Julian Sanchez header image 2

photos by Lara Shipley

Two Kinds of Innovation

August 17th, 2009 · 10 Comments

A commenter on the other day’s patent post makes an important point:

I also thought that even if an idea is obvious, it won’t be implemented until someone realizes that they have a need for it. I guess this falls under the question of whether there’s prior art, but it would also be a reason why obvious ideas might not be developed for a long time.

Let’s unpack this a bit. While there’s obviously not a sharp boundary here, we can speak loosely of two kinds of innovation that are hugely important to economic growth and general welfare improvements. Let’s call the one patents are meant to protect and incentivize technological innovation—building a better mousetrap or a faster microchip, devising new and better means to serve a particular need. But there’s also entrepreneurial innovation, which is centrally about perceiving or anticipating unmet needs, even if there’s no abstract or theoretical mystery about how to solve them. Maybe you observe that your neighborhood laundromat is overcrowded, that there’s often lots of bored people waiting for a free machine, and you get it in your head that the area could use another one—maybe one with pinball machines or movies playing. Maybe you discover some great Ecuadorian bands and become convinced their unique style will catch on in the U.S. market with a little push. Maybe you think growing demand for novelty justifies installing a newly invented printing technology in your T-shirt plant, so you can get designs based on the meme-of-the-day to market more quickly and cheaply.  The dichotomy I have in mind tracks Joseph Schumpeter‘s distinction between (technological) “invention” and (economic) “innovation” somewhat, though I’m probably using “entrepreneurial innovation” rather more broadly than Schumpeter would have, and I think I’ve seen Mike Masnick invoke a similar split. Probably the simplest cases—opening familiar type of business in an underserved area or bringing an existing product to a new market—don’t even fall within most people’s colloquial notion of “innovation.”  But I think the term fits: There’s still a species of creative leap involved in seeing the opportunity to profitably introduce something novel in a particular context, if not new on the face of the earth. As Learned Hand observed with respect to technological invention, after all “the test of invention is the originality of the discovery, and discovery depends upon the mental act of conceiving the new combination, for substantially every invention is only a combination.”

So to the extent these are distinct categories, or at least family-resemblance clusters, what are the differences? Rolling with the theme from the previous post, one is that technological innovation more often involves the application of abstract or theoretical knowledge, whereas entrepreneurial innovation is more likely to involve what Hayek called “knowledge of the particular circumstances of time and place.” (Tangent: Hayek tended to conflate this with what he called “tacit knowledge,” the sort of inarticulate knowledge embedded in practices rather than explicit, which Wittgenstein talked about as a “knowing how to go on.” These are importantly distinct concepts, and I always found it odd that Hayek lumped them together.)   That means we’re generally going to have different criteria for what counts as success within each type.

For a technological innovation, the question is just whether the gadget or process you’ve come up with actually accomplishes what it’s supposed to—which at least in the paradigm case is just going to depend on the laws of physics. You might invent something that accomplishes something nobody wants to do, or which there are already other better/faster/cheaper ways of getting done, in which case maybe you’d have a crappy or not very useful invention, but an invention nonetheless: If it works, it works. An entrepreneurial innovation is going to be much more context dependent: It’s a response to a particular constellation of factor prices, wage levels, consumer demand, and so on. A corollary of this is that there’s often no way of seeing whether an entrepreneurial innovation does work other than trying it.

As a result, there are a variety of ways in which it’s less susceptible to mechanical copying, though it certainly may inspire imitation.  For one, acting on an entrepreneurial insight seems much more likely to change whether it makes sense for others to do the same. If my innovative epiphany is that my neighborhood could support a comic book shop, and the corner of Florida and 7th would be a great place to open one… the corner of Florida and 7th is probably stops being a very good place for someone else to open a comic book shop unless they’re sure they can do a much better job of it. In more general terms, we can say that just as the evolutionary “environment” for social organisms consists largely of social organisms, the background against which an idea does or doesn’t count as an entrepreneurial innovation is the aggregate of all other entrepreneurial innovations. Finally, in sharp contrast to your basic mass produced widget, it’s a lot harder to disentangle the general idea from a particular instance.  It’s not enough to have the insight that a successful comic shop or café could be run on this corner—someone actually has to run it, which (again with the tacit knowledge!) is more about having a peculiar knack or set of non-transferable skills than about having the right blueprint.

Maybe the central distinction is where the innovative leap happens. A purely technical innovation could involve taking a very clear given problem or goal that everyone understands well enough—we want faster computers and thinner screens, and a cure for cancer would be nice while you’re at it—and then coming up with a clever means to that end. We may rest assured that when someone discovers the cure for cancer, whatever else it looks like, it will not be “obvious,” because a hell of a lot of smart people have been banging their heads on the problem for a long time.  In a case of purely entrepreneurial innovation, the tricky and innovative part is seeing that there’s a problem—an inefficiency to correct or a demand to meet—from which the solution often follows straightforwardly. A normally competent person would know how to do it, the innovator’s stroke is to see that it’s worth doing at all.

Largely because of these many differences, we have radically different policy responses to these two types of innovation. Genuine technological innovators, the conventional wisdom goes, must enjoy at least a temporary monopoly over their inventions, or they’ll lack the ability to recoup their costs or the incentive to invest time and energy in innovating in the first place. Entrepreneurial innovation, by contrast, is seen as an evil to be fought with all the force of law: The benefit of acting on the idea first or best is all the advantage the entrepreneur needs or ought to be allowed. I hope nobody will find me excessively boring if I say that, in these very abstract terms, the conventional wisdom here strikes me as pretty much correct.

Now of course, the problem—perhaps the problem at the root of a lot of complaints about contemporary patent law—is that in the real world it’s seldom quite this simple, and the two forms of innovation are combined. So consider those big clips for bags of potato chips you can get in most supermarkets—originally covered by, among others, US Patent 4394791: “Closure clamp for food bags.” It’s clearly an invention of a sort, pretty similar to clothespins and hairclips and binder clips, but with some important physical differences owing to its different function. Still, forget “skilled in the art,” I feel about as confident as one can be that if you went back to 1975, gathered a couple of perfectly ordinary dudes with no special training in food-container-closure-design, and asked them to brainstorm devices that would help to seal in the potatoey freshness of their chips, you’d get something that looked an awful lot like a chip clip. The design solution is nothing very special or clever; the insight is that there’s a problem there that can be solved with a product at all. And where did that come from? Well, maybe people used to use clothespins, but the spread of clothes dryers meant that a lot of households didn’t have any lying about anymore. Maybe, for any number of reasons, bagged potato chips got more popular. Maybe it became more viable because of the availability of cheap plastic materials, or the spread of chain supermarkets as a venue, or the overall growth in disposable income to spend on things that would previously have seemed frivolous. Now consider the patent for the improved chip clip, which purports to be superior to the old sort that use a torsion spring, because the exposed ends of the spring give a “cluttered appearance,” can snag on things, and sometimes get bent and leave the clip misaligned.  Would an ordinarily skilled tinkerer tasked with improving the torsion clip design come up with this particular improvement? Frankly, I don’t know. But it presumably takes a certain sort of eye to perceive that these little inconveniences are, indeed, problems crying out for a better design.

Much of this cross-applies to the patent discussed back in the original post. The mere notion of storing a document’s formatting instructions apart from the text content is surely too abstract an idea to be patentable in itself. But given that general idea, the method for achieving it—for turning a document with embedded formatting codes into a segregated one—is trivial. If there’s anything that looks like a genuine creative leap here, it’s in imagining that users might find this structure useful because it makes it easier to swap the formatting on the same text and display it in different ways—but that sounds an awful lot more like one of our entrepreneurial innovations.

