‘One doesn’t have to be a Marxist to be awed by the scale and success of early-20th-century efforts to transform strong-willed human beings into docile employees.’
– Gary Hamel
When John Endler first studied guppies in the streams of Venezuela and Trinidad in the 1970s, he noticed an intriguing pattern: guppies in the pools at the bottom of waterfalls tended to be rather drab, while those in pools further upstream were eye-catchingly gaudy. Endler suspected the likely cause of the difference: while guppies were able to swim upstream past the waterfalls, the voracious guppy-eating pike cichlid could not, so the upper pools were cichlid-free. The drab guppies were camouflaged because they had evolved in a dangerous environment. The brightly-coloured ones lived in Guppy Eden, safely separated from the cichlids by a waterfall, and their colours were simply useful for attracting the attentions of other amorous guppies.
Endler decided to test his hypothesis in a more controlled environment, and filled a large greenhouse with ten guppy pools. Some pools had pebbles on the bottom, other pools were lined with finer gravel. Endler released the dangerous pike cichlids into some of each type of pool, and other pools were stocked either with gentler predators or no predators at all. Within fourteen months, ten generations, the guppy population adapted. In the dangerous pools, only the most boring guppies survived to breed; what is more, the guppy camouflage fitted the lining of the pool, with larger patterning in the pebble-floored pool and smaller patterning in the gravel-floored pool. In the safer pools, it was the brightly-spotted guppies that bred more – female guppies, it seems, have a taste for colourful polka dot males.
Professor Endler’s guppy experiments are a modern classic in evolutionary biology, and a striking example of how a population adapts to a new problem, such as the appearance of a pike cichlid. Not only was the adaptation fast, it was sensitive to context: the right response to a pike cichlid depends on what sort of material the pool is lined with. It was a decentralised process, because no guppy planned the response. And it was driven by failure: some guppies were eaten, while others went on to produce future generations of well-adapted baby guppies.
If this was a straightforward guide to business success and personal growth, this would be the point at which the author would urge you to use the principles of adapting to gain wealth and success working just one hour a day, or to create the next Apple or Google. If only it was that easy.
Adapting is not necessarily something we do. It may well be something that is done to us. We may think of ourselves as Professor Endler, but we’re actually the guppies. No individual guppy adapted, but some guppies avoided being eaten and some did not. This book has so far taken John Endler’s view of things. We’ve seen how policymakers and the leaders of organisations can set up systems that either unleash or suppress adaptive behaviour: carbon taxes promote eco-efficiency; innovation prizes encourage new ideas; Donald Rumsfeld’s thought-policing of the US Army held back the process of adapting in Iraq; ‘too big to fail’ bailouts encouraged banks that were, well, too big. But this closing pair of chapters takes the guppy-eye view, asking how the principles of adapting can be applied in two areas: corporate strategy and personal life.
As the pike cichlid closes in for a meal, it’s little consolation to the polka-dotted guppy that its failure is helping clear space for a thriving population of pebble-coloured nieces and nephews. A struggling entrepreneur is just as unlikely to be comforted by the thought that the failure of her start-up is part of a wealth-generating process of creative destruction.
So let’s first acknowledge a crucial difference: individuals, unlike populations, can succeed without adapting. The guppy population evolved pebble-camouflage through trial and error, but no individual guppy did: each was born either with good-enough camouflage, or not. Similarly, many of the heroes of this book – Reginald Mitchell, Mario Capecchi, H.R. McMaster – are admirable not because they themselves adapted, but because they had the courage to experiment with new ideas in the face of overwhelming pressure to stick with the crowd. In business, if you’re in the right place at the right time and happen to hit the right strategy, you’ll thrive without much need for adapting. The basic story of, say, Amazon, is not of a business that selfconsciously experimented its way to success, but of a business with founders who had the luck or vision to spot the new opportunity of internet retailing and grasp it.
