“Ultimately, we’re all social beings, and without one another to rely on, life would be not only intolerable but meaningless. Yet our mutual dependence has unexpected consequences, one of which is that if people do not make decisions independently — if even in part they like things because other people like them — then predicting hits is not only difficult but actually impossible, no matter how much you know about individual tastes. The reason is that when people tend to like what other people like, differences in popularity are subject to what is called “cumulative advantage,” or the “rich get richer” effect. This means that if one object happens to be slightly more popular than another at just the right point, it will tend to become more popular still. As a result, even tiny, random fluctuations can blow up, generating potentially enormous long-run differences among even indistinguishable competitors — a phenomenon that is similar in some ways to the famous “butterfly effect” from chaos theory.” Duncan Watts
Network Effects are perhaps the most coveted, yet mismanaged and misunderstood concept in tech businesses. While many claim to know it, only few I feel truly know it. And those that can master it for their respective business achieve the greatest success, and can apply that knowledge into other strategies of their business.
The basic definition of network effects, at its most simplest form, is that a product gets more valuable as more users use it.
It is a key competitive advantage a company can have and even more deadly than traditional moats of defense like brand, regulation, and IP. The concept is a defining characteristic in successful software businesses.
There are different “types” of Network Effects. Here are the main ones:
Direct Network Effects:
Increases in usage leads to direct increases in value.
Metcalfe viewed the value of a network as proportional to the number of connected users as (N^2).
While David Reed in 2001 argued that Metcalfe’s law actually understated the value of a network, Reed saw that the true value of a network increases exponentially (2^N) in proportion to the number of users.
Indirect Network Effects:
Increases in usage of the product spawns the production of complementary goods, subsequently increasing the value of the original product.
Two-sided Network Effects:
Increases in usage by one set of users increases the value of a complementary product to another distinct set of users.
Think ride-sharing applications like Uber and Lyft. More riders does not improve your experience, but it raises the end for more drivers, which improves your “_____”experience.
Able to get a car quicker, faster etc.
“In marketplace businesses, sell-through rate can also go by “close rate”, “conversion rate”, and “success rate”. Regardless of what it’s called, sell-through rate is one of the single most important metrics in a marketplace business. As investors, we like to see a relatively high rate so that suppliers are seeing good returns on the effort they put into posting listings on the marketplace. We also like to see this ratio improving over time, particularly in the early stages of marketplace development (as it often indicates developing network effects).” Andreessen Horowitz, 16 Metrics
One great way of thinking about two-sided network effects are how they are characterized by a supply-side and demand-side user base. They are there for separate reasons, but are in a perpetual state of motion that adds complementary value.
It is important to highlight a more complex phenomenon of network effects that shall be addressed later, that is particular to Two-sided Network Effects, is that value can be subtracted. But then in Two-Sided Network Effects, indirect benefits can also offset the negatives.
There are also circumstances that can promote positive same-side network effects, where those in the supply-side can add value to each other.
2-Sided Marketplaces like Craiglist are difficult to disrupt as a value proposition must satisfy and be better for both parties.
Two-Sided Network Effects are complex and fascinating, beyond the scope of this article. It is also worth considering practically, that quantum improvements in technology can radically overhaul two-sided network effects. Ride-sharing applications risk being disrupted on the supply-side with the rise of autonomous vehicles.
The graph above provides a 2-Sided Asymptotic Marketplace where one side can differ significantly in the “value curve.” OpenTable had to grow the supply-side of restraints to a very high level before there was any value on the demand-side.
2-Sided Asymptotic Marketplace are susceptible to multi-tenanting.
Local Network Effects:
Increases in usage by one set of users increases the value of other subset users in the surrounding domain.
It is easy to confuse this as only specific to a literal location, or neighborhood. However, instant messaging is an excellent example where a product can display “local”network effects for those that might be geographically apart.
Folk Network Effects:
I call these folk Network Effects because they are intangible and cultural than anything. VCs correctly stress Engineering, Timing, Durability of technology/innovaction etc. Objective metrics are indisputably important, but it fails to give weight to a. social network effects—those that emphasize psychology and interactions between people b. language network effects—those that emphasize the importance of the protocol in all nodes of a network as they interact c. belief—beliefs become more valuable to believers the more people believe, this relates to bandwagon, people don’t want to be left out.
Don’t just think about language network effects in regards to literally language (Mandarin, German, Arabic or whatever) but even to vernaculars and jargons to specific groups.
- Successful startups can use network effects to monopolize on business category language and naming a company or product. “Googling…Hello?”
