What is 21 Inc. trying to do?

Ben Thompson today published an article on Bitcoin and the start-up company 21 Inc. I then somehow found myself in a twitter debate about Bitcoin and 21 Inc.'s business model with the notorious Walt French.

It turns out that it's very hard to express your ideas on Twitter. So below I've outlined my thesis:

  • Bitcoin is valuable for internet-of-things devices, allowing devices to algorithmically exchange property rights.
  • Ideally, IOT devices should be convenient to use with a "plug and go" design, just as normal devices exist today.
  • One solution to achieve this "plug and go" is for IOT devices to have access to bitcoins somehow embedded into its design, without the need for the user to "set it up" and link it to their own Bitcoin address. But how?
  • 21 Inc. aims to solve this issue by embedding IOT devices with its bitcoin-mining chips.
  • 21 Inc. will provide bitcoins on demand to the IOT devices that use its chips and in return, 21 Inc. will keep all bitcoins mined from those devices.
  • This works because the amount of bitcoins each IOT device needs is small and with high frequency. On the other hand, the amount of bitcoins each IOT device earns is large but with very low frequency.
  • Similar to a bank that borrows short-term and lends long-term, 21 Inc. will profit from the difference in bitcoins mined and consumed over the life of all the devices in its portfolio. This is also analogous to how insurance works by pooling risk.
  • In the end, 21 Inc. will be providing users a way to "pre-pay" for bitcoins that their IOT device will use over its lifetime without the hassle of set-up.

I think that's pretty much 21 Inc.'s plan at the moment.

Bitcoin has value from being a platform, not money

Like other fiat currencies, Bitcoin is a network-backed currency: It derives all its value from the network effect — also known as Metcalfe's Law. (Monetary economists call this the hot potato effect.)

Many people think that Bitcoin is flawed as a currency because it is not backed by something beyond itself. They are wrong. Bitcoin is flawed as a currency, but it is flawed because of its fixed supply, inherent in the protocol. Contrary to what many Bitcoin proponents argue, fixed supply is not a desirable characteristic of money because it would make monetary policy impossible. And remember, monetary policy is very important because it solves the problem of price stickiness and money illusion. In particular, bad monetary policy causes recessions (and disastrous monetary policy causes depressions).

Thus any currency that does not allow for monetary policy is inferior to the current model and is not likely to be mass adopted in the long-run — unless, of course, it finds some other solution to money illusion. But so far, I have not heard of any solution to money illusion other than monetary policy. To become widespread, Bitcoin needs a central bank that can adjust the money supply to prevent recessions. (Actually, a central bank is not needed if money supply adjustments are built right into the protocol, which would be very interesting, but I digress.)

Now, this does not prevent Bitcoin from being a platform for currencies. I have discussed before that it is likely that central banks will create their own crypto-currencies, and the easiest way for them to do so is to simply fork the existing Bitcoin protocol. Central banks would then be able to change the supply of money by leveraging the network effect (read the post to get a full description of how).

And if Bitcoin is forked to create official central bank money, bitcoins mined today will still have value in the future. But that value will be determined by central banks in the future, and means that today's value comes from Bitcoin being a platform, not Bitcoin being money.

Disrupting from Above by Extending Functionality (The iPhone Case Example)


  • It is difficult to analyze the iPhone as a low-end or high-end disruptor because of its diverse functionality. It just does so many things.
  • As functionality extends, the price of each function falls, making the iPhone a high-end disruptor to consumers valuing all of those functions. (Think of a college student using multiple apps.)
  • On the other hand, as the performance of each function improves, the iPhone becomes a low-end disruptor to consumers valuing one specialized function. (Think of a film director using the camera.)
  • This same characteristic applied to PC's decades ago, and is likely to repeat with the Apple Watch.


Recently I published a simple framework to explain disruption from above. The framework depends on measuring innovation using affordability instead of performance (the measure classically used in Christensen low-end disruption). In that post, I claimed that the iPhone can be viewed as a high-end disruptor to iPods. However, to be a high-end disruptor, the iPhone must have become cheaper than iPods over time, and many people have pointed out that they have not. Something similar could be said about PC's disrupting typewriters from above.

I stand by my claim, however, because the price of the iPhone has fallen — dramatically, in fact. Let me explain.

Remember, when thinking about disruptions, we have to think in the realm of the job-to-be-done. So the first question we need to ask is: What job-to-be-done did the iPhone disrupt? Well it disrupted many, actually. Many.

