Another look at network effects

A recent article in The Economist covering the current health status of some of the tech ‘darlings’ caught our eye:

Over the past year, the firms that epitomise these [new tech] business models — Uber and DoorDash; Netflix and Spotify; and Snap and Meta (which has tumbled spectacularly out of the trillion-dollar club) — have shed two-thirds of their market capitalisation on average …

their businesses all turn out to face the same main pitfalls: a misplaced faith in network effects, low barriers to entry and a dependence on someone else’s platform …

Network effects are real. But they also have their limits. Uber believed that its head start in ride-hailing gave it a ticket to riches, as more riders and drivers would mean less idle time for both, drawing ever more users into an unstoppable vortex. Instead, it encountered diminishing returns to scale: reducing average wait times from two minutes to one would require twice as many drivers, even though most riders would barely notice the difference.

Perhaps the issue in this case is not the limits of network effects, but that network effects alone are not enough. The companies that have made huge profits on the back of network effects have had another thing going for them: low incremental unit costs. The next search advert that Google posts, or software Microsoft sells, or workload hosted by Amazon Web Services costs those companies LESS than the one before.

As Robert Armstrong notes in a recent FT column:

Uber’s unit costs will never diminish as fast as [Google, MS or AWS], because it is a taxi company. It charges consumers for rides, pays most of that money to drivers and car insurers, and pockets a slice (about one-fifth of fares). Its much-vaunted food delivery business is little different: charge for delivery, pay most of that to drivers and restaurants, and take a cut. Uber doesn’t technically own the cars in its fleet, but ultimately it has to pay for their depreciation, by way of correctly compensating its drivers. In that sense, it pays for the gas burnt on each ride, too. And of course the first driver has to be paid as much as the last. Uber has the economics of an asset-heavy company because, in effect, it is one.

Ok, so network effects with low incremental costs are good, how does this relate to blockchain ecosystems and in particular Cardano?

We’ve talked before in this blog about the importance of network effects (or Metcalfe’s law as it is formally known), and how these are still the primary driver of valuations in the blockchain and crypto space (and likely to remain so for a while yet). In other words more users equals higher valuation for that cryptocurrency (all other things being equal), and this applies to Cardano in the same way as any other blockchain.

What about incremental costs and scaleability? Do blockchains like Cardano benefit from a reduction in incremental costs? It’s easy to see intuitively that they do:

  • The operating cost of the base network (Layer 1) is largely fixed, with at worst linear scaling for data storage, computational and network costs as transaction volumes increase.
  • The development cost of the base network is largely fixed irrespective of transaction volumes, therefore incremental costs reduce.
  • The development cost of smart contracts and dApps are largely fixed irrespective of transaction volumes, therefore incremental costs reduce.
  • Scaling technologies in Layer 2 such as Hydra in the case of Cardano promise to dramatically reduce incremental costs for certain use cases where transactions can be managed in a state channel and periodically synchronised with Layer 1.

Transaction costs are fixed on Cardano and therefore do not alter with volume and are scaling neutral. There is an argument that a tiering structure will be needed in future (high, low and no priority for example) but this will also be neutral. Competitors such as Bitcoin and Ethereum which operate fees markets see increased incremental costs as demand increases, the opposite behaviour needed for wide adoption.

To summarise, an overall reduction in incremental costs as shown by Cardano is very positive for adoption, and cryptocurrency ecosystems that have this trait and achieve critical mass in the coming years will be very hard to displace.

Especially L1 ecosystems with no dependence on other platforms, where barriers to entry will become relatively higher as the native assets that power them, such as ADA, appreciate in value.