Not Network Effects
Now that the network effects playbook is well known to founders and investors, founders are eager to claim it. Unfortunately the term is often commingled with things that are not network effects.
November 20, 2015
Venture capital firms, including the one I work for, like to invest in businesses with network effects. For many readers of this blog, the term “network effect” is already very familiar, but because this post is about how the term is often misused, here’s a brief and early aside on what does constitute network effects:
A network effect means that the value of a product increases with every new participant. It could be a network effect between users — when a new user of, say, Instagram, begins posting photos, it meant that every existing user has more photos to view and interact with. There are also technological network effects — a mesh network, for example, grows stronger with every node — and data network effects — machine learning gets smarter with every new data source it can interact with.
Network-oriented, internet businesses is that they are more defensible than businesses that don’t have network effects. Even in the face of stiff competition, or uncertain economic or technological conditions, networks are resilient. The abundance of success stories among network-effect businesses in the internet era speaks to why venture capital firms like to invest in them. (By the same token, the abundance of success stories among non-network businesses [e.g., Warby Parker, Casper] led by high-caliber, high-character founders speaks to why investors like to invest in those businesses as well.)
Because venture capital firms like network effects, businesses that approach us are keen to claim network effects . I wasn’t in venture capital in the mid-to-late 2000’s when network effects were a novel idea in the venture capital world, nor even in the past few years as almost every venture capital firm caught on. But in this past year, the one year that I have worked in venture capital, I’ve observed that founders are increasingly including a “network effect” slide in their investor pitch deck while commingling Network Effects with other things. (I’ve also made the n00bie mistake of asking founders about network effects when it wasn’t appropriate, and therefore railroaded founders into giving me an affirmative answer, so I’m as much to blame for mangling the definition as anyone I’ve spoken to.)
I’ve seen a few professions of network effects that go like this: “The more people who use our product and like it, the more other people will want to use it.” This is as true of a good pizza place as it is a good internet business, and that’s why there’s already a term for it: Word of Mouth. It’s not a network effect because every new customer of a pizza place or, say, Casper or Warby Parker, does not make the pizza, mattresses, or eyeglasses better.
I’ve also seen things which start with: “Our product creates something to share, which…” As in the example above, there’s already a term for it: Virality. It might be that founders are using “network effects” instead of “virality” because the latter term is somewhat tainted from the days of Zynga, Viddy, and Upworthy, whose reliance on virality for distribution proved unsustainable. It’s also worth noting Louis Vuitton is an business based on virality, and it is not a network (but a very good business!).
Another: “The more people who use our product, the better rates we can get on the services we provide, the more services we can afford, the better our product gets.” This is actually Economies of Scale. It’s true of every business on Earth and therefore does not contribute to defensibility. Remember, “network effects” only matter insofar as they give you a competitive advantage. Note how WeWork benefits from Economies of Scale, and might always be larger and more valuable than desk-sharing marketplaces, but desk-sharing marketplaces are still more defensible — read: more defensible, not necessarily more valuable — because unlike a centralized, asset-holding WeWork, those marketplaces have no liabilities.
The very last idea I often see labeled as “network effects” is murkier. Recently I looked at a mobile business that is building something you’d think Apple or Google would want to build and own themselves, and the founder argued that only a team who was exclusively focused on this single problem could build a truly transcendent version of the product. The network effect in this business, as he put it, was that the more data input they got from users, the better the responses they could output. This would yield a better product, which would bring more users, which would further improve the product, and so forth.
On the one hand, the skeptic would say, “There’s already a term for this: ‘Analytics’”. Like Economies of Scale, this is true of every business: the more data you have, the better you can separate the signal from the noise. Walmart, for example, surely runs lots of analytics on its business, and the p-values on its findings increase with every new location it opens or product it carries. But it’s too much of a stretch to say that every new location or product adds value to all the other locations or products; rather, the new locations or products mainly just help to increase efficiency.
On the other hand, the optimist would point to the data set this mobile business was creating and then point to Google, which clearly benefitted from the feedback loop between searching and then clicking on the most relevant result for that query. In other words, that’s the machine learning data network effect that I cited in the above definition of network effects.
Which one is it, the cynical or optimistic view? Well, really, a lot of that just depends on how popular this service gets. If it gets scooped up by Apple or Google in the next year, there won’t be much of Network Effect story to tell one day. If it grows up to become the next Google, then some analyst in 2025 will be citing it in a blog post like this as a paragon of Data Network Effects.
That’s one thing I’ve learned about the venture capital business: Oftentimes, not only do you not know exactly what you’re passing on, but you don’t entirely know what you’re investing in. Network effect businesses sometimes pivot away from the things that generate the network effects, sometimes for very good and maybe even very lucrative reasons. And of course that’s not an argument to just throw your hands in the air and make it up as you go along, because the clearer, more incisive your ideas and beliefs are as a venture capitalist, the better you evade inaction on the one hand, and profligacy on the other.