But unlike Amazon, or geniuses like Mitchell or Capecchi, or a pebble-coloured guppy, we don’t all get it right first time. Fortunately we have something that guppies do not: the ability to adapt as we go along. Guppies only get one shot at what colour they are: get it wrong, and they die in the jaws of a pike cichlid, or are unable to attract a mate. Few of our own failures are fatal. Within limits, we can experiment either sequentially or concurrently: we can try being pebble-coloured first and switch to polka-dot if it doesn’t work out, or we can split our time between them.
There are three essential steps to using the principles of adapting in business and everyday life, and they are in essence the Palchinsky principles. First, try new things, expecting that some will fail. Second, make failure survivable: create safe spaces for failure or move forward in small steps. As we saw with banks and cities, the trick here is finding the right scale in which to experiment: significant enough to make a difference, but not such a gamble that you’re ruined if it fails. And third, make sure you know when you’ve failed, or you will never learn. As we shall see in the next chapter, this last one is especially difficult when it comes to adapting in our own lives.
First, this chapter looks at how corporations can become less like an individual guppy and more like a population of guppies: trying things out, running with what works. We’ve already seen one example of how to do this in corporate life: the ‘skunk works’ idea of a Galapagan island of innovation. But there are other approaches, and organisations which have actively adopted Palchinsky-style principles of pluralism, gradual experimentation, and learning from mistakes, with great success. They do not offer the only path to business success, but they suggest what might be possible.
Let me describe a fast-growing company that embodies some of the key principles of adapting – we can call it ‘Difference Machine’ for now. The company operates on several different sites but it is more decentralised than that fact alone would suggest, being organised around small teams of which there are over half a dozen on each site. Within these specialised groups, staff have a lot of autonomy to decide what product features to offer to customers, at what price and with what sort of marketing push. These decisions are taken at a very local level rather than being passed to head office or even the senior managers on site – this allows new ideas to be tried out on a small scale and in response to a particular situation.
More radical still, the teams are self-selecting: a new recruit is placed with a team for four weeks for a trial period, at which point they can stay only if they win the vote of two thirds of their team members. (The mark of a good team is said to be its willingness to defy the advice of the team leader and kick out a new member who isn’t pulling his weight.) The team selection method is used not just for the site offices but for senior managers at head office, too.
Difference Machine has a suitably progressive philosophy of ethical business which helps provide guidance to this decentralised, experimental organisation. Coupled with the radical employee empowerment programme, that mission might make it sound soft-headed about the basic business of making money. Not at all: the company’s CEO explained some years ago in a blog post that ‘we can’t fulfil [our] mission unless we are highly profitable’. Employees are acutely aware of the bottom line. Many employees have stock options, but the profit focus is more local and immediate than that: every four weeks, each team receives a bonus if its profitability over the past month has exceeded a certain threshold. Healthy competition is promoted by a ‘no secrets’ policy of strict transparency: many of the company’s financial statistics are available to employees, and every team knows how every other team is performing – a mechanism that also allows bad ideas to be noticed and nipped in the bud, and good ideas to spread horizontally through the company.
This radical devolution of power and responsibility to frontline employees is working: the company has been a permanent feature on Fortune Magazine’s Top 100 ‘best companies to work for’; sales were $8 billion in 2009 and have been doubling every three years since the company floated on the stock market. The company’s market value is comparable to that of large competitors with ten times as many staff.
What industry might this paragon of corporate innovation inhabit? We might assume it’s one of the brash young software houses, or perhaps a green technology company, something big in genetics, or possibly some hyper-global outsourcing operation. In fact, ‘Difference Machine’ is an alias not for the next Google, but for one of the world’s most boring lines of business: a supermarket, an industry synonymous with dead-end jobs and disempowered staff, where every decision is taken at central office and mediated by a computer and a loyalty card. Difference Machine is actually Whole Foods Market, the high-end, organically-minded and lushly-stocked grocery chain. (The description of many of the management practices comes from Gary Hamel’s recent book, The Future of Management.)
