Technological change is soft, not hard
Our failure to see this is scuppering our policy response
One of the sentiments you hear most often in digital transformation is that the work is more about people than it is about technology.
People say things like: ‘digital transformation is about transformation first, and digital second’. And a similar point is reflected in Tom Loosemore’s widely-used definition of digital work as:
Applying the culture, practices, processes & technologies of the internet-era to respond to people’s raised expectations.
As Tom says, it’s not a coincidence that technology comes last in the list.
For people who work in digital transformation, it can be easy to think this idea now goes without saying. And at the level of individual organisational transformation, I do think good progress has been made recognising that you can’t treat digital work as a technology project.
But when it comes to the biggest transformation of all, the society-wide transformation of our political and social institutions, I can’t help but think we still haven’t taken this insight on board. We still treat technological change as if it’s mostly about hard technology and not about soft culture, and this is scuppering our public policy response.
To unpack this claim and see why it matters, it’s useful to start with a reminder of where that ‘people over tech’ idea comes from.
When people say digital ‘transformation is about people not tech’, they’re not just saying that organisational change is all about culture. They’re saying something deeper about the nature of technological change itself. Or at least they should be.
We know from economic history that when a new technology comes along it’s not the spread of the technology that has the impact; what has the impact is the spread of new behaviours that make the most of the new technology’s potential.
The classic example is the deployment of electricity in manufacturing in the late 19th and early 20th centuries. It wasn’t the electric motor that boosted productivity, it was the leap to new factory layouts that were made possible by the electric motor, and that unlocked its productive potential.
This is one example of a pattern that recurs predictably in technological progress; when a new technology comes along, we tend to start by just swapping it for the old one, putting the metaphorical motor on the patch of ground that used to hold a steam engine. Only later do we reimagine things to unlock the new tech’s potential — i.e. only later do we change our culture, practices, and processes.[1]
With computers we’ve done this all over again. For example, we took physical sheets of paper and letters and we made digital versions of these things in PDFs and Emails, which is a bit like leaving prison and keeping your handcuffs on. Forty years on, we’re only just starting to reimagine how we collaborate online, freeing ourselves from those long-unnecessary limits of the physical realm.
It’s this basic characteristic of technological change that explains why it’s all about people, not tech. So we might say that the ‘people not tech’ insight runs deep into the essence of what technological progress really is. It goes all the way down.[2]
Five flavours of error
So how does our misconception of technological change as hard, not soft, distort our public policy response? Here are five flavours of error that feel symptomatic of this underlying misconception.
1. We’re too impatient — and too passive
We tend to get impatient when it comes to the effects of technological change. Or, to be more precise, we tend to think the impact of a new technology will start more quickly than it does and then we also underestimate how dramatic the impact will be when it finally comes.
When electric motors were invented, it took ages for new practices of production to emerge. But when those practices did finally take hold, with the assembly line, they changed the world much more than anyone had anticipated, not least because no-one had imagined those new practices at all; they were unknown unknowns.
This human step in the process of technological change — the new mindsets that are required to unlock the potential of new tech — adds a lag to the process that we don’t tend to see. It’s as if a new technological discovery is really just a deep breath in, followed by a pause while we discover and spread new practices of production, before the winds of change blow out.
None of this is news; all this history is well known. And yet we’re still surprised by this lag today. The whole debate about stalling productivity and secular stagnation in the context of the digital revolution, for example, is really similar to a debate about stagnant productivity around the 1900s, when we waited impatiently for the upsides of electricity to be felt.[3]
The issue with this mindset from a public policy view isn’t just that we’re impatient; it’s also that it encourages a kind of passivity in our relationship to technology. We sit around waiting for the computers to kick in, when really the computers are waiting for us.
2. Technological change isn’t fast, it’s surprisingly slow
A related flavour of error is that we pay too much attention to the fast qualities of technological change and too little attention to the slow qualities. We’re distracted by the white water on the surface and we miss the more powerful currents underneath.
We get excited about iPhones and reusable rockets and this feeds our sense that technology is a fast phenomenon. But by far the most interesting thing about technological change isn’t the fastness of new tech but the slowness of its adoption.
We see this play out almost every day in a technology-saturated society. Why the hell is my bank still asking me to send them two copies of a bill as ID or to sign a piece of physical paper? Why did Blockbuster or Maplins or HMV go bust even though it was pretty clear what they needed to do?
The failure of organisations to adopt new practices of production, despite the potential of those practices, is one of the most fascinating things about the digital revolution. It’s like putting a kid in a room with a biscuit and then watching them not eat the biscuit. Even if the kid eventually eats the biscuit, the main question you’d want to ask is: why didn’t they eat it sooner?
The answer isn’t really that organisations don’t know that they should be using new technologies differently. After all, lots of other companies are already doing it, and there are endless books and management consultancies waiting to explain what needs to be done.
The issue is partly change management. Organisations can’t get from here to there because of all the stickiness of mindsets and habitual behaviours.
But there’s also a lack of what I’d call knowhow. Even if an organisation knows in theory how to make good use of new tech, and even if they’re up for it, they don’t possess that knowledge in a craft-like practical sense. They can’t apply the theory.
People sometimes liken this to riding a bike. You can read lots of books about bikes, but you won’t learn to ride one until you practice. And indeed this is one reason that iterative methods matter a lot in digital work. It’s not enough to know the theory of good digital work. Your organisation needs to build knowhow.
Another useful metaphor is to say that digital transformation is like trying to quit an addictive and harmful behaviour. You know what you need to do. But it’s not until you more deeply accept the need to change — until you really know it, deep down — that you will apply the sustained effort required to reset your habits. And of course lots of companies go out of business before they reach this moment of acceptance, just like people sadly die from harmful habits.
So it’s these slow and sticky qualities of technological change that are especially interesting and consequential and, in public policy, massively underplayed.
3. The future is easy to predict
Another thing we get wrong about technological change is that we talk about the future as if it’s highly unpredictable, associating technology with the dark art of futurology or even science fiction.
This is another example of us over-stating the importance of technology/science vs. culture/behaviour, in this case by overstating the importance of invention vs. application.
By far the main thing that’s going to happen over the next 50 years under the heading of technological change is the wider diffusion of ways of working to make better use of technologies we already have.
That’s not to say there won’t be big changes to our lives. The adoption of these new practices will unlock vast reservoirs of latent potential in today’s technologies and this will change the world.
But this process isn’t unpredictable in the way new inventions are unpredictable. It’s not about flying cars. It’s about taking practices that border on boring — mindsets like iterative working, or institutional forms like a peer-to-peer digital marketplace, or technologies like video calling — and applying them to the pre-digital parts of the economy and civil society (which in 2022 is still the majority of organisations).
In the Industrial Revolution, it took decades for the factory model to spread to complex products like shoes, where production was hard to automate. In a sense, it was easy to see that new practices of production would eventually be used to reimagine cobbling — you could have guessed in 1840 that at some point shoes would come to be made more like textiles were made already. It was just a matter of time, but quite a lot of time.
The same thing happened with managerial capitalism, when a new practice of production, this time based around administrative planning and the integration of small firms into large corporations, spread across the economy. It started with the railways, spread into infrastructure like the telegraph and shipping, then into finance, and then into wholesale, integrated manufacturing, and retail. And again this process took decades to play through.
All of which makes it kind of magical to live midway through a process of diffusion; it’s like we can see the future. We know that the digital practices now used at the vanguard of production will, in the next fifty years, be applied in a really insistent way to sectors where they have so far not been used very well. So we can get a good sense of where we’re headed just by working through these potential applications.
To my mind, this makes it all the more inexcusable that our public policy response to the digital revolution is still, even in 2022, so bafflingly timid and slow. It’s like it’s 1840 again and sewage is piling up in the streets and we’re arguing about tweaks to the horse tax to fund an expansion of night soil men.
4. We need sociologists as well as engineers
The fourth mistake that comes from treating technological change as hard, not soft, is that we overemphasise some academic/professional disciplines and underemphasise others.
If technological change is largely about people, not tech, then we need to draw heavily on disciplines that help us understand individual and group human behaviour and the dynamics of institutions.
If you hear the word technology you tend to think of scientists and engineers. But if you hear the words ‘technological change’ you should really think of sociologists, political economists, and organisational psychologists.
This matters for public policy because when governments make announcements about supporting technological change, they’re almost always talking about things like graphene research centres, or test centres for AI cars, or spaceports, or STEM.
These are all good things and they help to progress the frontier of discovery. But an equally important issue when it comes to technological progress — and particularly its transmission into living standards — is how we understand and accelerate the human/social process of diffusion, especially to make sure diffusion is broad and inclusive. And this is the domain of social science much more than it’s the domain of STEM.
Why do human systems take so long to adapt? How can you create institutions to accelerate the diffusion of new behaviours? Why don’t small businesses like corner shops use the latest payment, pricing, and stock management systems, even when it would make them more money? These are questions of sociology, political economy, or organisational psychology.[4]
So our public policy response needs to put far more emphasis on social disciplines, focusing on how we speed up and broaden the diffusion of today’s frontier of practice.
5. We need to get better at solving soft problems
More generally, our focus on tech as a hard thing means that public policy misses or downplays a whole bunch of technology-related problems that are what we might think of as soft.
I’m thinking of problems where the substance of the issue, and the constraints that hinder progress, are soft tendons like culture and habitual behaviours, rather than things that submit easily to economic analysis, or hard things like physical infrastructure.
An example would be how we ease society’s transition away from legacy technologies like PDFs or email. Or how we get academics to write better so that new ideas spread faster. Or how we encourage those corner shops to try out new technologies sooner.
We don’t explore soft technology issues like these enough in public policy; in a sense, it feels odd even to think of them as legitimate terrain for policy. But again, a lot of technological progress is dependent on us making headway on soft and sticky problems like these. And so we need a much fuller and more sophisticated strand of technology policy equipped to walk in this terrain.
Those are just five examples of errors that flow from our tendency to treat technological change as hard, not soft. There will be many more. The implications of the digital revolution are now so pervasive that no area of public policy will escape the fallout if we fundamentally misconceive the nature of technological change.
All of which is to say: we’ve still got a long way to go. We’ve made good progress maturing how we do digital transformation at the level of individual organisations. But when it comes to the biggest transformation of all, the reform of our social institutional settlement, we’re only just starting out.
This is the latest post in a year-long series exploring how we govern the future. To read along, follow me on Medium here. Or support this project for £3 a month (and get a free book) on Substack here.
For the big story behind all of this, from Victorian sewers to digital dragons, buy my book End State.
Footnotes
- As a side note, it’s interesting that this also gives imagination a central role in technological/social progress, much more so than is commonly accepted in public policy.
- I’ve stolen this turn of phrase from Roberto Unger after an Unger-binge in the last few weeks
- This point about the productivity puzzle is obviously a hornets nest and there are well-argued views on all sides of the debate. I’m just saying I find most sympathy with the ‘calm down and give it some time’ school of thought.
- The economist W. Brian Arthur deals with this challenge by expanding the definition of technology to include human systems. He says technology, broadly defined, isn’t just made of wires or lines or code; it’s also made of people and behaviour. And so institutions are technologies just like steam engines are technologies. If we take this approach, I suspect many of the top ten most important technologies in history would fall into the category of ‘soft tech’ (institutions) rather than ‘hard tech’ (machines).