The new components of the network effect

Network effects are behind the most valuable companies of our time. The sub-components of the network effect help us to understand why.

In his new book Cold Start Problem, Andrew Chen breaks down the classical network effect into three distinguished components. These components help us to build and grow businesses with network effects and give us a better view of the value that they generate.

Acquisition network effect → lower customer acquisition costs

Engagement network effect → increased interaction

Economic network effect → increased monetization

I’ll go through each of these with practical examples for optimization.

---

I write about growth and startups. Subscribe and tag along for further posts:

Why bother with the components of network effects?

Instead of running tests for increasing network effects, you are usually trying to improve some specific metric regarding customer activation, referrals, or monetization.

You usually focus on optimizing one metric by running different tests that you think that can have a positive impact on that metric.

Components of the network effects help us to move from the abstract concept into more concrete actions to grow and strengthen the business. Having of more granular understanding of the network effect can enable you to grow your business much faster and create more value.

Acquisition network effect

Product growing virally because of the strong network effect is probably something that comes to mind when thinking about network effects. It might be someone sharing something that is from TikTok or inviting you to join the Clubhouse.

All businesses with network effects usually have some sort of acquisition network effect. This natural virality comes from the fact that having more people in the network brings more value to the users, so referring the service to others is in your self-interest. That sharing might be prompted, incentivized by rewards, or just native to how you use the service.

Dropbox, Slack, and Pokemon GO are all products that have great acquisition network effects.

Referral programs are one way to optimize the acquisition network effect. Dropbox is famous for growing its userbase with referral incentives that gave you more storage space if you invited your friends. Many have tried copying that model unsuccessfully, but it doesn’t mean that you couldn’t make it work. For some companies, referral programs are a core acquisition channel.

Some products like Slack or other messaging products are natively viral. In order for you to use their product at all, you need to invite other people in that you can start talking to. Making the invitation easy and suggesting other people from your connections that you can invite to the service can increase this effect.

Having shareable moments in your service can also boost up your acquisition network effect. Pokemon GO had great virality from people sharing screenshots of the Pokemons in the real world using augmented reality. Being shareable can really increase the acquisition effect.

Engagement network effect

The second component of the network effect is the engagement network effect. It’s the effect that makes your customers come back after and keep using your service. Thanks to the engagement network effect, your users that have stopped using your service will come back and the ones that are currently using it will become even more active users.

The engagement network effect happens, when the different nodes in the network increase each other’s engagement. In a social network, the core engagement loop might be posting a picture to the website, which then leads others to comment and like the picture. That liking and commenting will bring you back to the service and make you use the product more.

Optimizing that core engagement loop with different notifications or making the posting easier can be a great source for growth for network effect business.

Another way to improve the engagement network effect would be to add new use-cases that will make the users more engaged. With a social network product, this could mean adding news functionality in terms of groups or events that would increase the engagement of the users.

Bringing back users that are not actively using your service is also an important piece of the engagement network effect that you can optimize for. You are more likely to click email about having a new friend request than a generic email to come back to the social network.

Economic network effect

The economic network effect means that the business can monetize better and reduce its costs when the network grows.

For example, a food delivery company’s operating costs per order get lower when the network is larger with more customers, more people doing deliveries and more restaurants. With that data and more possible nodes for each side of the product, there are more chances for optimizing and cost savings.

The better monetization due to the economic network effect with an example B2B SaaS can mean that when your product is fully in use in a large company, you have a better chance for selling a higher-priced tier with enterprise functionality.

An alternative point of view

I feel like these components are definitely giving new tools to better understand the network effects, but I don’t think these are yet the final form.

The lending company example, where the company was collecting creditworthy customers, would be classified as a data network effect if we follow the NFX’s categorization. Also, the acquisition network effects would not be categorized as network effects, but viral effects, if we would look at how NFX’s and James Currier have defined the network effects and viral effects.

So you can look at these things from multiple angles but all in all, I think the book was a great contribution to the whole scene and definitely recommend it to anyone who works with businesses with network effects.

I’m going to write more about the subjects of the book like market saturation and the more qualitative side of growing a network so remember to subscribe in order to tag along: