Network Effects

Recently, I’ve come across founders using increasingly liberal definitions of the term ‘network effects’. Often, these inaccuracies portend ill-focused strategy that belie grandiose claims, and gloss over important issues that are of paramount importance at the early stage.

Accordingly, this piece seeks to iron out some of the misgivings that surround this topic, and show how start-ups can ingrain the value of network effects into their companies.

By way of introduction, a definition:

A network effect occurs when a product or service becomes more valuable to its users as more people use it.

Historically, network effects were demonstrated by companies creating hardware products. The more people that had telephones, the more useful it became as a tool for communication.

In the software age these kind of network effects are, arguably, more important than they were in the past. Understanding their implications helps entrepreneurs build better products, protects them from competition, and reduces acquisition costs while simultaneously increasing engagement.

Although an oft-repeated trope: network effects create ‘winner-take-all’ markets. The preponderant player in the sector will have created significant barriers to exit for users, huge hurdles for competitor entry, and protected itself against price pressure.

To avoid alienating readers not wishing to wade through the boggy terrain of dense theory, it’s important to remember that what we are trying to determine:

What is the initial growth strategy that we can employ that will help us achieve scale whilst bootstrapping?

To provide any valuable guidance, we need to ask the right questions. Unsurprisingly, they aren’t the same for all businesses; so, a small exercise in taxonomy is required.

Most software companies looking to harness network effects can be classified under three broad categories:

  1. Network: A network company involves a group of interconnected people, or connected hardware systems. E.G Facebook (at it’s advent, now it displays features of both of the following)
  2. Marketplace: A marketplace is a type of network where transactions flow between two distinct groups of users on each side. E.G Ebay.
  3. Platforms: A platform is a network of users and developers. It allows customisation and programming to extend the parameters set by the original creators. E.G WeChat.

Now, we’ll explore each of these in more detail.

For network companies, the initial growth strategy should be focused around securing more users — nothing novel there. The idea is to reach a critical mass of users that begin to see value of having more people using the product. Founders should consider: what - and for how long - tactics are they going to have to employ to reach inflection point?

Facebook is the quintessential example of a company that’s bootstrapping phase was epitomised by strategic awareness of network effects. Their success can be analysed using the following loose methodology:

a) Value proposition: Contrary to assumptions, Facebook provided instant value for a single user who logged onto their website. It began as an online student directory with information that was immediately useful for a single person. Using this as an entry point, they quickly transitioned into connecting people who were studying the same courses or had joined the same clubs.

b) Growth strategy: To achieve growth, they had to continually add more information to their student directory. This was crucial in driving early adoption. Similarly, their product had virality (more on this to come later) — something they optimised for — which facilitated organic growth.

c) Engagement: After having attracted a number of users to their platform, Facebook recognised the importance of retaining them, hence developed a successful engagement strategy. They identified that engagement was significantly higher for those with a greater number of friends, so, they suggested friends, imported email contacts, built widgets, to ensure that a new user had 10 friends within two weeks.

Facebook has evolved significantly since then — relationship status, timelines etc. — and demonstrate the presence of a network effect by showing:

as user numbers rise, usage rates are also increasing.
For marketplace companies, the guidance given to founders is often hackneyed, but important. The key consideration in the early stage is how to build liquidity in the marketplace, often known as the chicken-and the egg problem. Similarly, these companies should consider which of the two ‘sides’ is going to provide money in the short-term, and which would benefit from subsidies.

Using a slightly tweaked version of our earlier methodology, Air BnB’s shows good execution of key marketplace principles:

a) Value proposition: Utility in saving the user between 30–80% of their short-term accommodation bill. Additional value provided in more differentiated, social residencies.

b) Growth strategy: Targeted certain cities at certain times where there would be a marked strain on affordable hotel space e.g Super Bowl, DNC convention. Used traditional marketing to drive the demand (believing, correctly, that supply will often followed when given the right support).

c) Critical mass: To reach the inflection point, they subsidised and supported the supply side. Notably, they offered hosts professional photography services free of charge. They also realised that a lack of trust was hampering their growth. So, they introduced a peer-review rating system.

Air BnB demonstrated network effects relatively late. From conception, it took the founders around three years of slow growth before they were set. Whilst virality can help achieve critical mass, Air BnB didn’t have a viral product, yet they still show network effects: more hosts, lead to greater availability for guests, who then in turn become hosts.

A similar example is Medium. Both sides of the network mutually reinforce each other, like Air BnB. More writers encourages more readers, which in turn leads to more writers. The highlighting features, mentions and replies have turned the site into a forum for debate that becomes more valuable to users.

Medium began by providing immediate functionality for the user. It can be used in ‘single-player’ mode by those desiring an intuitive, well-designed publication tool. In order to grow into ‘multi-player’ mode, where it could foster knowledge sharing, they encouraged influencers to write and share articles. They engaged customers with smart product development and were able to cluster communities based on their interests. One big lesson for marketplace companies is to encourage users to:

come for the tool, stay for the network.

Common misconceptions:

Network effects and virality are not the same thing.
  • Often, I encounter pitch decks where these two terms are used interchangeably. The two can often occur simultaneously, which may explain part of the confusion. To clarify:
Network effects increase the value for the user, the more that other people join. Virality simply refers to the increase in the speed of adoption of a product.
  • There are different types of virality. Traditional virality is often associated with word-of-mouth recommendations: i.e. a customer recommends a product due to a positive experience with it e.g Candy Crush/Angry Birds. Other types of traditional virality derive from good customer acquisition practice e.g encouraging recommendations with a financial incentive (Dropbox), giving the appearance of exclusivity (Gmail), or clever hacks (Hotmail). Whilst all may have grown quickly, and may subsequently develop network effects, this type of virality provides no assurances of future network effect. Take Angry Birds, while it displayed viral growth, and is still a valuable brand, it doesn’t increasingly benefit current users with further user growth (arguably, the same could be said for Pokemon Go).
  • Conversely, a product may have inherent virality, that is: by virtue of it’s very use, it encourages rapid adoption from other people e.g Snapchat. This is incredibly valuable as it means CAC is basically $0, and it leads to better retention through increased engagement.
If a product has scale, it doesn’t necessarily mean that it has network effects.
  • Again founders sometimes confuse economies of scale (“EOS”)with network effects. In EOS products become cheaper to produce as businesses increases in size. Essentially, unit costs drop as fixed assets can be distributed over a greater number of goods. The reduction in price has no correlation in value attributed by the user.
  • That’s not to say that EOS aren’t incredibly useful: Amazon’s purchasing power and ability to dictate terms is a testament to their success in this field. Their e-commerce platform is accordingly very valuable. However, they have network effects in their P2P marketplace (where individuals can buy and sell to other individuals) and, interestingly, this is growing much faster than the e-commerce side.

Finishing up, a few take-aways. Network effects can be super useful in B2C products or services. So, it makes sense to build your company in a way that maximises your chance of harnessing this opportunity. The upshots are huge: moats from competition; better engagement and accordingly retention; winner-take-all markets. Looking at some of the best practice in the sector leads us to a few suggestions: a) identify what kind of network effect you are looking to achieve and ask yourselves the right questions; b) provide single-player utility and then hook users with the network; c) as with any start-up: have a defined value proposition, growth strategy and engagement triggers; d) use a clustered approach to acquire segments of your target market; e) build products to optimise for engagement — it is increasing usage, not just growth that shows network effects; f) set yourself goals to determine what point you think you have reached critical mass, and do everything you can to get there!

Best of luck.