How defensible are Zoom’s network effects?

Why ease of adoption and coordination can undermine network effects and what to do about it

Apps have become easier to install and more intuitive to use, and they are making it easier for groups of people to coordinate and join together (e.g. one person can invite their friends to join any video chat app via a link sent by text or email). This is a double-edged sword for network effects. On the positive side, it makes it easier to solve the cold-start (chicken-and-egg) problem and get more adoption faster. The downside, however, is that it makes those network effects less defensible. Few investors and entrepreneurs seem to recognize the downside, which is the topic of this post. We will also discuss some design features that can help strengthen network effects and their defensibility.

It’s useful to first look at a few examples:

• Video chat apps like WhatsApp, Skype or Houseparty. The initiator of any given call chooses the app and can send an email or a text to the participants who don’t already have it, inviting them to download the app and register.

• Videoconferencing apps apps Zoom, Google Meet or Microsoft Teams make it even easier for meeting organizers to ensure participants join by sending them an email and/or calendar invite. By now all three apps allow participants to join a call in a browser, simply by clicking on the link sent by the host and without needing to install the relevant app or creating an account (the host must always do both).

• Evite and Eventbrite enable event organizers to create and send invitations, and then manage communication with their guests, while Pigeonhole Live makes it easy for conference organizers to interact with their audience (e.g. live polls, Q&A, etc.). In all three cases, one side (the organizer) makes the initial platform adoption decision and sends the invite to the other side (participants) via email, web link or QR code.

All these examples feature network effects: the value to a user (host or participant) of adopting a given app is increasing in the number of users that have adopted the same app. To be clear, adoption means downloading the app, creating an account, or at the very least learning how to use its features. The problem is that all these examples also feature very easy adoption, as well as very easy coordination by the host and participants of any given call/event/conference on the app they like the most. This suggests that network effects do not provide a lot of defensibility in these examples. For example, if Pigeonhole Live had a competitor called Seagullhole Live (we made this one up) with slightly better features and/or slightly lower price, any conference organizer reached by Seagullhole’s marketing should choose it over Pigeonhole – and all her conference participants would simply follow.

The only caveat is that hosts may want to choose the app that is most widely used in the general population if somehow that makes participants more likely to join a given event (call, meeting, conference), or the interactions in that call better (because more people have experience with the relevant app). For example, one of us shuns meeting invites using something other than Zoom in order to avoid having to figure out how to get other apps to work smoothly. And when deciding which video conferencing app to use for an online seminar series, we decided to go with Zoom because we thought more of our likely participants will know how to use Zoom’s features for asking questions, chatting with other participants, and adjusting video settings, which ensures the seminar experience is better for everyone.

From the discussion above, we can extrapolate an important and fundamental principle. To create strong defensibility, network effects must be accompanied by

a) high switching costs

OR

b) high costs of coordinating adoption among users that end up interacting with one another.

The most powerful network effects have either a), or b), or both.

For a), switching costs are high when users must incur high costs (monetary and non-monetary) to adopt competing services. These costs are the time, effort or money that users have invested to get more value out of the service in question, and that they would need to invest again in order to obtain the same value out of a competing service. Examples include the hundreds of dollars a user would have to part with in order to switch from an iPhone to an Android or from PlayStation to Xbox, as well as the time spent learning how to use the new device and the money spent buying apps or games for it.

For b), think of how important the discovery of sellers/buyers is on Airbnb, Alibaba, Craigslist and eBay. Switching costs from Craigslist to an alternative (e.g. OfferUp) are not very significant for buyers and sellers, but a buyer can’t coordinate to move onto that alternative with sellers she doesn’t yet know! As a result, she goes to Craigslist, because that’s where she expects having the best chance of finding a suitable seller. And vice versa from the perspective of the sellers.

In some contexts, the costs of switching (a) and coordination (b) are in large part driven by exogenous factors. Smartphones and videogame consoles are expensive pieces of hardware to produce, whereas accessing an online service is inherently cheap. Buyers who go to Craigslist or eBay are intrinsically looking to discover new sellers, whereas on Zoom people want to communicate with people they already know and there is no demand for discovery (no one wants random strangers to be able to video bomb their Zoom call). This is of course different from Clubhouse and other social apps, where users are specifically looking to spontaneously discover and join conversations with strangers.

However, in other contexts, these costs can be significantly affected by design. This means companies can make decisions that enhance the defensibility of their network effects.

To achieve a), one should add features that allow users to get more value when they use the service more often. That value will act as a switching cost when considering using other apps. For example, video conferencing apps can allow users to save frequently used contacts or groups, customize virtual backgrounds or skins, create convenient integrations with other apps and services (e.g. calendar, polling software, social networks), set-up payment details so a participant can easily pay a host, etc.

To achieve b), one should add features that facilitate the discovery of new interaction/transaction partners. This makes it harder to coordinate on switching to a different app, simply because it is impossible to coordinate with interaction partners one does not yet know about. Eventbrite has improved its defensibility by adding the ability for users (organizers and participants) to discover other events they might be interested in. Pigeonhole Live could try something similar for conferences (live and virtual). Another example is Freeplay, which allows existing groups of friends or co-workers to work out together at participating fitness facilities. The problem is those groups can just as easily join a different platform that gives them access to suitable facilities. To make its network effects more defensible, Freeplay can enable users to discover and join new groups that they were not part of when they joined Freeplay. Similarly, Playbook enables users to take virtual fitness classes or programs from well-known trainers. Playbook encourages trainers to bring their existing social media followers on to the app. That is of course a great way to overcome the cold start problem and get the network effects going. However, those network effects are not very defensible unless there is a strong element of discovery, i.e. users follow one trainer onto the app and then discover other trainers they like.

To sum up, having low costs for users to adopt and coordinate on a product makes it easier to launch products with network effects and for them to go viral, but later in their lifecycle, such businesses should focus on adding features that create stickiness and/or facilitate discovery so that their products enjoy long-term defensibility.

So how defensible are Zoom’s network effects? Zoom pioneered almost frictionless adoption in the videoconferencing space. While this approach has allowed Zoom to build up a massive user base, it has been copied by its competitors and Zoom’s defensibility remains limited. To illustrate, imagine Zoom tried to charge individual users a monthly subscription fee to initiate Zoom calls (or a fee for each such call). We expect such fees would lead many such users to shift to one of the many alternative free apps, given how easy it is to coordinate on and adopt one of these. And after Zoom’s security flaws came to light at the start of Covid, some enterprises users did switch over to more secure alternatives such as Microsoft Teams.

On the other hand, Zoom does enjoy some defensibility among business (or pro) users who may particularly care that their employees, business partners or customers are familiar with using its features. That preference will only grow as Zoom adds more features that enhance the value users derive from using its service (e.g. creating and managing breakout rooms) and that would take some effort to relearn on a competing service. Moreover, since 2018, Zoom has been building up a marketplace of third-party tools that complement its core product and create additional network effects (between users of Zoom and the third-party providers) as well as adding further stickiness to its core product. All these efforts should help Zoom build a more defensible moat in the years ahead.

---

If you enjoyed reading this, sign up now so you will receive our posts directly in your inbox as they are published.

In the meantime, tell your friends!