Computer Guy

Computer Guy
Sunset at DoubleM Systems (DBLM.com), Del Mar, California

Tuesday, June 16, 2020

Metcalf's Law

Metcalfe’s Law: Why big networks produce colossal winners

In 1980, Robert Metcalfe — one part of the duo behind Ethernet, a technology for connecting computers together — observed that communications networks increased in value in proportion to the number of users.
One telephone in a city was useless. Even a few hundred, dispersed across millions of inhabitants, wasn’t very useful. But once you had friends and family who owned telephones — and restaurants and movie theaters and stores all had telephones — it was suddenly very desirable to own one.
Metcalfe was working at 3Com, a computer network company he co-founded, when he came up with the idea. At the time, he was trying to understand why his company’s local area network (LAN) starter kits — which allowed 3 PC users to share a printer and a hard disk — weren’t selling.
What he found wasn’t necessarily that the cost was too high, but that 3 people didn’t constitute a sufficiently large network to justify the purchase.
Like early installations of the telephone, the cost of connecting the network initially exceeded the value generated — but there would also be a point of critical mass where the benefits of connecting did exceed the upfront investment.

HOW FACEBOOK LEVERAGED METCALFE’S LAW TO BECOME A SOCIAL MEDIA GIANT

When Facebook first launched at Harvard University in 2004, it was like a singular telephone: pretty useless. But it wasn’t useless for long. Within a month, 50% of the student body was on Facebook.
This was partly a case of pent-up demand: every year, Harvard released a physical “face book” that included every student and faculty member on campus, designed to help everyone at the school get to know one another. “TheFacebook,” as the social network was then known, was an attempt to digitize this already existing physical object and make frictionless the process of snooping on other people’s photos.
The effect, however, was this: with every new Harvard student that joined the platform, it became more valuable for other students to join too. Each new student promised new people to look at, learn about, and “poke.”

As Facebook grew, the company added features that were designed to tap into the “more is better” mechanics of Metcalfe’s Law. Photos, groups, likes, comments — they all leveraged this concept to bring users back into Facebook again and again. And the more people there were on Facebook, the more photos were uploaded and tagged, the more groups were created, and so on.
The non-linear effects of Metcalfe’s Law would also contribute to the long-term viability and success of Facebook’s advertising business. The more people on the platform, the more data and connections, the greater the revenue opportunity in targeted advertising.
Metcalfe himself acknowledged the impact of this law on Facebook’s growth, observing that if Facebook’s revenue was used as a proxy for its value, it was impossible to deny that the network’s value had risen exponentially compared to the steady linear growth of its user base.

DEATH SPIRAL: WHAT HAPPENS WHEN NETWORKS DON’T REACH CRITICAL MASS

According to Metcalfe, a network’s critical mass is a function of the cost of a new connection (e.g. the cost of acquiring a user), the number of users, and the value of each connection.
This approach describes a few mechanics key to network effects. The lower your cost per connection, for example, the lower the number of users that you’ll need to hit critical mass — and the same goes for a higher value per connection.
Andrew Chen, partner at Andreessen Horowitz, has written not only about how the mechanics of Metcalfe’s law are key to growth at companies like Facebook, Twitter, and Snap, but also how the law can work against these networks if user retention isn’t high enough.
In one scenario, each new user that comes to a platform makes the experience better for everyone else. Growth is self-perpetuating because not only are users encouraged to stick around, but there’s an increasing value proposition to entice new users at a steady clip.
In another, not-so-good scenario, new users may still be joining and helping generate the benefits of network effects (for a time), but the network’s retention isn’t great.
If a network starts losing too many users, it can get stuck in what he calls a “social network death spiral.” Because while the forces described by Metcalfe’s Law can help startups gain lots of users quickly, they can also cause those startups to lose users at the same pace. In other words, “as you lose users, the value of your network decays exponentially.”
What’s lacking in this reverse-network effect scenario is the value offered by joining the network. If that value is too low, retention will be low, and it becomes difficult to reach critical mass. Meanwhile, if the value is high, your retention will likely be high as well.
Chen’s death spiral provides a way of understanding what happens when social networks take off for a time but ultimately don’t give each new user sufficient value to stick around — a fate that has befallen prospective social networks from Ello to Path to Peach. These apps often launch with a bang, attract initial interest, and then rapidly lose their user base as the promise of the network fails to materialize.

TAKEAWAY

The telephone, the fax machine, and early LAN networks had what venture capitalists today call “network effects.” The more people that were on these networks, the more valuable it was for new users to join.
This is one of the most important dynamics for businesses in the 21st century, especially in tech. Many of the most valuable businesses of the last several decades, from Microsoft to Facebook to Airbnb to Uber, have succeeded in part thanks to the power of network effects.
While building a business through network effects was valuable before the internet, the internet has made it possible for companies to grow their user base at exponentially faster speeds, often building their entire business models on this growth.


Consider this interesting example of Metcalfe's Law that happened to software I created before the internet existed: As the number of users of TeleMagic grew, it became easier to sell it because there were more people who were trained on the software and therefore easier to hire someone. And when a user left his/her employer for another similar job in another company, they would always recommend the product they knew best... TeleMagic! Other software products were left in the dust.