System Wide Network Effects in Mobile
In tech and in business (and in economics), network effects measure the value of a product or service against the number of people using it. When a network effect is present, the network becomes more valuable to each user when a new user is added.
The internet is perhaps the most famous stack with network effects. Web browsing wasn’t possible until technologies like TCP/IP proliferated in the 90’s, routers were installed underground, CDNs helped with content delivery etc. As the internet was built out, software was invented that could take advantage of having an open pipe to your home—e.g. Netscape.
In the modern era we think of services like Snapchat or Twitter. Networks effects in the mobile era have led to fast adoption and viral growth and have created hundreds of billions of dollars in economic value overnight. The reason mobile has accelerated network growth correlates with sheer size—magnitudes more people have smartphones than owned PCs, more developers exist on iOS and Android, and bandwidth is available anywhere, any time.
But to analyze mobile’s potential we should actually try to measure uncaptured value, not captured value. When you do this, it’s easy to tell that network effects will dramatically accelerate innovation over the coming 5 years, and also drive entirely new platforms such as augmented and virtual reality.1 The reason for this is something I call system wide network effects.
System wide network effects are network effects that take hold when adjacent parts of an overall system are built out—e.g. smartphones, wearables, sensor networks, new physical layers, blockchain etc. Network effects at these layers are incredibly powerful as they effectively unlock compounded value from previous layers that was ready to be extracted. For example, for your smart device to provide accurate indoor context, a developer-friendly sensor / location stack needs to be built out.2
In mobile networks, which effectively weren’t architected for data, and completely broke down for the few years following the invention of the iPhone, the stack is in fact extremely immature. Some layers are more mature than in 2007—e.g. LTE has fixed signaling issues with 3G and ‘solved’ bandwidth constraints. But many other layers have not been built.
A lot has been argued with respect to bitcoin / blockchain, but computer scientists tend to universally agree—some version of this is the future. Just as the invention of TCP/IP was core to the internet, blockchain could be prodigious for services where trust and fraud are central issues.
In the sense of system wide network effects, you can actually think of blockchain as another layer of the internet stack. Why? Because intelligence in networks now extends much deeper into network architecture, and allows us to do much more with the bits we move around. Computer science breakthroughs are extending deeper and solving previously unsolvable problems. This is fact.
Yet today’s reality is correctly argued by Fred Wilson in this piece on network decentralization—the blockchain layer is in relative immaturity and existing layers (apps and at-scale services and connected devices themselves) cannot yet take advantage.
Platform history tells us something very important here: only after infrastructure is laid do developers write software. Different pieces in the stack need to work harmoniously, tied together by software. And only after the infrastructure is in place can developers invent new software. Much of today’s ‘bubble’ narrative can interestingly be explained away by system wide network effects.
Just look back at ideas in the late 90’s and ideas today—almost every service that was invented then is now emerging as successful. Webvan vs Instacart, Pets.com vs BarkBox, Metricom vs SpaceX. The similarities in business idea are remarkable. One reason is the mobile stack has enabled this software to actually work. A lot of the carnage of the 2000 bubble was because people were getting too much venture money to build out services before the converged internet stack was mature.
And this is why we are now entering the golden era for mobile—breakthroughs will be absolutely staggering to witness over the next 5 years as new stack layers mature… And it’s also precisely why investment in new technology such as sensor networks and blockchain and wearables is actually dramatically low today. A new post-mobile future is arriving much faster than many think, and there will be many new winners along the way.
these will inevitably be called post-mobile when they arrive to the masses ↩
Benedict Evans of a16z and I did an awesome podcast on this topic. ↩