Product design and the myth of faster horses

Product design often encounters a tension between solving observable customer needs (reactive design), and inventing novel experiences without concrete basis in current customer behaviour, but which designers believe will be valuable (proactive design). Both reactive and proactive design can produce successful results. However, the practical question remains: given that most product design teams have finite time and resources, which should be prioritised as the default approach for design practice?

In this article, I explore the differences between the reactive and proactive approaches, note some of their advantages and limitations, and argue that reactive design, grounded in empirical studies of users, is almost always the better choice.

The debate between reactive and proactive design

Reactive design seeks to improve people’s current experiences by listening carefully to what they need and continuously evolving a product. Proactive design seeks to invent new categories of experience based on the designer’s skill and intuition and making large, disjoint leaps in the product. While reactive design results in fairly straightforward and predictable improvements, there is always an element of uncertainty around whether a proactive design will succeed. Reactive and proactive design therefore differ in their aims, the resources they draw upon, and the impact on product.

A comparison of reactive and proactive design.

At first glance, proactive design sounds more impressive and attractive, and in contrast, reactive design sounds dull, incremental and boring. Who wouldn’t want to be inventing new categories of experience based on their craft skill and intuition? This approach puts the designer in the role of a creative hero-saviour, supported by our cultural myth of the lone creative genius.

Reactive design resembles applied research

The differences between reactive and proactive design bear a resemblance to the much older debate around basic versus applied research. In particular, which to fund. Perhaps the oldest example of organised funding for research (at least in the West) is the British Board of Longitude, founded 1714, which disbursed several grants over its more than 100-year history to those seeking solutions to the problem of accurately determining longitude at sea. This was a decidedly applied problem that nonetheless resulted in the production of much basic knowledge in astronomy, as well as applied knowledge in horology.

As government-funded research programmes gradually came to be seen as indispensable tools of nation-building, the approach of funding only solutions to specific and pertinent current problems was derided for its naïve short-sightedness by those who saw the value of fostering innovation for its own sake. By the mid 1800s, spearheaded by William Whewell, the British Association for the Advancement of Science had pioneered the disbursal of scientific grants to works-in-progress towards all sorts of scientific aims, a pattern that gradually spread throughout Europe (but which the elitist Royal Society lagged in adopting, in part because such grants enabled ordinary people and not just the leisure class to participate in science).

In 1979, Cosmic Search magazine gleefully mocked reactive research through the tale of Lord Allsen, set in the early 1800s. In this story (fictitious, as far as I can tell) Allsen imperceptively advocates only for funding research with immediate benefits, brushing aside Hans Christian Oersted’s discovery that a current passed through a wire affects the needle of a nearby compass:

It would be a waste of the taxpayers’ money, he pointed out, to spend even one pence to find out more about what a wire would do to a compass. What was needed, he said, was more practical research like developing longer burning, brighter candles that didn’t need their wicks trimmed as often or breeding faster horses so that messages could be carried between cities more swiftly. Lord Allsen promised to do all he could to support research for better candles and faster horses.

[… nonetheless, the Danish Parliament decides to fund Oersted …]

Oersted extended his research and this led to further work by Faraday, Maxwell, Hertz, Edison and many others so that today we have electric motors, generators, and lights, the telegraph and telephone, radio, television, electronic computers and a host of other devices. It is interesting to speculate that if Lord Allsen and other Lord Allsens had had their way we might enjoy none of these today, the world would have been spared the blight of everything electrical but we might have better candles and horses so fast that, working in relays, a message could reach Chicago from New York in no more than 3 days.

The tale convincingly makes the case against the reactive approach. Indeed I am personally in favour of funding basic research. However, product design is not research; the objective of the former is to make a product for people to use, the objective of the latter is to make knowledge. Because research is an imperfect analogy for design, the arguments for basic research cannot be directly interpreted as arguments for proactive design.

Nonetheless, proactive design and basic research share an underlying logic: that cues from the present environment alone cannot tell you what is useful, interesting or worth pursuing. Conversely, reactive design and applied research both start from the premise that in order to be worth pursuing, an idea must be grounded in cues from the environment. The former looks within, the latter looks without, showcasing the timeless tension between rational and empirical approaches to knowledge-making.

Related debates, which we will not enter here, concern whether human development reflects “continuities or mutations”, and the “push and pull” models of innovation.

In a corporate research lab where product teams work closely with research teams, the different objectives of research and design (i.e., making knowledge versus making products) create an inherent tension both between the teams but also within the aims of the work itself. Managed well this tension can produce good research and products, but managed poorly it can destroy both the research and the product.

A digression: Henry Ford and his “faster horses”

Henry Ford is said to have summarised another case for proactive design thus: “If I had asked my customers what they wanted, they would have said a faster horse”. This argues that people’s needs are latent, and because they can only articulate their needs in terms of their existing experiences, it is impossible to produce genuinely new ideas simply by listening to what they say.

The 1908 Ford Model T: far from unimaginable by the ordinary customer. Source: Wikimedia Commons

I’m not a fan of the “faster horses” quote, for at least four reasons. First, there is no evidence that Henry Ford actually said this. Second, it doesn’t accurately reflect public opinion and technical knowledge at the time. Benz was selling automobiles in the 1890s, and the Oldsmobile was mass produced from the 1900s, many years before the Model T. And while at this time cars were still luxuries inaccessible to most, fast and affordable passenger steam trains had been operating for nearly 100 years. There is therefore no reason that in 1908, the year the Model T was released, a customer interviewed by Ford about their travel needs would be so blinkered as to only ask for faster horses. The third reason I’m not a fan of the “faster horses” metaphor is that it implicitly positions automobiles as being an unambiguously good solution to the problem of public transport. However, with hindsight, it is clear to see that the personal automobile has been a disaster for city planning, the environment, and societal fabric. The final reason is that Ford was a well-documented anti-semite who used his power, influence and the Ford dealership network to publish material precipitating anti-Jewish sentiment throughout America (and later the Weimar republic); viewed through this lens, his disregard for customer input in the design process can be viewed as another manifestation of the idea that some people’s thoughts and opinions are inherently better than others.

All these issues notwithstanding, because it is so widely known, the metaphor of faster horses is useful to communicate the spirit of the conundrum between reactive and proactive design.

The perils of being proactive

What the parables of the faster horses and Lord Allsen, and in general the history of success of basic research would suggest, is that the proactive approach is preferable and produces better results. However, this is not the case, for at least two important reasons.

Most design work is evolutionary

The first major problem with the proactive approach is that most of the time, the everyday work of design is not revolutionary, it is evolutionary. The science historian Thomas Kuhn refers to the activity of “normal science” as “puzzle-solving”, where a well-defined problem is addressed using a well-defined set of rules. Most of the problems designers face are of this kind: How can we help users understand this aspect of the system? Is it clear to the user what to do here? Are we overwhelming the user with information? Could we make this process faster and less error-prone? These questions can only be answered by grounding design decisions in actual user behaviour, i.e., by taking a reactive stance.

Even the poster child of “revolutions” in the technology sphere, the iPhone, has been developed through the gradual accumulation of evolutionary ideas. As Apple analyst John Gruber notes:

That’s how Apple builds its platforms. It’s a slow and steady process of continuous iterative improvement—so slow, in fact, that the process is easy to overlook if you’re observing it in real time. Only in hindsight is it obvious […] Apple’s iterative development process doesn’t just add, it adapts. […] We may never see an iPhone that utterly blows away the prior year’s, but we’ll soon have one that utterly blows away the original iPhone.

The absurdity of calling anything produced by Silicon Valley a “revolution” notwithstanding (to see why it is absurd, one need only compare tech product “revolutions” and their societal impact to those of the Agricultural, Industrial, and French varieties…), even if we take the original iPhone to be an instance of revolution, it is clear that most of the design work done on the iPhone in its now 15-year history is evolutionary and reactive in nature.