Of course, this kind of business model is not the only way to succeed in the supermarket trade. Far more centralised supermarkets such as Wal-Mart in the US and Tesco in the UK are clearly very profitable – they still experiment but have managed to centralise and automate that experimentation. Yet Whole Foods demonstrates that even in this most regimented of industries it is possible to succeed with a radical, employee-led management model that would not seem out of place in a utopian Silicon Valley start-up.
Whole Foods isn’t unique, either. Almost every management innovation described above applies equally to one of the UK’s least glamorous brands, Timpson. Timpson has several hundred small branches which adorn many British high streets, offering a bric-a-brac of services such as key-cutting, shoe and watch repairs, and engraving. Like Whole Foods Market, Timpson has a ‘no secrets’ policy, sending round a frequent newsletter to all staff explaining how the business is doing and how much money there is in the bank. Like Whole Foods, the staff in an individual shop are in charge of deciding what exactly goes on the shelves and whether to offer deals or promotions – the company’s chairman, John Timpson, calls it ‘upside-down management’. If a boy comes in trying to engrave something for his grandmother, and doesn’t have enough money to pay the usual price, it’s up to the local staff to decide whether to offer him a price he can afford; if a customer has a complaint, the lowliest shop assistant has the authority to spend up to £500 putting it right. Timpson doesn’t have a large complaints department at head office: it doesn’t need one. And the small team of staff at each shop is paid a performance bonus every week based on the performance of that particular team. No wonder they know exactly how the shop is doing when Mr Timpson drops in – and he drops in frequently, because he spends four days a week on the road doing nothing but visiting his shops to talk to staff.
The first thing Timpson does when it buys another business is to rip out the electronic point-of-sale machines (there are always EPOS machines) and replace them with old-fashioned cash registers. ‘EPOS lets people at head office run the business’, explains John Timpson. ‘I don’t want them to run the business.’ EPOS machines empower head office but they make it harder to be flexible and give customers what they need. John Timpson describes one instance where he couldn’t buy half-price happy-hour drinks at a hotel bar, because midway through giving his order, the hour ended and the bar’s computerised sales system refused to allow the half-price offer to be applied. He fumes at the idea of powerless staff telling irritated customers that ‘I can’t put that through the till.’
John Timpson, and John Mackey of Whole Foods, have learned the same lessons as H.R. McMaster in Iraq: the best computer systems in the world cannot substitute for being there, talking about what’s going on and responding at once to subtle situational clues – or in Hayek’s now familiar words, ‘knowledge of the particular circumstance of time and place’. The correct balance between centralised control and decentralised experimentation depends on circumstance: at a nuclear power station, we want engineers to keep an eye on each other but we don’t want them improvising with new ways to run the reactor. Nor do we want to allow a situation where a company such as AIG, with 120,000 employees, can be devastated by a unit employing barely a hundred.
Overall, though, as we saw in chapter 2, more and more companies are decentralising, flattening their hierarchies and paying performance bonuses to more junior staff, and they are doing this because the world is increasingly rewarding those who can quickly adapt to local circumstances. H.R. McMaster criticised the idea that ‘situational understanding could be delivered on a computer screen’; John Timpson might put it more bluntly, but these two very different men with very different responsibilities have reached very similar conclusions.
‘We only have two rules,’ explains John Timpson. ‘One: look the part. Wear your tie, turn up on time, be nice to the customers. Two: put the money in the till.’ The second rule is an intriguing one: with so much autonomy, it is not hard for employees to steal money from the company. It is part of a broader problem: if an organisation grants radical autonomy to its members, how does it guarantee that the members will respect the organisation’s interests rather than simply pursuing their own?
Partly this is a matter of trust. Timpson’s company training manual describes the twenty easiest ways to defraud the company, making it clear that the company understands the risks it is running and trusts its employees anyway – and many people respond to being trusted by becoming more trustworthy. Partly it’s the strong focus on performance: both Timpson and Whole Foods Market monitor performance closely and reward it frequently. But in large part, these systems work because staff keep an eye on each other and have a strong incentive not to tolerate slackers and crooks.
‘It made us pay more attention to the people themselves because our way of running a business only works if you’ve got the right people,’ says John Timpson. And he emphasises that poor performers don’t just damage the company, they damage their colleagues. ‘If somebody is just not interested, just turns up for work, we don’t want them. Nor does anyone else working with them.’ Half of Timpson employees join through a ‘refer a friend’ scheme – in other words, Timpson uses its own employees to recruit the ‘right people’. At Whole Foods Market, remember, new team members have a four-week tryout, after which they need to have won the confidence of two thirds of their colleagues.
Whole Foods and Timpson both use a system of peermonitoring. This makes sense: if power is delegated to the front line of the organisation, then that is where good ideas need to be separated out from bad – and good people, too. It’s the ‘worm’s-eye view’ we saw advocated by Muhammad Yunus. And there’s also a parallel with the whistleblowers of the last chapter: it is the people who work regularly on a particular site or in a particular division who notice that something is amiss. The problem is persuading them to speak up, so it’s not hard to see why both Timpson and Whole Foods Market place such a premium on team performance and measure it, publicise it and reward it every month or even every week. Peer monitoring does not always work, of course; peer groups can turn into self-serving or even corrupt cliques. (It is no wonder that John Timpson spends most of his working life visiting Timpson stores.) But it offers a subtlety and sensitivity that monitoring from corporate HQ simply cannot match.
Peer monitoring can take many forms. At Timpson and Whole Foods it’s about making sure everyone does their bit. But the same approach prevails at Google, where peer monitoring is all about maintaining an atmosphere of intellectual challenge. Eric Schmidt, Google’s chairman and, until recently, its CEO, views his role at Google as mediating a debate and forcing other people to reach decisions, rather than making the decisions himself. (The company, in any case, gave him few of the trappings of authority: on his first day at Google he discovered that his assigned two-desk office had been spotted and colonised by an engineer; Schmidt took the second desk without protest.)
At W.L. Gore, the company that developed Gore-Tex, the CEO was effectively elected by her peers: the board of directors polled Gore ‘associates’ about who in the company – anyone – they were willing to follow. The name that came up was Terri Kelly, and she was duly appointed. Gore associates find their own colleagues and their own projects to pursue, and they must rely on the power of their case rather than the authority of the org chart. One Gore associate comments, ‘If you tell anybody what to do here, they’ll never work for you again.’ John Timpson speaks of ‘upside-down management’, but the practice obviously has its uses well beyond the high street.
Peer monitoring is closely associated with the virtual world: it’s the fundamental building block of Google’s search algorithm (giving weight to how popular a site is with other sites), phenomena like eBay (which relies on buyers and sellers rating each other’s reliability) and Wikipedia (in which anyone can edit anyone else’s articles), and the open-source software movement which has delivered such successes as Firefox and Apache. But as Timpson shows, it’s applicable far behind the cutting edge of crowd-sourced technology.
I witnessed a striking example of peer monitoring on my visit to the Hinkley B nuclear power station. I’d just received a briefing on Hinkley’s safety culture from Peter Higginson, an avuncular physicist from Shropshire who was responsible for the safety of Hinkley’s two massive advanced gas-cooled reactors. The safety culture sounded impressive, and depended heavily on peer monitoring. All significant actions, such as flipping a switch in the reactor control room, were double-checked by a colleague. Every employee – receptionists, security guards and press officers included – took a course on nuclear safety; everyone was responsible for taking care of everyone else. It all sounded great – but also a bit too good to be true.
Then we changed into overalls and toecapped boots to prepare for a visit to the turbine hall. Just as we were about to leave the meeting room, a portly middle-aged lady in a hard hat walked in pushing a trolley laden with sandwiches. She took one look at us and politely but firmly admonished our host that we’d left our shoes in a place where they constituted a tripping hazard, and asked us to move them. Perhaps the incident was unusual, and of course a tripping hazard is a far cry from a failure in the reactor core. Yet it was hard to forget seeing peer monitoring in action: the instant correction of a problem, no matter how small and no matter what the hierarchical relationship might be between head of safety and tea lady.