In relation to the iPod, the iPhone disrupted the job of personal mobile music. Think back to 2007 when the iPhone was introduced. If you bought an iPhone ($1000) simply as a replacement for an iPod ($400), it would have been a very expensive iPod. However, if you bought an iPhone as a replacement for an iPod and a feature phone, the cost for the iPod function would only be $500 (the other $500 allocated to the phone function). That's still more expensive than the stand-alone iPod cost, but that's just the start. Now add in the iPhone's function as a digital camera and the cost of each function drops even further to $333. Now add in the functions of a GPS, mobile browser, email device, calendar/scheduler, alarm clock and all of the millions of apps in the App Store. As functionality extends, the price of each function falls drastically. In fact, for most current iPhone users, the cost of their iPod is $0, since they would have likely still purchased their iPhone without the music app. And this applies to nearly all of the iPhone's other functions. The iPhone is the ultimate high-end disruptor.

But the iPhone is also a low-end disruptor if looked at from the right point of view. Imagine a user that needs only one specialized function of an iPhone. For example, a documentary film-team might purchase an iPhone simply for its video camera function. From the view of professional film-makers, the iPhone is a very low-performing, inexpensive video recorder. But the video function is much better than it was in 2007, and it is getting better every day. Just last week, the TV series Modern Family aired an episode shot entirely with the iPhone. This is classic low-end disruption. 

It is thus difficult to analyze the iPhone as a low-end and high-end disruptor because of its diverse functionality — it just does so many things. As functionality extends, the price of each function falls. At the same time, the performance of each function is constantly improving. The iPhone has so much functionality and is constantly improving that it disrupts everything, from multiple angles. The same thing can be said of PC's decades ago, and explains why PC's were eventually preferred over typewriters for even the most cost-conscious consumers. And it is likely that we will see a similar phenomenon play out with the Apple Watch.

Extending functionality doesn't mean better performing functions, necessarily, it means more functions to solve more jobs. This is why I do not subscribe the theory of new-market disruption, which I think is just another misclassification like Ben Thomson's obsoletion. Instead, you can think of the term new-market disruption as describing a product that disrupts several jobs-to-be-done, from both the low- and high-end, by extending functionality and improving performance. 

The iPhone example reinforces the fact that low-end disruption and high-end disruption are idealized special cases of a general disruption phenomenon which occurs by innovating along both performance and affordability. It also reinforces the need for a synthesized framework.

If you liked this article, please share, follow me on twitter or subscribe to RSS.

High-End Disruption: Using Affordability to Measure Innovation


  • The key insight to understanding high-end disruption is that innovation improves products either by increasing performance or by increasing affordability.
  • While low-end disruption occurs through improvements in performance, high-end disruption occurs through improvements in affordability.
  • While incumbents chase profit margins in low-end disruption, incumbents chase market volume in high-end disruption. In both cases, incumbents are acting rationally by improving their profitability through sustaining innovations — at least at first.
  • Christensen's low-end framework can be modified to visualize high-end disruption by changing the vertical axis from performance to affordability.
  • In low-end disruption, consumers are price-choosers and performance-takers. While in high-end disruption, consumers are performance-choosers and price-takers. Low- and high-end disruption are idealized special cases of a general phenomenon of disruptive innovation. 

So, in short, the master plan is:
  · Build sports car.
  · Use that money to build an affordable car.
  · Use that money to build an even more affordable car.
Don't tell anyone.
— Elon Musk

The purpose of this article is to present a case for disruptive innovation from high-end products. This article is intended for readers who are already familiar with the classic theory of low-end disruption as described by Clay Christensen.

Intuitively, we know that industries can be disrupted with (what are initially) high-end products. For example, iPods were a high-end disruptor to CD players, and iPhones were then a high-end disruptor to iPods (among many other things). And there are countless other examples (see the end of this post). In these cases, incumbents were displaced by disrupting entrants, despite the strength, market power and resources of the incumbent. However, the classical theory of disruptive innovation developed by Clay Christensen only explains disruptive innovation from the view of low-performing products. We need a way to approach disruptive innovation from the view of high-performing products.

To understand high-end disruption, we must first take note of a stylized fact: Innovation improves a product by either improving performance or by improving affordability. There is, of course, a direct relationship between product performance and price, but innovation does not imply better performance, necessarily. Instead, innovation can also manifest in the form of a lower price. This distinction is the key insight to understanding high-end disruption: Whereas low-end disruptors enter the market with a low-performing, low-price product and then work to improve performance, high-end disruptors enter the market with a high-performing, high-price product and then work to improve price. In each case, the entrant eventually takes over the entire market with a high-performing, low-price product.

