How to operationalise 'network effects'
The promise of ‘network effects’ is a common feature in startup pitch decks, explaining how a platform-based product will grow. Businesses dream of the miraculous flywheel off 'network effects' that converts more users as they interact with their product.
This is undoubtedly a great idea to scale the use of products and services, but how is a ‘network effect’ actually created?
Over the last few months, I’ve had several conversations with scaling businesses about how to build more stickiness into their products and services. While some have been further along the journey than others, they are all trying to build a new system that aims to be picked up by large numbers of customers across many geographic regions. In these conversations, I’ve found myself referring to the same counterintuitive idea about what it takes to get things moving: that to build a network, you need to start by building stability at the smallest atomic unit of the network.
To build a network, you need to start by building stability at the smallest atomic unit of the network.
This idea is at the core of Andrew Chen’s book The Cold Start Problem. As a general partner at leading venture capital firm Andreessen Horowitz and a former Uber operations executive, Chen has deep experience in observing and understanding network effects. Why is starting small to build a network counterintuitive? Networks sound big. You picture mass take-up of your product, so you think of strategies on this scale (e.g., regions, markets, your whole target customer demographic). But, as Chen describes, no ‘big bang launch’ and ad spend will help if the fundamentals don’t work at the smallest viable network scale. Chen outlines many examples of this observation from many well-known technology businesses. What I found insightful in these examples is the simple truth that some of the most ‘scalable’ network effects in businesses must be approached in a very focused way. Paying attention to local nuances that may not be obvious to an outsider. I’ll explore this concept through three of Chen’s examples, Zoom, Slack, and Uber. Through Chen’s examples and some additional reflections, I’ll outline why building robust and sticky networks where users must be focused on the smallest viable scale. If these ‘atomic’ networks are strong, growth will emerge.
In the case of Zoom, you might think a whole organisation has to adopt zoom to be sticky, but you simply need two people for the product to work. First, you need a single person to start inviting people to join calls on zoom rather than on another platform. The network can scale quickly once users begin to see the simplicity of connection and the quality of the call. This contrasts with the traditional model of Skype, for example, which by design precluded calls to people that didn’t have an account and the app downloaded. Zoom also outperformed in call quality, which made users keep using it. Zoom just worked. It is a shining example of product-led growth, where the product’s design drives the network’s expansion. For the first two years of its existence, Zoom had no marketing team at all. The stability of the network comes from users who could convert new users through an open system, and once people use it, they keep using it.
For the communications platform Slack, you might think it needs to be sold at an enterprise level to be adopted by an organisation. But unlike its incumbent competitors like Microsoft and others, Slack took a different approach. Slack can be adopted by individuals and small project teams with a free tier. Its growth came from small teams using Slack independently of the rest of their firms. A team of two could get benefits using Slack, but with three or more users, the stickiness would multiply. For every person who joins the network that Slack creates gets more valuable, increasing the likelihood of others following. Slack started by getting their friends to try Slack in other silicon valley startups, not by attempting an enterprise sales approach. The stability of the network comes from small teams who became addicted to Slack, pulling new users into their conversations.
Uber’s ride-sharing success in recent years is astounding. Today they have achieved dominance in many cities. However, contrary to a general understanding, Chen outlines that Uber’s victories have not been won at a country or city level but have been won tactically at a street level. It has been their responsive operations team to win over competitors. Uber has been competing against the incumbent networks of taxis, walking, public transport and people driving themselves, and other direct new competitors, Lyft, Didi, etc. The new competitors like Lyft and Didi required the most complex and tactical approaches to overcome them. In this case, the ‘hard side’ of the network is the limited pool of drivers pulled from different networks. With a swipe on their phone, drivers could switch and take jobs from other rideshare companies. Uber had to ensure that they had drivers working for them and only them as much as possible. They used a range of incentives rolled out in close to real-time to persuade drivers to work with them at critical locations and times. This dynamic would reduce the quality of their competitor’s services while increasing theirs. The stability of the network comes from the repeated availability of more uber drivers than competitor drivers at a specific time and location.
Overall, Chen encourages leaders to think through how to increase acquisition, engagement, and retention from their details. If any of these are letting you down, your network effects will be short-lived, or more likely, never eventuate. Every product and service relationship will have a different emphasis across elements of user growth. Rather than architecting the whole and trying to build the ‘complete’ system at once, focus on smaller ‘atomic’ networks first. Be specific about how your first networks become stable and grow.
I find with systems concepts, there is always a metaphor in nature to bring these ideas to life. This one is quite simple: scattering seeds across an open field gives no certainty that any will germinate. When resources are limited, you need to ensure a higher chance of survival. You need to look for the right places to nurture them when they are in this fragile state. Your efforts are better spent focused on the germination and stability of just a few, giving them the effort in ensuring correct soil, watering, sunlight, clearing weeds, and protection from pests. This combination of resources creates the conditions for the network to begin. Over time these will grow and make the enabling conditions for many others to flourish.
So, what does this mean?
The implications of this insight are broad and can impact your product strategies, partnerships, communications, marketing and recruitment strategies. While the examples above are technology-based, I believe the same patterns apply to any other type of business. Getting a network started from scratch is extremely difficult. Scaling that network effectively is even harder.
Across the Snowmelt client base, each organisation has different approaches to addressing this problem, but I believe that Chen’s model is an excellent way to think about what to do next. It’s a balance between increasing acquisition, engagement and retention.
Reflecting on Snowmelt as a business, we don’t offer a platform-based digital product. Instead, our atomic networks are built with individual clients, their trusted colleagues and friends in their network that they may listen to for guidance. We are continuing to explore how our work can increase the stability of networks around new ideas, strategies and transformations.
The idea discussed here is just one of the many that Chen builds on through the book. If you're responsible for products and service growth, I highly recommend Chen's book the The Cold Start Problem as a way to think through what this might mean for you.
Want to keep up to date? Subscribe to Snowmelt News.