At Hinkley Point, the key priority is ensuring that the power station operates exactly as planned, without deviation. But at other companies, the challenge is to do something new every day, and nowhere is this truer than at Google.
The company’s CEO Eric Schmidt had a surprise when he walked into Larry Page’s office in 2002. Page is the co-creator of Google and the man who gave his name to the idea at the company’s foundation: its PageRank search algorithm. But Page had something rather different to show Schmidt: a machine he’d built himself which cut off the bindings of books and then scanned their pages into a digital format. Page had been trying to figure out whether it might be possible for Google to scan the world’s books into searchable form. Rather than instructing an intern to rig something up, or commissioning analysis from a consulting firm, he teamed up with Marissa Mayer, a Google vice-president, to see how fast two people could produce an image of a 300-page book. Armed with a plywood frame, a pair of clamps, a metronome and a digital camera, two of Google’s most senior staff tried out the project themselves. (The book went from paper to pixels in forty minutes.)
Larry Page regarded the time he devoted to the project not as something he could do because he was Google’s founder and could do whatever he wanted, but as something to which he was entitled because every engineer at Google had the same deal. Famously, Google has a ‘20 per cent time’ policy: any engineer (and some other employees) is allowed to spend one fifth of his or her time on any project that seems worthwhile. Google News, Google Suggest, Adsense and the social networking site Orkut are all projects that emerged from these personal projects, along with half of all Google’s successful products – and an astonishing portfolio of failures.
Whole Foods Market would have little to gain from letting its employees noodle around on whatever project took their fancy, but Google’s 20 per cent time is a practice made successful by the same basic mechanism that Whole Foods relies on: peer approval. Managers stay out of the way of personal staff projects. It is other engineers who determine which projects gain momentum and which languish: if you can’t persuade your peers to help you with your idea, it will go nowhere. Managers can provide the space for innovation, but it is peers who provide most of the time and energy. More recently, Google has grown so big that Eric Schmidt, Larry Page and Sergey Brin have formalised a process of supporting promising innovations. Even so, the aim is not to stifle more projects but to give extra funding and resources to projects that might otherwise be lost in the noise of 20,000 employees.
It’s hard to imagine two more different companies than the shoe-repair chain Timpson and the internet search giant Google, but look at the similarities in the language: Google wants to maintain a ‘bozo-free zone’, Timpson is insistent about keeping ‘drongos’ out of the business. Bozos are less-than-brilliant engineers; drongos are shop assistants who don’t care about the business and don’t pull their weight. The basic idea is the same: in a company where the selection mechanism is your teammates rather than top-down rules, there is no room for people who don’t play their part.
The 20 per cent time policy isn’t unique to Google: not only is it being widely emulated across Silicon Valley, but it long predates the creation of the Googleplex. A similar deal has been standard practice for half a century at W.L. Gore, where all employees get a half-day each week of ‘dabble time’. Again, we see that while the experimental approach may be perfectly exemplified by the denizens of Silicon Valley, and even more by the online communities they make possible, the basic ideas have been around and successful much longer than the World Wide Web.
A serial innovator such as Google or W.L. Gore knows that if you give smart people some space, you may get a Spitfire, the solution to the longitude problem, the technique for knocking out genes in mice – or Gmail. A few such successes justify a lot of slack time. One example is W.L. Gore’s ‘Elixir’ range of acoustic guitar strings, which now dominate the market. They emerged via a long period of experimentation when a W.L. Gore engineer, Dave Myers, applied the Gore-Tex polymer first to cables on his mountain bike and then to guitar-strings. Gore had no experience in the music industry and Myers had no management approval for what he was doing. He didn’t need it.