We can adjust the Christensen low-end disruption framework to visualize high-end disruption simply by changing the measure of innovation from performance to affordability. If you are already familiar with the Christensen low-end disruption framework, everything is straightforward once you make this change.

To start, we can simply think of affordability as the inverse of price. That is:

Affordability = 1/Price

Similar to the performance needs distribution, consumers will fall into the affordability needs distribution.

In the Christensen low-end case, the left-tail of the performance needs distribution comprises consumers that value affordability and are willing to pay for it with low performance. In the high-end case, however, the left-tail of the affordability needs distribution comprises consumers that value high performance and are willing to pay for it with low affordability. In both cases, the entrants are competing against non-consumption — although under-consumption may be a better term for the high-end consumers.

In Christensen's low-end theory, low-end disruptors enter the market with a product that is low-performing but highly affordable. Low-end entrants initially under-serve the performance needs of the majority of the market, but are good enough for the low-end which values affordability. Established incumbents are under-serving the low-end of the market because it is simply not attractive in terms of profit margin. For the entrant, however, the profit margin of the low-end market is attractive. Entering the market with an already highly affordable product, low-end entrants move up the market by increasing performance through sustaining innovations. As low-end entrants improve performance, their product becomes good enough for more and more of the market, eventually over-taking the entire market. As Christensen describes, established incumbents are unable or unwilling to adopt the new disruptive technology because they are incentivized against competing with low-end entrants, and instead move further up the market towards higher performance in order to attain higher profit margins. Incumbents facing low-end disruption are seemingly acting entirely rational by ceding the lower-end of the market, since achieving higher profit margins through sustaining innovation boosts profitability — at least at first, and until it doesn't, when it is too late to pivot to the disruptive innovation.

On the other hand, high-end disruptors enter the market with a product that is low in affordability but high-performing. High-end entrants initially under-serve the affordability needs of the majority of the market, but are cheap enough for the high-end which values performance. Often times, this niche high-end market is comprised of hobbyists and enthusiasts  although not always. Established incumbents are under-serving the high-end of the market because it is simply not attractive in terms of volume. For the entrant, however, the potential volume of the high-end market is attractive. Entering the market with an already high-performing product, high-end entrants move down the market by increasing affordability through sustaining innovations. As high-end entrants improve affordability, their product becomes cheap enough for more and more of the market, eventually over-taking the entire market. Established incumbents are unable or unwilling to adopt the new disruptive technology because they are incentivized against competing with high-end entrants, and instead move down the market towards higher affordability in order to attain higher sales volume. Incumbents facing high-end disruption are seemingly acting entirely rational by ceding the higher-end of the market, since chasing sales volume through sustaining innovation boosts profitability — at least at first, and until it doesn't, when it is too late to pivot to the disruptive innovation.

It is important to note that both low-end and high-end disruptors enter the market by serving consumers that are under-consuming given their preference for performance and affordability. Whether starting from the low- or high-end, the disruptor eventually takes over the market with a product that is both better performing and more affordable. High-end disruption could thus also be described as ‘low-affordability disruption’.

In low-end disruption, consumers are price-choosers and performance-takers: they choose their desired price level and receive the corresponding performance. In high-end disruption, consumers are performance-choosers and price-takers: they choose their desired level of performance and pay the corresponding price. This distinction between low- and high-end disruption is similar to the distinction between Cournot competition (quantity-choosers/price-takers) and Bertrand competition (price-choosers/quantity-takers) in oligopoly theory. Readers familiar with oligopoly theory will know that both Cournot competition and Bertrand competition are useful for understanding different types of markets in the real world. Bertrand competition is useful for understanding Pepsi and Coca Cola, for example, while Cournot competition is better for understanding cartels such as OPEC. Similarly, low-end disruption and high-end disruption are useful for understanding different types of innovations for different jobs-to-be-done.

The famous key phrase for thinking about low-end disruption is good enough, while the key phrase for thinking about high-end disruption is cheap enough. To reiterate: Established incumbents that are being disrupted from the low-end chase profit margins, while incumbents being disrupted from the high-end chase sales volumes. The same incentives that encourage incumbents to move up the market in search of higher profit margins in low-end disruption also encourage incumbents to move down the market in search of larger sales volumes in high-end disruption. In each case, incumbents are (initially) increasing their profitability, until it is too late. And in each case, the entrant is competing against non- or under-consumption in the consumer needs distribution.

Let me thus emphasize that low-end disruption and high-end disruption are idealized cases of a general phenomenon of disruption, where price and performance both interact within consumer preferences. That is to say, we need to synthesize the theories of low-end and high-end disruption. More to come on that in the future.

Christensen famously dismissed the iPhone as a non-disruptor since it was not starting from the low-end. I have recently noticed that Horace Dediu dismisses Tesla’s strategy in the auto market for similar reasons. Of course, looking through the lens of affordability, we can see that both the iPhone and Tesla are high-end disruptors. (The iPhone case is a little tricky  maybe I will explain in a future post. Update: here it is.) This mistake is actually unfortunate because Christensen’s low-end theory never provided a case against high-end disruption, but only provided a framework that explained low-end disruption. Measuring innovation using affordability instead of performance, however, explains high-end disruption very well. I have graphed several cases of high-end disruption at the end of this post. You will notice that high-end disruption is actually quite common. The best example of high-end disruption is the general-purpose computer — viewed as an entire group, including mainframes, PC's and smartphones — which has disrupted many industries from the high-end on the back of Moore's Law. Indeed, the importance of Moore's Law is due to the phenomenon of high-end disruption. An example of high-end disruption outside the realm of transistors is the light bulb (with electricity), which disrupted candles from the high-end. A modern example is Chipotle, which is disrupting McDonald's from the high-end. (Yes, Chipotle has become more affordable over the years — you must remember to account for inflation and income growth.)

Ben Thomson distinguishes between disruption and obsoletion, and also makes the case that disruption does not apply to personal consumer markets because people have infinite wants — but that’s not right either. Infinite wants simply means that there are infinite jobs-to-be-done, whereas disruption theory only exists in the realm of one job-to-be-done at a time. I think that what Thomson classifies as obsoletion is simply the phenomenon of high-end disruption.

To finish, below I have graphed case examples of high-end disruption for several jobs-to-be-done. If you liked this article, please share, follow me on twitter or subscribe to RSS.

Will Bitcoins Ever Become Money? A Path to Decentralized Central Banking

So, will bitcoins ever become money? It’s tricky… I would like to discuss a scenario where bitcoins (or something like it) would be money, why central banks would want to create their own crypto-currency protocols, and how monetary policy would be conducted in such a world.

Imagine that a private entity with significant resources and credibility implements a crypto-currency (like Bitcoin) that it backs with real, hard, good old-fashioned central bank currency. So imagine Google creates a protocol similar to Bitcoin called gCoin, and publishes the protocol publicly so that anyone can use it. In particular, anyone is able to mine gCoins by contributing computational power to the network. At the same time, Google says that it will buy unlimited gCoins in exchange for dollars, say at a price of 1 for 1, and that it will also sell gCoins that it has already mined for the same price. Thus gCoin would initially derive its value from Google's promise to exchange gCoins for real dollars at any time.

First off: Why would Google do this? Who knows! Google is very well known for building useful services that it isn’t yet sure how to monetize. And gCoin is useful: Like Bitcoin, using gCoin to buy and sell goods on the internet is faster and much cheaper than using a credit card. A simple way to monetize gCoin would be for Google to provide a ‘wallet’ service that helps facilitate the use of gCoin, which you could just think of as a chequing account for gCoins. You would “deposit” your gCoins with Google, use their online service to keep track of them and transact them, and Google would charge you a monthly fee just like a regular bank. Google would also have a reserve amount of gCoins that it initially mined before publicly releasing the protocol, which would be very valuable if the use of gCoin became widespread. I’m sure there are other and better ways to monetize crypto-currencies that we can’t even imagine yet — we don’t need to discuss this.

Now, who would use gCoins? Only someone who had faith in their value. If people believed Google would always convert their gCoins into dollars, they would gladly accept gCoin as payment. But as you see, we are in the same situation as we were with the gold-backed bank notes in the 19th century: What if there is a bank run on Google? And if other organizations like Amazon, Facebook, Apple (and PayPal?) have their own digital currencies, a bank run on just one of them could cause bank runs on the others — the same financial contagion effect that paper currencies experienced. Privately issued crypto-currencies are susceptible to the same problems as privately issued bank notes.

You can thus imagine that, just like what happened with paper currencies, central banks will eventually step in to create their own crypto-currency protocols and forbid the use of any others. For simplicity, let’s call the central bank crypto-currency protocol BitDollar. Of course, these BitDollars would always be redeemable in regular dollars by the central bank, at least at first.

