4 ms·

> Network effects are unnatural online. I find this line ironic since the Network Effect (or Metcalfe's Law) was created to describe literal computer networks.

> Network effects are unnatural online. I find this line ironic since the Network Effect (or Metcalfe's Law) was created to describe literal computer networks. Network effects don't happen in the physical world nearly as easily because the bar, grocery store, or whatever is limited by geography. They can only address a population within a certain radius and can't grow beyond it. Transportation hubs - like airports - may be an exception because the more places you can get from one airport, the more likely you are to use that airport.. if you even have a choice.

cookiecaper 10y ago
>I find this line ironic since the Network Effect (or Metcalfe's Law) was created to describe literal computer networks. The distinction is that it refers to physical networks (and iirc, referred initially to phone systems, not computer systems per se). At layer 1, the network effect is still very much valid and caused by the capabilities of the physical hardware. This is forgotten because today everyone is plugged into the same network. This was not always the case. Because Metcalfe's Law exists, people have misunderstood it to apply to cyberspace and incorrectly believed it's simply the way things have to be. This is not so. The compatibility concern at the core of real network effects, where the hardware buy will be dictated by what the most neighbors have, are not relevant once you have a physical hookup. In cyberspace, layers 2 and up, there is no physical limitation on compatibility. There are only digits (hence the term "digital") upon which computations are made (hence "computers"). If the correct computations can be performed (i.e., if you possess hardware that can convert the electrical pulses to the expected digits and software (drivers, browsers, etc.) that can decode the digits to match expectations), the desired data can be presented. Since network effects refer exclusively to physical compatibility, we only see this in the digital realm because it's artificially imposed. The law says that anyone who tries to extricate data from, e.g., Facebook will be severely punished, despite the fact that Facebook does not have a copyright interest in the data, and that as long as such extrication is done responsibly, it would have 0 impact on Facebook's ability to conduct its operations. That's an artificial barrier that we've concocted, and it's the only mechanism by which network effects can work in cyberspace. Without the legal risk, people would read the data out from Facebook and recast, repopulate, etc. in alternate hypermedia. People would also make writable interfaces available. Facebook would then have to compete on its ability to provide a superior user experience, like most other businesses, not the availability of the data that they've misappropriated from users. Social networks like Facebook have forced themselves between the open network and the data that used to be exposed to it, insisting that all traffic must be subject to their conditions. There's no reason to allow that! Once the online marketplace is subject to the same types of competition that exist in the physical marketplace, i.e., once everyone is allowed to compete over the same core of knowledge, things are going to drastically improve.
- dredmorbius 10y ago
  I appreciate your thinking, but your understanding and application of networks, as well as your decidedly wishful thinking that they don't apply online, is simply wrong. Mind: I'd much rather this not be the case, but it simply is. Metcalfe's function is, by the way, also incorrect. Very few networks are comprised of nodes of equal value, instead the nodal (and inter-nodal) value tends to follow a power function (Zipf's Law), with one suggestion being that the value function of a network with peer-to-peer connections then be: V = n * log(n) (Tilly-Odlyzko). I see even that as incomplete, as networks also generate negative interactions, though by their nature, Metcalfe's formula applies, with some constant, such that the network cost imposed by nodes is: C = k * n^2 Putting those together, a clear corollary is that the maximum feasible size of the network is dependent on the cost function constant. What you're failing to consider is that the members of a network, including both the users and advertisers, are themselves components of that network, and have their own value and cost calculations to be applied. For large advertising networks, the fact that 1) larger markets with 2) more customers can be approached with 3) fixed costs for the advertiser and 4) efficiencies of scale for the network operator (e.g., Facebook or Google, approximately 2/3 of all online advertising), then there will automatically be an advantage to larger networks, provided k is sufficiently small. There's nothing artificial about those barriers. Incidentally, you can find network or dendritic structures in any system in which you have nodes or elements, and links or relationships: transport and communications systems, cities, even knowledge or information itself (think of the terms "knowledge web" or "information web"). I've never cared for Facebook and am finding Google increasingly evil and/or incompetent (and am torn as to whether that last is a good or bad thing on balance), but there's very little validity in claiming that the dynamics which have benefitted them are artificially imposed.
  - wyager 10y ago
    Worth noting that Metcalf didn't actually think his "law" was particularly correct or insightful; he made it up for Ethernet sales pitches.
    - dredmorbius 10y ago

