
2026 update: network effects in the AI era
A lot has happened since the book came out in 2022. Here's what I'd add to the chapter today.
1. AI has weakened content networks built on generic answers. Chegg built a huge library of homework answers and the Google traffic that came with it. Then free AI answers arrived. By Q2 2026 its revenue was down 51% year on year. The Chegg case study is the warning: if a model can produce your content for free, your content network effect is gone. Community, accountability and unique data are what survive.
2. Data network effects are now the prize. The most valuable AI-era networks collect data nobody else has. Arena, the AI model leaderboard, built its rankings on more than 10 million user evaluations and sells the resulting data to AI labs. TechCrunch reported in June 2026 that it had become a $100 million business (Arena itself says its revenue is consumption-based rather than recurring). Ask yourself: what does every one of my customers generate that, pooled and anonymized, would make the service better for all of them?
3. Shareable outputs are the new referral loop. Several 2026 fast growers spread because what customers make carries the brand. Meshy's free 3D models carry an attribution license, so every published model credits Meshy, and paying users get private ownership instead. Emergent lets users share and clone the apps they build.
4. Platforms now open up to AI agents. Google's Looker added a Model Context Protocol (MCP) server so third-party AI agents can query its trusted data definitions. That's the platform network effect from the book, updated: the more agents and tools plug into you, the harder you are to remove. My AI moats guide goes deeper.
5. Marketplaces monetize the pro side. Zillow charges agents and landlords, not home buyers. StyleSeat charges stylists for new clients, not for repeat ones. Keep the side you need to attract cheap or free, and charge the side that profits from access.
What to do next
Here's how I'd put this chapter to work this week:
- Score yourself. Run your business through the six types using the Network Effects Scorecard. Pick the one with the most potential, not all six.
- Shrink the network. Choose the smallest geography or niche where a network would already be useful, as Trulia did with Silicon Valley.
- Win the supply side first. List three things you could give suppliers (or members, or partners) before you have an audience: software, leads, traffic, a standard template.
- Decide how you'll get paid. Subscription for pros, a capped finder's fee, a slice of each transaction. Write it down before you launch.
- Plug the leaks. Run the five-question leakage checklist above.
- Make it multiplayer. Add one visible social element: a leaderboard, a member count, a streak, a share button.
Network effects are also a great companion to referral programs, which are often the cheapest way to get the first loop turning.
Read the whole book free
The Subscription Playbook as a free ebook and audiobook. Australian business owners can claim a free hardcover, shipped free.
Claim the free hardcover Get the ebookAction exercise: How could you use network effects in your business? Write down the type, the first side you'll win, and how you'll get paid. Then take it to the next chapter on land and expand, where we lift revenue without hiring a huge sales team.
Frequently asked questions
Can a small service business have network effects?
Yes, though usually a local or niche version. A wedding planner with a trusted supplier list, an accountant who runs a client community, or a gym whose members train together all get more valuable as more people join. You don't need millions of users. You need each new member to make the experience better for the others, and a reason for them to stay inside your network.
What is the difference between network effects and going viral?
Virality is about how fast a product spreads: each user brings in more users. A network effect is about value: each user makes the product better for everyone else. They often travel together, but not always. A funny video can go viral with no network effect, and an industry-standard form can have a strong network effect without ever going viral.
How many users do you need before a network effect kicks in?
There's no universal number. It depends on how small you can make the network. Trulia started with one region, Silicon Valley, so it only needed enough listings to be useful there. Shrink the geography or the niche until a few hundred active users, or a few dozen suppliers, already deliver real value. Then expand area by area.
What is multi-tenanting?
Multi-tenanting is when suppliers or customers use several competing networks at once, such as a seller listing the same product on eBay, Amazon and Etsy. It dilutes your network effect because you no longer have exclusive supply. The fix is to give suppliers so much value on your platform (tools, buyers, reputation) that listing elsewhere feels like extra work for little return.
Do network effects and subscriptions work together?
Very well. A network effect makes the product more valuable each month, which lowers churn, and the subscription turns that value into predictable revenue. Common patterns are a free network with paid pro tools (Strava), a paid supplier membership (eBay Stores) or a subscription for professionals who want access to the audience (Zillow's agents).
Get both books free
The Subscription Playbook (ebook + audiobook) and Feed a Starving Crowd, instantly. In Australia? Claim the free hardcover.
Robert Coorey, MBA
Robert is the author of The Subscription Playbook and Feed a Starving Crowd and a Director at BPAG Accountants in Sydney. More about Robert.
Sources
- NFX, The NFX of Trulia: building a $3.5 billion marketplace
- Zillow Group, Q4 and full-year 2025 results (10 Feb 2026)
- REA Group, FY26 results announcement to the ASX (6 Aug 2026)
- StyleSeat, StyleSeat Review 2026: What Professionals Actually Pay
- Vox/Recode, Homejoy shuts down after battling worker classification lawsuits (17 Jul 2015)
- Duolingo, Q2 2026 shareholder letter (Aug 2026)
- Strava, Strava to acquire Runna (17 Apr 2025)
- TechCrunch, Arena, the AI leaderboard everyone uses, is now a $100M business (29 Jun 2026)
Facts and prices checked 27 October 2026. Spotted something out of date? Tell us.