Apologists for the status quo patent regime—a group that seems to consist entirely of patent attorneys—love to assert that hindsight is 20/20, and that anyone who regards a patented invention as obvious must just be a victim of this phenomenon. The claim has a certain surface plausibility, but it’s usually awfully hard to keep a straight face when contemplating particular examples, and the grounds for insisting that your retrospective intuition of obviousness is mistaken almost always amounts to pointing out that the first guy to file the patent was—they’ll keep saying it until you grasp this thorny concept—the first guy to file the patent. One (ahem) obvious problem with this line of argument is that someone can do something that is genuinely innovative, can really have had an insight that nobody else at the time had, without having done any non-obvious technical innovating.  The heart of the innovation might well have been entrepreneurial in character. And again, our general practice is not to protect entrepreneurial innovators from competition, but if anything to encourage it as a spur to still further innovations.

Since the real world is messier than these simple categories, it’s probably too glib to say that courts or examiners just need to classify each element of a patent claim as one or another type—there will be elements of both in many cases, and which predominates may depend on the vagaries of the specific process by which the inventor hit on it more than the intrinsic properties of the innovation.  Still, it seems like it would at least be a step forward for the patent system to pay more attention to the possibility that the novelty of a claim may have more to do with entrepreneurial than technical insight. Insofar as its the specific properties mentioned above that justify the difference in policy treatment, patentability criteria more closely attuned to those properties may help tweak the system to generate higher-quality patents.

Tags: Economics · Law



10 responses so far ↓

  • 1 RickRussellTX // Aug 17, 2009 at 5:35 pm

    Once again, you’ve put into words a concept that’s been banging around in my head for some time. If only I had applied for a patent!

    Seriously, there seems to be considerable scholarship devoted to “developing innovation” — in particular, methods to hothouse and grow the kind of entrepreneurial service innovation that you discuss and keep it within the organization.

    When I was getting my MBA, I tried to approach the parallel question: what is required to foster *technological* (scientific, engineering, medical) innovation? In fact, there seems to be less scholarship devoted to this issue, I think because it’s such a thorny problem. Clearly there is money, and facilities, and getting the best minds, and… precious little detail on how to fit them together, except to throw them in a lab and see what happens.

  • 2 JustinOpinion // Aug 17, 2009 at 5:47 pm

    There is something you’re leaving out from your (otherwise insightful) analysis. A patent is in some sense a deal between the public and the inventor: “We’ll grant you a short-term monopoly (with the force of government behind it), if in return you tell us all the gory details of your invention.” The idea is that the public is getting something out of the patent beyond just giving businesses an “incentive” to produce things. I’m not sure about the extent to which this kind of reasoning, in practice, applies to patent applications. But I submit that it was, and should remain, a key justification for the social burden of the patent system.

    In this context, the “obviousness” criteria is derived from the social contract. Only if the patent application provides useful information to the public (which would otherwise have been lost to a corporate trade-secrets vault) is it really a good ‘deal’ for the public. A patent for a generic chip-clip, for instance, doesn’t disclose anything that someone else couldn’t have easily figured out without the patent. It’s obvious. The details of a manufacturing process to make an alloy that has superior strength and flexure properties (and could be used as a spring in a chip-clip) is very valuable to the public. In exchange for granting a monopoly on the production of that alloy, the public (and other companies) get to see the details and (eventually) build upon the discoveries.

    It seems to me that many ‘bad patents’ could be avoided if this social contract were kept in mind (vagueness and obviousness are immediate disqualifies). Returning to your technological vs. entrepreneurial divide, it’s clear that disclosing details of a technology is valuable. Whereas an entrepreneurial innovation is, as you point out, totally obvious–the value to society is only in the doing.

  • 3 LP // Aug 17, 2009 at 5:56 pm

    “A purely technical innovation could involve taking a very clear given problem or goal that everyone understands well enough—we want faster computers and thinner screens, and a cure for cancer would be nice while you’re at it—and then coming up with a clever means to that end.”

    I don’t disagree with the categories you’ve distinguished here, but would like to observe that the above is untrue. An awful lot of really unexpected invention happens at the basic research level, which is less solution-oriented or goal-oriented than “anomaly-oriented” – here is an odd fact that doesn’t seem to fit into what we’d predict would happen, so let’s start pulling at it, and pulling and pulling, et voila: a cure for cancer. Or whatever. Obviousness can get tricky here – some particular insight might be completely novel, because no one has ever studied that particular gene pathway, enzyme interaction, etc., but still look completely obvious in retrospect, aside from the fact that no one ever looked there before. In other words, the number of smart people banging their heads against a fundamental scientific problem is not necessarily correlated with the lack of obviousness of the solution, once found.

  • 4 LP // Aug 17, 2009 at 5:58 pm

    JustinOpinion: Good point. Many people are mistakenly debating patents vs. public domain, when the real alternatives are patent vs. trade secret (non-optimal, especially for new drugs and other dangerous things).

  • 5 Julian Sanchez // Aug 17, 2009 at 6:03 pm

    Hmm, while certainly it’s true that the implicit social contract is an important part of understanding the “obviousness” criterion, I want to question the suggestion that entrepreneurial innovation is generally “obvious.” In some cases, of course, it will necessarily be—you can market a new alloy while keeping the production process secret, but you can hardly keep secret the fact that you’re marketing it. In other cases, though, the reason you don’t want to set up the quid-pro-quo is that it’s actually the very opposite of obvious: Maybe someone finally runs a successful café in a location where others had failed. Maybe it’s not one innovation, but a process of a hundred little changes, made continuously, that keeps it appealing to the local community. You can’t have the contract there because, in fact, the innovation is really built into the person’s particular skills and business sense and can’t easily be taught to others or turned into a formula. (In those cases, of course, protection is clearly not necessary either, and I doubt many patent claims are occasioned by innovation of this sort.)

  • 6 Julian Sanchez // Aug 17, 2009 at 6:12 pm

    Yes, well, the “could” there was something of a deliberate hedge. But I don’t know that the kind of case you mention presents special difficulty: If the innovation is a matter of exhausting a huge search space, maybe even in a pretty mechanical manner, surely you want to grant the protection and provide incentive to do the search. But that seems like it’s probably just a question of being sure to index obviousness to the search space rather than serving up the result and asking whether it’s obvious that it has the desired properties.

  • 7 southpaw // Aug 17, 2009 at 10:32 pm

    I think what we might be butting up against is the desire to make innovation an act of personal commission when invention so often involves elements of chance and serendipity and path dependency. The heroic conception we want to attach to the innovator doesn’t always fit.

    There are tens of thousands of brilliant biological researchers out there, but only Fleming had the wit to cause penicillium notatum to float in the window.

    That’s not to say the state shouldn’t incentivize people to develop the skills that could allow serendipity to reward them. It’s just that the whole scheme bears an uncomfortable similarity to the state lotto much of the time.

  • 8 LP // Aug 18, 2009 at 12:55 am


    Looks like the Supreme Court agrees with your ‘number of smart people who’ve unsuccessfully banged their heads on this problem’ criterion (Graham, KSR). I’m fascinated by the EU’s ‘would/could’ method of determining obviousness, as well as by the UK’s warning to consider the perspective of a ‘skilled but unimaginative’ person.

  • 9 sam // Aug 18, 2009 at 6:24 am

    I also thought that even if an idea is obvious, it won’t be implemented until someone realizes that they have a need for it.

    Dunno if it was all that obvious, but Les Paul said he didn’t set out to be a grand innovator — he just wanted and needed an electrified guitar to make a certain kind of music, and there wasn’t anything around that satisfied the need. So he set about to fill his own need.

  • 10 Stones Cry Out - If they keep silent… » Things Heard: e81v2 // Aug 18, 2009 at 8:57 am

    […] An interesting post on innovation. […]