cookiecaper 10y ago

>I find this line ironic since the Network Effect (or Metcalfe's Law) was created to describe literal computer networks. The distinction is that it refers to physical networks (and iirc, referred initially to phone systems, not computer systems per se). At layer 1, the network effect is still very much valid and caused by the capabilities of the physical hardware. This is forgotten because today everyone is plugged into the same network. This was not always the case. Because Metcalfe's Law exists, people have misunderstood it to apply to cyberspace and incorrectly believed it's simply the way things have to be. This is not so. The compatibility concern at the core of real network effects, where the hardware buy will be dictated by what the most neighbors have, are not relevant once you have a physical hookup. In cyberspace, layers 2 and up, there is no physical limitation on compatibility. There are only digits (hence the term "digital") upon which computations are made (hence "computers"). If the correct computations can be performed (i.e., if you possess hardware that can convert the electrical pulses to the expected digits and software (drivers, browsers, etc.) that can decode the digits to match expectations), the desired data can be presented. Since network effects refer exclusively to physical compatibility, we only see this in the digital realm because it's artificially imposed. The law says that anyone who tries to extricate data from, e.g., Facebook will be severely punished, despite the fact that Facebook does not have a copyright interest in the data, and that as long as such extrication is done responsibly, it would have 0 impact on Facebook's ability to conduct its operations. That's an artificial barrier that we've concocted, and it's the only mechanism by which network effects can work in cyberspace. Without the legal risk, people would read the data out from Facebook and recast, repopulate, etc. in alternate hypermedia. People would also make writable interfaces available. Facebook would then have to compete on its ability to provide a superior user experience, like most other businesses, not the availability of the data that they've misappropriated from users. Social networks like Facebook have forced themselves between the open network and the data that used to be exposed to it, insisting that all traffic must be subject to their conditions. There's no reason to allow that! Once the online marketplace is subject to the same types of competition that exist in the physical marketplace, i.e., once everyone is allowed to compete over the same core of knowledge, things are going to drastically improve.

dredmorbius 10y ago
I appreciate your thinking, but your understanding and application of networks, as well as your decidedly wishful thinking that they don't apply online, is simply wrong. Mind: I'd much rather this not be the case, but it simply is. Metcalfe's function is, by the way, also incorrect. Very few networks are comprised of nodes of equal value, instead the nodal (and inter-nodal) value tends to follow a power function (Zipf's Law), with one suggestion being that the value function of a network with peer-to-peer connections then be: V = n * log(n) (Tilly-Odlyzko). I see even that as incomplete, as networks also generate negative interactions, though by their nature, Metcalfe's formula applies, with some constant, such that the network cost imposed by nodes is: C = k * n^2 Putting those together, a clear corollary is that the maximum feasible size of the network is dependent on the cost function constant. What you're failing to consider is that the members of a network, including both the users and advertisers, are themselves components of that network, and have their own value and cost calculations to be applied. For large advertising networks, the fact that 1) larger markets with 2) more customers can be approached with 3) fixed costs for the advertiser and 4) efficiencies of scale for the network operator (e.g., Facebook or Google, approximately 2/3 of all online advertising), then there will automatically be an advantage to larger networks, provided k is sufficiently small. There's nothing artificial about those barriers. Incidentally, you can find network or dendritic structures in any system in which you have nodes or elements, and links or relationships: transport and communications systems, cities, even knowledge or information itself (think of the terms "knowledge web" or "information web"). I've never cared for Facebook and am finding Google increasingly evil and/or incompetent (and am torn as to whether that last is a good or bad thing on balance), but there's very little validity in claiming that the dynamics which have benefitted them are artificially imposed.
- wyager 10y ago
  Worth noting that Metcalf didn't actually think his "law" was particularly correct or insightful; he made it up for Ethernet sales pitches.
  - dredmorbius 10y ago