In Part 1, I laid out nfx’s 16 network effects. Which is a pretty useable framework. And I agree with about half of them (which is pretty high for me).

In Part 2, I want to lay out how I take apart network effects. And it starts with rejecting the graphic below. This is how network effects (i.e., demand-side economies of scale) are often presented.

The right side is pretty accurate. As scale increases, you can get lower per unit costs. But other types of costs increase with scale, so you end up with a sweet spot in terms of scale. This is why most industrial businesses only get to a certain size.

But then you hear the analogy that network effects are “demand side economies of scale”. And you see the graphic on the left. Where people get excited about the exponential curve. And NFX’s curve (below) has somewhat of the same character. Its “typical marketplace” has an exponential shape.

Now you can see this in the early stages of companies, which is where venture capitalists live.

But any time anyone shows you an exponential curve, you should call bullshit. There are no exponential curves in business or reality. 

Think about the tree analogy I used in Part 1. Do trees just keep growing forever? What powerful mechanism goes on for very long? What exponential phenomenon continues very long?

The answer is none. 

Linear and especially exponential increases in value or performance are only true for a limited range. 

How to View Network Effects

Here’s the graphic I use for network effects:

I think this is a much better for visualizing network effects. It’s not sexy. There’s no exponential. Network effects are a phenomenon with a starting point and an ending point. Just like with trees.

And most of the emphasis in the graphic is on the starting point and the point at which it flatlines (additional users or activity add no more utility or value). The range where the utility and/or value service increases with more users or activity is actually not huge on the graphic – which is what usually happens in practice.

Note the carefully chosen words in the headline. For network effects, I always start with this question:

“How does the marginal user or activity increase the value and/or utility to current and potential users?”

That language is very specific.

We are looking at what happens when there is additional users or activity. The marginal value or utility added to the product or service.

• If 10 customers eat chicken, how much does the value and/or utility of the chicken increase when an 11th eats? Answer is none.
• If there are 10 users in a messenger service, how much does the value and/or utility of the messenger service increase when an 11^th joins? About 10% for each user (which you have to add up).
• If there are 100M users in a messenger service, how much does the value and/or utility of the messenger service increase when another 1M joins? Not much.

So you can see:

• Connected products can show increases. 
• There can be a lot of value in the early stage. You need a minimum scale and then it starts to appear as a phenomenon.
• The effect disappear after a certain scale.

And which users are we talking about?

• Is it users of WhatsApp?
• Is it more merchants on Shopee?
• Is it more consumers on Shopee?

Network effects are going to be different for each user group on the same platform. And they will each change in impact based on scale, time, and other factors. For each user group.

Networks Effects Can Impact Utility and/or Value

I argue that one of two things can change. The value of the service can go up. And/or the utility of the service can go up.

• Utility is easier to understand. If more people can be called by your phone line, it has greater value to the user. I can send money to more people.
• Customer value is more complicated. It can be real (I have more merchants I can chose from on Shopee). Or it can just be perceived (I think this site is better). And the value can be tangible (there are more merchants). Or it can be more emotional or psychological. Keep in mind, we are talking about value to the user, not the economic value. My favorite products always have a core utility (you need water to survive) and then a layer of emotional value (Coca-Cola tastes good).

You also need to think about current versus potential users. Adding lots of Indonesians on Facebook may not be valuable to Americans already on Facebook with most of their friends. But could be very valuable if Indonesia is earlier in market penetration.

My Checklist for Network Effects

Next take a look at the shape of the curve between the beginning and ending point.

I try to draw the shape of the curve for each user group on a platform. Is it linear? Does it take a long-time to get going? When does it flatline?

I like to know how additional value is created as a network gets bigger or more active.

• Some companies like Airbnb have a high threshold for viability. You need a lot of apartments listed in a city before it is a viable service. And really in lots of cities. After that, the platform increases in value for consumers pretty linearly as the platform adds more and more accommodations around the world.
• Companies like Uber have a very low threshold for viability. If you get 50 cars in a neighborhood, it is probably viable as a transportation service. But then the value increase flatlines pretty quickly. You don’t need 1000 cars to get to the airport. 
• Other companies go exponential at the beginning and then flatline.
• Some go up in a step function (especially on the supplier side).

Basically, it can be different for each company when you are dealing with platform business models (not just direct network effects). So, you need to take it apart.

For the network effects for each user group on a platform, you need to know:

1. What type of network, platform business model and network effect is this?
2. What is the asymptotic scale?
3. What is the minimum viable scale?
4. Within this effective range, rate (1-10) the marginal value and/or utility of each additional user or activity?
5. What is the shape of the curve?
6. What is the scale differential with rivals? Now versus later?

That’s my basic checklist. 

And really, there are other important questions for network effects, which I have listed in detail in my Moats and Marathons books.

There Are (At Least) 4 Big Effects in Network Effects

Here the 4 big effects of network effects:

1. Network effects can cause rapid increases in the real or perceived value and/or utility to customers. A product or service that increases in value TO THE USER is great. And this is usually what we are talking about with network effects.
   1. This is usually thought of as consumers. Or business customers. But it can also be for other user groups, like content creators and developers.
   2. Utility is easy to understand. Think communication networks. These are usually commodity services.
   3. But “value” can be complicated. It can be everything from a marketplace or videos to watch. Or a community or tribe to join.
   4. This doesn’t necessarily mean growth. You can have a great and improving service in a small or flat market.
2. Network effects can increase economic value. This is not the same thing as customer value (real or perceived). If a platform business model has attractive unit economics and growth potential, then you can see increasing economic value and shareholder returns with network effects. But not always. You can have a fantastic service with increasing customer but not economic value (by network effects)
3. Network effects can create a competitive advantage. This is demand side economies of scale as a moat. This is what collapses the market to a monopoly or oligopoly. However, this doesn’t necessarily mean the creation of economic value. You can dominate an unattractive business. And it doesn’t necessarily mean growth. You can dominate a stagnant business.
4. Network effects can create a barrier to entry. In digital, this is mostly by indirect network effects, which have a chicken and egg problem. That is hard for new entrants to overcome with an incumbent present. There is less of a barrier to entry with direct network effects. We also see barriers to entry in physical networks which require lots of tangible assets. Replicating a railroad is almost impossible in a developed country.

All of these can be happening with a network effect.

***

Ok. That is most of how I view network effects. In Part 3, I’ll talk about some of the softer versions of this. Like tribal, belief and psychology network effects.

Cheers, Jeff

• Can You Get Network Effects from Belief? What About from Tribes, Expertise, and Language? (3 of 3) (Tech Strategy)
• Nfx’s 16 Types of Network Effects (1 of 2) (Tech Strategy)

——–

I write, speak and consult about how to win (and not lose) in digital strategy and transformation.

I am the founder of TechMoat Consulting, a boutique consulting firm that helps retailers, brands, and technology companies exploit digital change to grow faster, innovate better and build digital moats. Get in touch here.

My book series Moats and Marathons is one-of-a-kind framework for building and measuring competitive advantages in digital businesses.

This content (articles, podcasts, website info) is not investment, legal or tax advice. The information and opinions from me and any guests may be incorrect. The numbers and information may be wrong. The views expressed may no longer be relevant or accurate. This is not investment advice. Investing is risky. Do your own research.