- Succesful startups can also reduce their potential entry in creating a more accessible jargon.
Here’s to developing on a few key elements and implications to Direct Network Effects:
- A new entrant, even if its a significantly better product, cannot compete if it does not have a comparable network effect that produces a comparable amount of value as its competitor for the users.
- There are Physical Direct Network Effects—tied to physical nodes and links (eg. physical networks are utilities that monopolize into winner-take-all markets), Protocol Direct Network Effects—created when a communications or computation standard is declared and all nodes and node creators can plug into that network (eg. Bitcoin, Ethereum and more). It is worth noting that protocols can be extremely difficult to replace, and the success of such protocols depends a lot on competitive moats—which now seem outdated—like marketing, social engineering, and market niche.
- There are other variants of Network Effects. But, in my view, these are more “subsets” of Direct Network Effects; personal utility, which are built upon the personal identity of the users; personal, where personal identity and reputation is at stake (IRL) and enables you to create and maintain a public image; and market networks, which combines the identity and communication aspects of a Personal Network with the aspects of a two-sided network, and are intended to promote the building of long-term relationships.
- Some further subsets of Direct Network Effects; Data Network Effects occurs when nodes can feed useful data to the central database and depends on the aggregate data accretion; Tech Network Effects, runway advantage in that tech performance network effects become faster and cheaper the bigger they get.
Network Effects are often enhanced by introducing a concept of virality. A viral product often occurs when a rate of adoption increases with more adoption; this affects not really the network effect, but particularly its speed.
While network effects and virality magnify the value and growth as more users adopt the product, there are limits. This relates to a Two-Sided Network Effect and marketplace, and can be explored on a separate post. Regardless, I believe CEOs of large network effect enterprises need to begin considering sustainability of product.
As it relates to growth and virality, here are why network effect are so powerful. Value = Exponential and Cost = Linear.
One trend that is prevalent are the rise of content networks versus connect networks. Think (Twitter, Facebook, and LinkedIn) versus (Behance, Pintrest, Instagram, Dribble and Scoop)
Something I want to stress is not all network effects are created equal.
I believe, like I mentioned above, in two-sided network effects and their sustainability. Certain systems, network effect or otherwise, have a threshold capacity and can collapse or significantly stagnate as their are simply too many users who lack value in a system.
Furthermore, I believe that founders in startups and CEOs need to be clear on their product and its network effects. Certain network effect systems, like two-sided market place network effects can offset their negatives. However, the overlap of various Network Effects—when Two-sided network effects combine in potency or direction, to the product’s overall detriment, with Data Network effects.
One of the most common concepts somewhat—I say somewhat because I think questions of sustainability and positive growth in related to value proposition are not stressed enough with critical mass—relating to the idea expressed above, and Network Effects as a whole, is critical mass: the size a company needs to reach in order to efficiently and competitively participate in the market, or the level of users that are required to help create a set of network effects that are so strong, that a moat is erected for that business.
The application of a mental model from physics to other disciplines – as well as business or life questions — is intended to be a method of improving general thought. But it’s pure folly to assume that formulas from physics can be applied in human affairs and produce the same predictive outcomes. As Richard Feynman says, electrons do not have feelings like people. The real world — which is a nest of complex systems — can’t be modeled like a physical system. The trick is to apply the basic ethos of physics in your metal model without assuming that the real world can be modeled with formulas containing Greek letters. Two Powerful Mental Models: Network Effects and Critical Mass, Tren Griffin
Network Effects can be discussed completely incorrectly when they are oversimplified.
I believe the following discussed in this blog are best discussed in a separated post:
- “Butterfly Effect” and Network Effect. Related to Critical Mass and Sustainability.
- Dicussing Network Effects in theory and practice.
- Folk Network Effects.
These are the most important arenas of network effects that I belive, elevate from discussions.
There was a lot of key words in this article. Here are three takeways from Network Effects:
1. There are five key network effects to know: Direct, Indirect, Two-Sided, Local, and Folk. Significant overlaps in overall “themes” from the supply-side economies of scale and demand-side economies of scale, complementary value and sub-networks, like Reed’s Law to look into.
2. Virality and Critical Mass are two other concepts to know when discussing Network Effects. Virality refers to speed of adoption and Critical Mass refers to the amount of users where there is a moat— like brand, embedding, and scale—replace old ones like IP or even technology.
3. Network Effects are incorrect when viewing a pure increase users as immediately translating to value growth. There are subsets in network effects, as there are for categories.
This article is great: