In an increasingly remote region of cyberspace called USENET, a highly committed group of volunteers works to help people they’ve never met with computer problems. These problems might be simple; some are quite complex. Yet these volunteers spend many hours helping lost cyber-souls find digital salvation. In this particular space, there are more than two million contributors a year, with more than forty thousand making more than 36 annual contributions, and about eight hundred making contributions just about all the time.
The striking thing about this is not that here we have people helping other people, or that they are helping people they’re never going to meet. It is that all this pro bono effort is devoted to a very peculiar end: making Microsoft’s customers happier. For these volunteers operate within Microsoft ‘newsgroups’. They’re not paid by Microsoft; the vast majority are not even recognised by Microsoft. But they all work (and some work quite hard) to make Microsoft richer by solving its customers’ problems.
Microsoft knows this. In one of its two buildings devoted to research, Marc Smith, the head of the Community Technologies Group, leads a team that studies what goes on within these groups. The company has developed elaborate technologies to measure the ‘health’ of these and other virtual communities. Is there enough balance to the contributions? Are the contributors constructive, or smothering? It is constantly reflecting on how to help the communities work better. The problem isn’t easy: there are scores of variables to consider. But the one variable that remains off the table is money. Not because Microsoft isn’t wealthy enough to pay its help, or because the communities aren’t creating things of value to Microsoft. Money is off the table because Smith doesn’t believe that ‘cash exchange for community support’ helps. ‘There are numerous social benefits that amply incentivise contributions to communities,’ he explains. Money isn’t one; indeed, money would probably be harmful.
To people with a certain view of human nature, this story is bizarre. What would motivate people to give up their time to help the richest software company in the world get richer? An increasingly salient view of human behaviour offers an explanation. We’ve been driven to recognise one kind of economy: the quid pro quo (‘qpq’) economy, in which wages are exchanged for work, or CDs for $20 bills. But examples such as Microsoft’s USENET suggest we should consider a different model as well: what the internet visionary and Japanese venture capitalist Joi Ito calls the ‘sharing economy’, in which unrelated individuals, often in remote parts of the world, ‘work’ together to produce private and collective goods. The sharing economy is no less an economy for that. As Microsoft recognises, it, too, produces significant wealth. But the way this wealth is created is different from the ways of the quid pro quo economy. And by understanding the difference, companies like Microsoft benefit from the value produced.
Two recent books help us understand these sharing economies better. They’re the beginning of an important trend in social research that governments, and corporations, would be wise to keep track of. Steven Weber’s The Success of Open Source explores the economy that has produced the ‘open source’ and ‘free software’ movements. Eric von Hippel’s Democratising Innovation looks at an economy involving a much wider range of user-supplied innovation. Both books address similar questions: how come these economies actually function? How do we understand them? How could we improve them? And both make it clear that we all – citizens, businesses and governments – lose if we ignore the wealth these economies create.
The sharing economy in software (called ‘FLOSS’ – Free/Libre/Open Source Software’ – by Rishab Gosh, a leading researcher), produces a puzzle. How is it possible that thousands of people from around the world could work voluntarily to build and improve on massively complex software projects? Weber’s answer is a convincingly subtle account of the ‘political economy’ of open source production. But he begins by framing the question in legal terms. ‘This is a book about property,’ he says, ‘and how it underpins the social organisation of co-operation and production in a digital era.’ Unlike the proprietary software of, say, Microsoft, ‘property in open source is configured fundamentally around the right to distribute, not the right to exclude.’ These rights are effected through software licences, and the suggestion is that these licences matter to FLOSS development.
Little in Weber’s account actually turns on the particulars of FLOSS licences. Indeed, much in it should lead us to wonder whether the particulars matter much at all. For while FLOSS projects are licensed under terms very different from those attaching to proprietary software, they are also licensed under terms that differ from case to case. The two most prominent FLOSS successes – the Apache web server, which is likely to be the software you interact with when you browse the World Wide Web, and the GNU/Linux operating system, which is the FLOSS equivalent to Windows – operate under radically different licences. While the GPL (governing GNU/Linux) preserves the ‘right to distribute’ by restricting the ways in which modifications can be distributed, the Apache licence imposes no substantial restrictions on its users’ freedom to distribute Apache code however they want. So if it’s the licence that is doing the important work, how significant can that work be when these two radically different licences have both produced wildly successful FLOSS projects? Such differences make it difficult to conclude much about whether it is a particular notion of ‘property’ that ‘underpins’ FLOSS development – at least beyond the now undeniable fact that massively complex software projects don’t need proprietary licences to succeed. Research by Gosh suggests that many key developers depend on the assumption that the others involved will share and share alike. But there is more to be done to show just how much the law here really matters.
The core of Weber’s analysis, however, is not the law. Instead, he addresses two fundamental questions: first, what motivates those who contribute to this sharing economy? And, second, given the complexity of large software projects, how are their contributions successfully combined?
It is the first question that puzzles most people. We all have views about human motivation; but as most of us know little about technology, we’re quick to imagine that fancy tools must exist in order to effect this wide ranging co-ordination. In fact, as Weber demonstrates, the second question may be harder to understand than the first, for by carefully unpacking the particulars of individual motivation, it becomes clearer what drives the sharing economy as a whole.
Most developers, Weber observes, code FLOSS projects to fix a problem they are themselves experiencing. (As the anthropologist Eric Raymond puts it, they code to ‘scratch an itch’.) A programmer at a major bank, for example, might have discovered a problem with the way a program handles many simultaneous users. That programmer then ‘opens the hood’ (to borrow again from Raymond) to investigate the problem. If the coder fixes it, he has served his own private need – either for himself, or for the company he works for. That service needs no explanation from the perspective of the qpq economy. The puzzle is, why does he then release his fix to others for free? Why, after expending private resources, would he not demand further compensation first?
The answer has something to do with the individuals concerned, and something to do with the nature of software. It’s ordinarily hard to understand why anyone would give away something of value, but that’s because usually, giving it away means having less yourself. But software in particular, and knowledge in general, is not like food: when I reveal to you how best to install Word on your computer, I don’t lose that ability myself. As Thomas Jefferson put it nearly two hundred years ago, ‘he who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine, receives light without darkening me.’ Software, like ideas, is ‘non-rival’. Your using it doesn’t harm me.
Yet even if I don’t lose anything, what do I gain? It still requires some effort to reveal the information about my fix. What reason do I have to make that effort? Weber points to some familiar reasons – the kudos, for example, that flows to a programmer from others in his community – but he adds another that is often missed. It’s not just that code is non-rival; it’s that code in particular, and (at least some) knowledge in general, is, as Weber calls it, ‘anti-rival’. I am not only not harmed when you share an anti-rival good: I benefit.
The reasons are obvious when we consider a more pedestrian example: the English language, say. Language is an anti-rival good: not only does your speaking English not restrict me, your speaking it benefits me. The more people who speak a language, the more useful that language is, at least to those who speak it. We therefore don’t tax foreigners who learn our language; we encourage them, since we benefit from their use of it just as they do. Language in this sense is a commons – everyone has free access to it – but this commons not only doesn’t produce a ‘tragedy of the commons’; it results, as the Yale law professor Carol Rose puts it, in a ‘comedy of the commons’.
The same is true of at least some FLOSS. A programmer not only wants the code to work well for himself, he wants it to work better for everyone else. For if it does, more people will use it; and the more who use it, the greater its value to its users, including himself. ‘Free-riding’ in such a world is beneficial. Eliminating free-riding, it follows, can sometimes cause harm.
This does not mean that a FLOSS project could flourish with free-riders alone. But FLOSS is a clear example of what Weber calls the ‘twist’ to what is ordinarily assumed about the collective action of large groups: ‘Under conditions of anti-rivalness, as the size of the internet-connected group increases, and there is a heterogeneous distribution of motivations with people who have a high level of interest and some resources to invest, then the large group is more likely, all things being equal, to provide the good than is a small group.’
This is important progress in our understanding of motivation within sharing economies in general. The real difficulty with FLOSS economies, however, is to understand the way contributions are co-ordinated. For not only are the contributors spread across the world, the project to which they are contributing is insanely complex. It is hard enough to produce software if you can, like Bill Gates, command all contributors to toe a certain line. How is it even conceivable when the only commands in the system are those given by programmers to the computers they run?
Weber’s answer is that a combination of norms and technology replaces the ability to issue commands. These are, we could say, the inventions mothered by the necessity of having no Bill Gates. And the most intriguing suggestion is that in many important contexts, these substitutes could work better than what they replace.
The norms are many, and familiar: they concern who leads a project, how decisions get made, ‘technical rationality’ (meaning practical results decide), and transparency in decisions. Most interesting is a commitment to the freedom to ‘fork’, meaning to split a project if its current leader pushes it in a way that many don’t like. This keeps alive the possibility of ‘no-fault divorce’ in all FLOSS projects. It keeps pressure on partners to continue to perform – and in particular, on project leaders to manage the project well.
The less familiar, but perhaps more important tools enabling co-ordination are technological: the design principles that enable code to be developed by many different people at minimal cost. Because extensive co-ordination in FLOSS is difficult, developers build projects, for example, to a disciplined modular design. So long as the interface of a module plugs into the rest, no one need worry about how the internals work. That means there needs to be no negotiation, or agreement, or command structure. Modules simply compete on the basis of the efficiency with which they achieve some particular end.
These technical features are necessary in the absence of anyone with the capacity to command, but this necessity in turn produces benefits. Modularity is necessary because, without leaders, co-ordination costs are high; but a consequence of modularity is that projects can innovate more easily, as new functions can simply be plugged into the most recent version. It is here that Weber rightly links these architectural features of FLOSS to parallel features in the design of the internet: it, too, could not depend on centralised co-ordination; it, too, had to adopt an architecture (called ‘end-to-end’) that permitted simultaneous and unco-ordinated innovation at the edge of the network; and it, too, has taught us that, at least sometimes, decentralised innovation trumps innovation at the core.