The management guru Gary Hamel argues that Google in particular is actively pursuing a Darwinian strategy of pushing out the largest possible range of products – not a single guppy but a greenhouse full of different guppy strategies. Google is quite simply an evolutionary organisation: it began with a search engine, then turned site hits into revenue when it teamed up with AOL and Yahoo, then developed a system of displaying adverts alongside search results. Google then stumbled upon the idea of Adsense, the ability to make adverts relevant to any web page. This discovery was made serendipitously while developing Gmail, and trying to deliver context-sensitive adverts alongside the Gmail inbox, and then expanded into Google apps and other projects. Hamel comments that ‘like an organism favored by genetic good fortune, Google’s success owes much to serendipity’. That is true of many successful companies – John Mackey, the CEO of Whole Foods, calls himself ‘the accidental grocer’ – but Google have elevated it to a guiding principle.
If any company can be said to embrace trying new things in the expectation that many will fail, it is Google. Marissa Mayer, the vice-president who helped Larry Page bodge together the first book scanner, says that 80 per cent of Google’s products will fail – but that doesn’t matter, because people will remember the ones that stick. Fair enough: Google’s image seems to be untarnished by the indifferent performances of Knol, a Google service vaguely similar to Wikipedia which didn’t seem to catch on; or SearchMash, a testbed for alternative Google search products which was labelled ‘Google’s Worst Ever Product’ by one search expert and has now been discontinued. According to the influential TechRepublic website, two of the five worst technology products of 2009 came from Google – and they were major Google products at that, Google Wave and the Android 1.0 operating system for mobile phones. Yet most internet users know and rely on Google’s search, Google Maps and Image search, while many others swear by Gmail, Google Reader and Blogger. As long as the company doesn’t pour too much money into failing products, the few big successes seem to justify the many experiments.
This is fundamental to the way Google does business. Google has established its own equivalent of John Endler’s guppy ponds and is seeing what emerges. The company’s corporate strategy is to have no corporate strategy.
Some years ago, a craft and fabric chain called Jo-Ann Fabrics offered its customers a surprising deal. It wasn’t surprisingly creative or surprisingly generous. In fact, it was surprisingly lame: buy one sewing machine, get a second one at 20 per cent off. Who on earth wants to buy two sewing machines? But the deal was also surprisingly successful. Customers found that the prospect of saving 10 per cent per sewing machine was tempting enough to make it worth having, so they went hunting for friends who might also want to buy a sewing machine. In short, the quirky offer turned out to be an unexpected way to recruit amateur sales staff.
More interesting than the special offer was how it was discovered: Jo-Ann Fabrics was using its website, JoAnn.com, as a laboratory. Different customers would automatically be shown different website designs and different offers, the particular combination chosen at random by a computer. In line with the first two Palchinsky principles, Jo-Ann Fabrics was prepared for many such offers to fail, and could afford for them to do so. The deal for bulk-buyers of sewing machines was one of the unlikely successes that this random process discovered, and the use of randomised experiments on the website more than tripled revenue per visitor.
As Ian Ayres explains in his book Supercrunchers, stories like that of Jo-Ann Fabrics are becoming more and more commonplace. Credit-card providers have long used combinatorial experiments in their junk mail – these experiments turbocharge the randomised trial method we’ve already encountered by layering multiple randomisations on top of each other to generate a very rich set of data. The results are all used to refine the mailshot and to hook more customers. But whereas these experiments once required statistical experts and cutting-edge computer technology, they are now very easy to run online. Anyone can buy two or more advertisements on Google AdWords and see which one works best. (Ian Ayres did just that, which is why his book is called Supercrunchers and not his own favourite, The End of Intuition.) Or for bigger projects, professional help is available to unleash the full power of combinatorial experiments.
Such experiments aren’t limited to the web. Supermarkets can and do randomise their price offerings, their shelf placement, the vouchers they send to customers with loyalty cards, or the design of the advertisements they place in local newspapers. Fast-moving consumer goods companies play with the packaging of key brands. Publishers sometimes offer several different covers to a magazine or a book and see what sells.
