Virtually all startup pitches include a conversation about competitive advantage or “moats,” and rightfully so; these moats are what ultimately allow companies to become compounding franchises and accrue outsized profits over the long run.
The most commonly sought moat in consumer technology is the “network effect”, which is when the value of a product increases directly as a function of how many others use that product. There are several types of network effects consistent with this definition, such as demand-side increasing returns, marketplace effects, and platform effects.
Network effects can be one of the deepest and most enduring types of moat and many great technology companies have benefited from network effects: Facebook, LinkedIn, WhatsApp, Snap, eBay, Airbnb, and Uber to name a few. If a startup can architect network effects into the product or business model, that is worth its weight in gold, and plenty of blog posts have been written about this topic.
However, there are real world problems to be solved and businesses to solve them which, no matter how hard one spins the story, will not be candidates for a network effect approach. The nature of these problems or products simply doesn’t allow for the value proposition to be directly tied to the number of others using the product.
The decision to stay in a Marriott isn’t driven by how many other people are staying there. Tesla owners do not derive utility from how many drivers around them are also in Teslas (in fact one could argue the opposite). When subscribers turn on Netflix or Spotify, the subscriber count has little impact on their experience. Developers choose to implement Stripe because it is the easiest payment gateway, period. Investors buy stocks on Robinhood because it is the lowest cost and best designed platform for doing so. Even the “act 1” of internet giants like Google, with its PageRank search algorithm, and Amazon, with its e-commerce business, were not products where users directly benefited from more users adopting the product.
I believe that important businesses will be built in areas without hard-coded network effects and that there are ways to become extremely valuable without them. So, when I meet with founders to talk about a business that can not be built on classic network effects, rather than dwelling on that topic, I find that the more interesting conversation is to ask the question: what kind of moat can you build?
Here is a list of non-network effect sustainable competitive advantages that are relevant today:
Economies of scale are nothing new; they’ve been a feature of business building since at least the industrial revolution. But the internet magnifies the importance of scale effects in many ways; most obviously by removing the constraint of geography and reducing the capital intensity of growth. The canonical example of scale in the age of the internet is Amazon, which has used its scale to achieve market leadership first with e-commerce and then with the AWS cloud business. Scale creates several types of competitive advantage:
- Negotiating leverage. Scale allows for procurement of inputs at the lowest cost, which can be passed to the consumer in the form of lower price. The lower price attracts more consumers, which in turn increases the company’s scale and gives them more negotiating leverage, etc. MoviePass, with their all-you-can-watch movie subscription, is attempting this playbook; by subsidizing the cost of the service today, its long-term strategy is to use a large subscriber base to negotiate revenue share with theaters. Walmart is a more traditional example of using power over suppliers to drive cost advantage.
- Amortization of fixed assets. Companies with more customers can spread the cost of fixed assets over a larger base. Digital subscription businesses like Netflix are examples of this. When Netflix produces or purchases the rights to a media asset, it can spread the costs of that over their tens of millions of subscribers. It would be difficult to start a Netflix clone today because the economics of buying or producing media would be far less attractive than they are for Netflix.
- Product density. In some product verticals, larger scale makes the product fundamentally more attractive, all else being equal. This phenomenon is similar to marketplace network effects but is not unique to marketplaces. The dockless bike and scooter startups (Ofo, and Mobike in China, LimeBike, Jump, and Bird, in the US) do not have traditional network effects, but are better products for users than smaller competitors because they have higher density of vehicles, which makes the service more convenient.
Economies of scale do not affect all industries equally. In fact, some industries exhibit dis-economies of scale, and have remained cottage industries for this reason. So, it’s important to think through the effect of scale on business model carefully to determine if there is sufficient operating leverage to justify scale as a moat.
Brand is one of the longest standing sources of defensibility. I believe that the internet has increased, rather than decreased, the importance of brand. Traditionally brands were built through advertising, word of mouth, and shelf placement at major retailers, all of which made it difficult for startups to compete on brand. The internet elevates the importance of trust in brand, which explains why social media (i.e., what are my friends and role models using?) has become an important channel in which to amplify brand.
Coinbase’s brokerage product is an example of how brand drives defensibility. An online brokerage business does not have classic network effects because it is built on top of exchanges (like GDAX, which Coinbase happens to own) where liquidity consolidates, so value is not directly tied to the number of users on the platform. But trust, and therefore brand, is extremely important in financial services generally and crypto specifically. By developing a trusted brand and reputation in an otherwise murky space, Coinbase has become the de facto way to access digital currencies. (disclosure: Greylock is an investor in Coinbase.)
The concept of brand can appear to be an obstacle (i.e., strong brand of incumbents) to a startup as opposed to a future competitive moat. That’s why some of the most successful startups have focused on creating a brand in a category where there was previously none.
WeWork, the 8-year old co-working startup which was recently valued at $20 billion, is an example of this strategy. WeWork’s business model lacks traditional network effects, but by becoming the category-defining brand in co-working they have been able to create an extremely valuable company.
Non-network Switching Cost
Switching costs is the one-time inconvenience or expense a user incurs to change over from one product to another. The strongest form of switching cost comes from network dynamics — other users being on a platform. But even in non-networked products, there are several forms of switching costs:
- Personal data. Some products are architected in a way where users build data within the product over time. Data has gravity and the more data a user generates within a product, the harder it will be for the user to switch. Strava (ignoring the social functionality) is an example of this phenomenon. After having used Strava for several years, switching to a new fitness app would mean orphaning my fitness history.
- Habit. Products that are used with high frequency tend to exhibit habit formation and once habit is formed, users are reluctant to switch. For instance, there is a shockingly small percentage of users that switch from either iOS to Android or vice-versa when they get a new phone, in part because users habituate to navigating a particular operating system. Because of user preference for the status quo, to replace a habitual product a new entrant must be vastly, perhaps even an order of magnitude, better, which translates to a moat.
- Embedding. This is a phenomenon whereby a product becomes integrated with a user’s other products and workflow. This type of moat is most often seen in enterprise software products; many enterprise software giants get their defensibility from high switching costs. Arguably the large cloud platforms like AWS and Microsoft Azure have embedding lock in. But embedding can work in consumer products as well. For instance, consumer finance products often exhibit embedding since users connect their bank accounts and payments cards to other applications.
The above is not a comprehensive list of sustainable competitive moats; others include regulatory, technical advantage, and cornered access to limited supply. But the takeaway is that, while network effects remain a powerful driving force of defensibility, they are not the only source of competitive moat. Founders should stay focused on solving a problem that they are passionate about and building defensibility as they scale.