The exchange rate between BitDollars and regular dollars does not necessarily have to be 1 for 1. In fact, the central bank could adjust this exchange rate in order to conduct monetary policy, although there is a better, more natural way to do monetary policy — more on that in a minute. Assuming the exchange rate is fixed, say 1 for 1, the distinction between dollars and BitDollars no longer exists. Just think of dollars as an abstraction which can be manifested in traditional paper form and now digital form. And like original gold-backed currency, we eventually won't even need regular dollars to back BitDollars, as long as the network effects of BitDollars exist to justify their value  though this transition will happen gradually.

Indeed, a similar type of network effect is what most Bitcoin startups today are betting will give bitcoins their value in the future, and they are building services that allow customers to easily transact regular bitcoins instead of building their own protocols. You can think of these startups as Bitcoin banks: They provide banking services for bitcoins. Currently these banks operate with 100% reserve ratios because they record your ownership of bitcoins on the official Bitcoin blockchain. We will have to wait to see how the value of bitcoins plays out and what banking business model emerges, but it is clear to me that Bitcoin banks and the Bitcoin network will be just as prone to bank runs as Google's gCoins would be, or traditional banks were before the days of central banking. More simply: Bitcoin needs a central bank. And BitDollar would be the answer to that.

But what type of protocol would BitDollar be? Well certainly, the central bank needs the ability to conduct monetary policy, so it would need a protocol where it could easily adjust the money supply at will. It would also need the ability to be the lender of last resort, which means access to an unlimited supply of BitDollars.

The simplest way for a central bank to create its own crypto-currency is for it to fork the Bitcoin protocol into a new protocol that is unchanged in every way except that, going forward, the central bank would set and adjust the block mining reward at its discretion. (Remember, the block reward is arbitrary in Bitcoin and other crypto-currency protocols, and can change to anything that the network agrees upon going forward.) In this world, the central bank can conduct monetary policy by changing the block reward, which in turn changes the future supply of base money. In particular, increasing the block reward corresponds to loosening monetary policy, and decreasing the block reward corresponds to tightening monetary policy. A more intuitive way to think about this is that increasing the block reward decreases transaction costs, while decreasing the block reward increases transaction costs — for every single transaction. For central banking, this is the key to understanding crypto-currencies. And like paper currency, the central bank's crypto-currency would be both decentralized (in transaction) and cenralized (in supply).

Why would people start using the BitDollar protocol and not the original Bitcoin protocol? Again, it’s simply a matter of network effects, and the fact that BitDollar would be legal tender under the law, including the requirement to pay taxes with them. The answer to this question is the same answer to: Why do people US dollar bills instead of notes printed by private banks? It's tautological: People use US dollar bills because people use US dollar bills.

We still need to resolve how the central bank can access an unlimited supply of BitDollars so that it can continue to act as a lender of last resort. The central bank can create unlimited BitDollars for itself by leveraging the the network effect — let me explain. Remember, BitDollars can only be created by solving a block in the blockchain (i.e. processing transactions) and, as we have discussed, the reward for solving a block is now set by the central bank. But also remember, the transaction history in the blockchain is impossible to change, so the central bank can only change the reward for future blocks. So to create BitDollars for itself, the central bank would announce two things to the network, both at its discretion: (1) It announces a future transaction block, and (2) It announces a one-time reward amount for that block. Once that transaction block comes to pass, the central bank will solve the block and announce the updated official blockchain to the network, walking away with the reward. Anyone else that tries to solve that block will not be recorded as such on anyone else's copy of the blockchain because everyone else will be using the official blockchain released by the central bank. Again, why would people using the protocol simply agree to this? Because…networks. The central bank said so. In fact, the central bank doesn't even need to solve the block itself for this process work, simply by also providing an additional transaction reward for whoever processes the block.

Of course, with the ammunition of unlimited BitDollars, there is no need for the central bank to change how it conducts monetary policy by using the block reward method I described previously. But I think that adjusting the block reward is a much more natural way to change future money supply and conduct monetary policy because it affects every transaction in the economy — it is very broad based.

A smooth transition from paper-currencies to crypto-currencies will require central banks to understand the importance of crypto-currencies early on, as they develop. If central banks wait too long, there will be risks of bank runs and financial instability from privately issued crypto-currencies. It is important that central banks recognize this and respond accordingly.

If you liked this article, please share, follow me on twitter or subscribe to RSS.


This is Part 4 of a discussion on the economic consequences of Bitcoin. You can access the remaining parts here:

1.  A Friendly Introduction for Economists to the Bitcoin Protocol
2.  A Monetarist View of Money and Bitcoin
3.  A Brief History of Paper Currency and Central Banking
4.  Will Bitcoins Ever Become Money? A Path to Decentralized Central Banking