Media theorist Steven Johnson posits that innovation usually happens in the realm of the “adjacent possible”. In order for an idea to arrive, the foundations for that idea need to already have been built: before we got the iPhone, we first needed computers. Moreover, in order for an idea to thrive, we need an ecosystem capable of playing to the strengths of that idea: powerful palmtop computers were available well before the iPhone, but only revealed their true potential with the growth of fast mobile internet and the introduction of 3G in the mid 2000s.

Even true revolutions, such as the meteoric rise of China in the decades following the 1980s, need to be reactive to succeed. Unlike Mao’s disastrous (and ironically named) “Great Leap Forward”, his successor Deng Xiaoping approached the problem of modernisation much more cautiously. The international development debate in the 1980s and ’90s raged between the mainstream view of “shock therapy”, that countries should liberalise swiftly and thoroughly, and “gradualism”, that liberalisation ought to happen slowly and in the correct order.

The US treasury and IMF strongly advocated for shock therapy, and the model was adopted in developing countries around the world, from Latin America to South-East Asia. The results were initially promising, but it is now clear to see that shock therapy produced only short-term gains, and many of these countries are not significantly better off than they were 50 years ago. In contrast, Xiaoping took a gradualist stance, and developed the policies of “socialism with Chinese characteristics” through careful study of successful models (in particular, the Japanese economic miracle following World War 2). Moreover, the transition was marked by experimentation with liberalisation through incremental policies (such as dual-track pricing, which sets prices for commodities up to a planned target, beyond which they are exposed to the market price; a compromise between protectionism and liberalism), or pilot tests of market economies within certain cities.

Unlike Western liberal politicians as in the USA and the UK, who are expected to have a concrete vision from the day their election campaign begins, Xiaoping’s economic legacy is his reactive approach, summarised in the Chinese folk saying with which he is most strongly associated: “one must cross the river by feeling the stones”.

As a design tool, and an antidote to the hubristic tech saviour complex of Silicon Valley design practice, I propose what we might call the humility razor. The humility razor is the principle that the solution to any given design problem is usually an evolution of what has come before. It is named by analogy to Occam’s razor (prefer the simpler explanation) or Hanlon’s razor (never attribute to malice that which is adequately explained by stupidity).

Technology and society shape each other in unpredictable ways

The second major problem with the proactive approach is that technology shapes society (Veblen’s technological determinism) and society shapes technology (via Durkheim’s social determinism). The design of technology shapes and persuades users to behave in a certain way, but conversely users adapt, modify, extend, subvert, and repurpose technology for use in ways completely unanticipated by the designers. The combination of these two forces is sometimes referred to, perhaps unimaginatively, as mutual shaping.

Mutual shaping means that there are some aspects of design that can only be understood and determined by observing the interaction between technology and society. Examples abound as to why this must be the case.

The television remote control was not part of the initial television viewing experience, but could only have been developed in time in response to user behaviour and desires (namely, not sitting within arm’s reach of the television) as well as changes in the media landscape (having many television channels meant having more reasons to interact with the television).

The Internet was primarily intended for government use, with restrictions on commercial use until the mid-1990s; the Web (which is the part of the Internet that delivers websites, and which most people think of as being synonymous with the Internet) did not arrive until 30 years after the first government networks. The Web could only have developed in response to the growing adaptive use of the Internet, mostly by academics, as a place to store, browse, and navigate documents, for which the previous paradigms of internet use (centred around the retrieval of static text files) were grossly inadequate.

The rigid conformity of the modern social network (highly limited personalisation, finite well-defined user interactions) exemplified by Facebook, could only have been invented in response to the decade-long experiment in the “cacophony of pimped profiles” that was MySpace. MySpace laid the groundwork for Facebook in at least two ways. First, it normalised the transformation of Internet users (in the West) into immaterial labourers. Second, it showed how excessive personalisation enabled bad actors and also leeched user time and attention from their primary “job” of creating monetisable, advertising-friendly content.

Cases such as the television remote, the Web, and modern social networking demonstrate that the need for observation runs deeper than simply releasing an imperfect version of a product and refining it in response to user feedback; they show that entire features and categories of user needs, impossible to anticipate, can emerge from the interaction between people and technology. Under a mutual shaping regime, product design needs to react to survive.

Product designers: study your customer

Given the issues posed by mutual shaping and the maxim of the humility razor, it should come as no surprise that for most design work, I would advocate for a reactive approach: listening to customers, observing real data about the interaction between people and technology, and trying to solve their problems.

Don’t just build what the customer asks for

However, reactive design can also be done poorly. Asking customers what they want, and then literally building what the customer asks for is a bad version of reactive design. Customers can only articulate their needs in terms of familiar reference points; technologies they already have interacted with or know of. They do not (necessarily) understand the capabilities or constraints of technology. The design researcher’s job is to interpret these articulations (“faster horses”) and discover the underlying need (“efficient transport”). One way to do this is through repeated questioning during user studies; the ‘five whys’ technique attributed to Sakichi Toyoda (of Toyota Motor Corporation) can be effective here. When the customer says they want feature X, ask why. When they respond, ask why again. And so on, until the root cause of the problem is determined.

Another issue with reactive design is that because customers are not familiar with the capabilities of technology, they may not realise that some aspect of their technology use can be improved; they cannot problematise (i.e., view as a “problem”) aspects of their technology use that a trained designer can spot. Here again, relying on what customers say will betray the design researcher. “Customers didn’t say X was a problem” doesn’t necessarily mean that it isn’t worth improving the experience of X.

“Problems customers don’t know they have” or “unknown unknowns” can usually be detected with a careful analysis of behavioural observations. Design researchers are trained to consider the entire user workflow, and spot opportunities where steps can be simplified, time and effort saved, errors reduced, joy and delight introduced, and more value added. These opportunities may not be apparent to the user, but in observing users interact with the system, they become apparent to the researcher. Observing users is key, and is what makes the approach reactive.

Proactive design can still be a valuable tool

The reactive approach does not dismiss the importance of the designer’s craft skill and intuition. These are valuable tools both in determining the root cause of a problem as well as detecting unarticulated opportunities from user observations.

Nor does the reactive approach dismiss out of hand any idea that does not have its basis in empirical observation. Design ideas, even good ones, can arise ex nihilo (although introspection often reveals that an external observation was the source of inspiration). In such cases, the spirit of proactive design can be captured quite well using contemporary research techniques. To test a potentially revolutionary idea, it is not necessary to invest enormous time and effort into building and launching a product. Rather, these ideas can be evaluated by investing as little as possible in communicating the idea effectively, to see if it resonates with customers.

Paper prototyping is an inexpensive way of testing a design idea without investing a lot of engineering effort. Source: Wikimedia Commons

Techniques such as paper prototyping (mocking up an interface out of paper sketches), Wizard of Oz studies (where you pretend that a system works, but behind the scenes a person is manipulating the interface to make it look like it works), and even design fictions (stories written to envision an alternative future) can help customers experience and react to an idea. If the idea is revealed through such studies to be less brilliant than the inventors think (as is often the case), then the crisis that would have arisen had they plunged headlong into building it has been inexpensively averted. In this way, it is possible to incorporate the best of both reactive and proactive approaches into design practice.


Reactive design emphasises direct responses to observable user needs; proactive design draws on the inspiration and intuition of the design practitioner. We have seen how, despite the lustre of proactive design, it is flawed because most daily design work is evolutionary, and also because technology and society shape each other in unanticipated ways.

Reactive and proactive approaches are both useful implements in a designer’s toolkit. The challenge is combining them in practice to best solve the customer’s problem. A reactive, empirically-grounded method is effective for most design work, and lightweight, low-investment methods such as paper prototyping can help test ideas proactively.

Notes and references

… British Association for the Advancement of Science had pioneered the disbursal of scientific grants … see Snyder, Laura J. The philosophical breakfast club: four remarkable friends who transformed science and changed the world. Random House Digital, Inc., 2011.

… tension between rational and empirical approaches … see, e.g. Markie, Peter and M. Folescu, “Rationalism vs. Empiricism”, The Stanford Encyclopedia of Philosophy (Fall 2021 Edition), Edward N. Zalta (ed.),

An earlier version of this blog post had presented this as a tension between “Aristotelian” and empirical approaches, which was based on a mischaracterisation of Aristotle, common since the middle ages. See, e.g., Cushing, J. (1998). Aristotle and Francis Bacon. In Philosophical Concepts in Physics: The Historical Relation between Philosophy and Scientific Theories (pp. 15-28). Cambridge: Cambridge University Press. doi:10.1017/CBO9781139171106.004

… activity of “normal science” as “puzzle-solving” … see Kuhn, Thomas S. “The Structure of Scientific Revolutions. Chicago (University of Chicago Press) 1962.” (1962).

… human development reflects “continuities or mutations” … see Lewis Mumford (1946). Garden Cities and the Metropolis: A Reply. The Journal of Land & Public Utility Economics, 22(1), 66–69. doi:10.2307/3159217. The historian and architecture critic Lewis Mumford published this in reply to an article by Lloyd Rodwin. Mumford disagreed with Rodwin’s statement that “thinking, however imaginative, must reflect continuities, not mutations, if it is to find practical expression”. Hat tip to this Quote Investigator article on the Henry Ford quote.

… the “push and pull” models of innovation … Di Stefano, Giada, Alfonso Gambardella, and Gianmario Verona. “Technology push and demand pull perspectives in innovation studies: Current findings and future research directions.” Research policy 41, no. 8 (2012): 1283-1295.

… international development debate in the 1980s and ’90s … Stiglitz, Joseph E. Globalization and Its Discontents. New York: W.W. Norton, 2002

… “adjacent possible” … Johnson, Steven. Where good ideas come from: The natural history of innovation. Penguin, 2011.

… society shapes technology (via Durkheim’s social determinism) … researchers have proposed different models for how society determines technology, including social shaping of technology (SST), social construction of technology (SCOT), and actor-network theory (ANT).

… the “cacophony of pimped profiles” that was MySpace … Gehl, Robert W. “Real (software) abstractions: on the rise of Facebook and the fall of MySpace.” Social Text 30, no. 2 (2012): 99-119.


4 thoughts on “Product design and the myth of faster horses

  1. This was a really cool write up Advait!
    I agree with most of what you said. Reactive design sometimes feels/looks like Kaizen.
    Do you think that there is a limiting factor to reactive design in the sense that incremental changes are exactly that. SMALL changes to improve the overall efficiency of the system/usability. Innovation doesnt happen here (not radical innovation at least).
    Then there is also how a designer asks and gets responses from their users. i.e: being able to identify that the customer didnt actually want a faster horse, they wanted to arrive at their destination in lesser time. And the nature of these question will sometimes define whether the solution can be achieved reactively vs proactively.

    Looking forward to hearing what you think. Great essay, insightful and thought provoking.

    1. Hey Visal! Firstly, it’s great to hear from you, and I hope you’re doing well! Thanks for reading and your thoughtful comment.

      You are right to draw comparisons to Kaizen (the folks at Toyota really did come up with a lot of great stuff). The Kaizen philosophy adds a time contingency to design (i.e., that it must be done *continuously*), which was not previously widespread in the software industry but which seems to be increasingly important, given the meteoric rise of subscription SaaS and CI/CD practices.

      I think there are two ways in which radical innovation can happen through reactive design. The first is that while each individual change might not be radical, the cumulative benefits of many small changes can (this is the point Gruber was making about the iPhone). Second, sometimes the observation of user needs does spawn an idea that is a radical departure from the technology context in which those observations were made (I guess this was what I was trying to illustrate with the examples of the television remote, the Web, and modern social networking).

      You’re also right to observe that the question (whether it’s a design question or a research question) should determine the methodology applied to answer it. For example, “does design A make users faster at completing their tasks than design B” can’t be answered through interviews; you need a controlled experiment. But methods that rely on gathering empirical data are inherently reactive. Perhaps these can be viewed as different stages in a cyclical process: proactive design occurs during the ideation phase, and these ideas can then be tested and refined reactively. I do believe that both are important tools for designers. The reactive approach seems undervalued to me relative to the proactive approach, in the world of professional product design. The emphasis seems the wrong way around, and the strong advocacy for reactive design in this article is intended to offset that 🙂

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s