While much in Weber’s account will be familiar to anyone concerned with this debate, his book should make this extraordinary phenomenon understandable to a much wider audience. But the real contribution of his intricate analysis is to make a more general point: that just as non-centrally directed Western societies outperformed centrally planned Communist societies, so may production that is not centrally directed outperform centrally planned production. The challenge is to inspire these decentralised economies to produce and to develop norms and techniques that make their production effective.
The software development that Weber describes is alien to most of us. But von Hippel naturalises it, showing us that aliens are everywhere. Something very much like FLOSS is happening in a wide range of commercial contexts. Sharing economies, in other words, are not just for geeks.
‘Users’ of both products and services, von Hippel argues, increasingly ‘innovate for themselves’ in the design and development of these products and services. Windsurfers buy sailboards, for example, and tinker with them; cyclists alter the design of mountain bikes, to make them usable on a wider range of terrain. That part of the story is understandable in terms of the qpq economy. The puzzle is that when innovating, both individual and corporate ‘users’ freely reveal their innovations to others. This ‘free revealing’ is especially puzzling when it involves corporations charged by law to try to make money. Von Hippel writes:
The ‘private investment model’ of innovation assumes that innovation will be supported by private investment if and as innovators can make attractive profits from doing so. In this model, any free revealing or uncompensated ‘spillover’ of proprietary knowledge developed by private investment will reduce the innovator’s profits. It is therefore assumed that innovators will strive to avoid spillovers of innovation-related information. From this perspective . . . free revealing is a major surprise: it seems to make no sense that innovators would intentionally give away information for free that they had invested money to develop.
One might dismiss the phenomenon as a fluke, were it not so widespread. In the areas that von Hippel considers, between 10 and 40 per cent of users adapt the products they use, and freely reveal their innovations. From skateboarders to building engineers, this is not just carelessness on the part of insufficiently self-interested users; it is a critical part of the innovation economy.
Why do they do it? Von Hippel identifies a number of motives. As with FLOSS, the innovator may benefit personally from the diffusion of an innovation because it increases the value of the product (anti-rivalness). The clearest example is academic publication, where ‘free revealing’ (publishing in open access journals) increases the number of citations the academic receives. And as with FLOSS, the innovators may benefit simply because they value the process of innovation. The hobbyist tinkering with his mountain bike may well be producing something of value to himself, to other cyclists and to the company that makes the bike. But he wouldn’t therefore call his tinkering ‘work’, at least in the qpq sense of that term.
So in what contexts might we expect user innovation to happen, or to matter most? The examples that interest me, as a lawyer, are those in which users innovate because if the producers tried the same innovation, they might become legally liable: it is not surprising, for example, that high-performance windsurfing – in which surfers perform acrobatics in mid-air – was developed through innovation by users, rather than by sailboard manufacturers. It’s not difficult to imagine the terror that company lawyers will have felt the first time they saw their products launched high into the air.
But in the more general, and important, cases the product itself has a heterogeneous range of uses. Here, the disadvantage the company faces is the cost of taking account of all possible uses. As von Hippel demonstrates, in such cases the company is likely to innovate less effectively, typically limiting its innovations to marginal improvements in known uses, rather than identifying wholly new ones. (Innovations by lead users of 3M’s products have been estimated to be eight times more valuable to the company than innovations made by the company itself.)
Von Hippel doesn’t argue that ‘free revealing’ is something new. It isn’t. Historians have found similar behaviour in the 19th-century ironwork industry, and there are other early examples as well. But the increased technical capacities of users increase the frequency and importance of such behaviour. This represents an opportunity for efficient innovation that should lead firms, and governments, to remove obstacles to it. Firms out to maximise profits should realise this quickly enough, and once given the hint, competitive businesses will soon follow, if von Hippel is right in his analysis.
The real problem lies with governments, which are too often disciplined by a market that is more interested in private rather than general welfare. The market that controls today’s policymakers is keen to keep them from grasping obvious truths that would add substantially to the general good.
It is here that von Hippel’s work has the greatest potential: by making apparent to a wide range of businesses the benefits of user innovation, he has given them an incentive to recognise obstacles to it. These derive from intellectual property laws that have become impervious to economic reasoning. An industry of lobbyists now works around the world to strengthen those laws, despite increasing scepticism among economists as to their value.
The lobbyists’ view is based on assumptions about innovation that von Hippel’s work flatly contradicts. Intellectual property regulation increasingly locks information in, so increasing the costs of encouraging or enabling user innovation. Though ‘spillover’, as William Baumol has argued, increases economic efficiency, the assumption of increasingly extreme IP regulators is that all spillover is waste. Current regulation benefits those who would reduce competition, but it’s only justified if it helps spur innovation. And if user innovation is important, then this extreme form of IP regulation may well be undermining its own justification: to encourage progress.
Both these books deserve the careful attention of a wide audience, including, especially, governments. It’s time for them to recognise officially what we all know individually: that the economies of human behaviour are more than one; that these other economies too produce extraordinary human wealth; and that they require a different set of norms and techniques. The communities that Microsoft husbands are important and genuine; the wealth they produce for Microsoft is great. The wealth similar communities could produce for society generally is even greater – if only we could learn what these authors preach, and what Microsoft, and many others, already practise.