Experiments have been going on in corporations behind the scenes for over a century. Thomas Edison may have been known as the Wizard of Menlo Park, but his experimentation hit a systematic, industrial scale in 1887 after he built large laboratories a few miles north in West Orange, New Jersey. He employed thousands of people in an ‘invention factory’ and made sure the storerooms were well stocked and that the physical layout of the laboratories allowed the largest number of experiments in the shortest possible time. He was the father of industrial research. Famous for saying, ‘If I find 10,000 ways something won’t work, I haven’t failed. I am not discouraged, because every wrong attempt discarded is just one more step forward,’ he also commented more directly on the industrialisation of the trial-and-error process: ‘The real measure of success is the number of experiments that can be crowded into twenty-four hours.’
That number can now be dozens, hundreds or even tens of thousands thanks to the introduction of cheap supercomputers and other techniques for systematising experiments. Pharmaceutical companies use ‘combinatorial chemistry’ to search through a colossal range of possible drugs: thousands of different chemical compounds can now be synthesised on the surface of a single silicon chip, or bonded to the surface of polymer beads to allow easy mixing and further synthesis, or synthesised in larger quantities in robot labs without human intervention. The resulting compounds can then be tested in parallel to answer simple but vital questions: Are they toxic? Can the body absorb them? Silicon chip manufacturers design bespoke chips in a virtual environment before testing and refining them experimentally. The faster computers become, the faster new computer chips can be designed and tested. The same process is applied to the aerodynamics of a car, or its safety during a crash. And the fundamental point of all these massively parallel experiments is the same: when a problem reaches a certain level of complexity, formal theory won’t get you nearly as far as an incredibly rapid, systematic process of trial and error.
We saw in chapter 4 that randomised trials make some people queasy in medicine and in foreign aid, and the same is true in business. Some years ago, a consumer-goods company approached Dan Ariely, a marketing professor at Duke and MIT, for advice in running some experiments on its own customers. It was a sharp move: Ariely has since become one of the most celebrated behavioural economists after the success of his book Predictably Irrational. Ariely uses experiments all the time to develop and test ideas in psychology and behavioural economics, such as the hypothesis that ‘free’ isn’t just a price of zero; ‘Buy one get one free!’ feels different to ‘Buy two get both at half price!’, even though the offers are identical. Real-world applications of this insight were something the company might be able to use, and Ariely could use the collaboration to harvest data for his academic research.
All went well at first. An experiment, with multiple websites and various combinations of offers, was due to go live when suddenly some higher-ups at the company began to raise concerns. Their objection was much the same as the long-standing complaint about randomised trials elsewhere: that some customers would miss out on the good stuff. ‘Because we were extending differing offers,’ explains Ariely, ‘some customers might buy a product that was not ideal for them, spend too much money, or get a worse deal overall than others.’ In some ways, the executives’ worries were more valid than those we dismissed in chapter 5. The two key counterarguments – that subjects of trials can be approached for their informed consent, and that trials bring wider social benefits – don’t apply in business. You can’t inform a customer that they’re being charged full price to see how much difference it makes to offer another set of customers a discount. And customers are not necessarily the ones who will benefit from research that is designed to make the business more profitable.
But these concerns could have been addressed very easily. If a retailer is simply testing out whether a discount will pay for itself in increased sales, then there’s a simple way of compensating customers in the full-price arm of the trial: after they have made their purchase decision, give them the discount anyway, either immediately or as a cash refund when the trial is complete.
In the end, the executives decided to revert to a way of doing business they were more comfortable with: they asked Dan Ariely simply to tell them the best marketing technique. This was Archie Cochrane’s ‘God complex’, with Dan Ariely cast in the role of God. But Ariely didn’t think his expert opinion was worth much compared with the insight that would emerge from a proper experiment: ‘Companies pay amazing amounts of money to get answers from consultants with overdeveloped confidence in their own intuition,’ he marvels. The project was canned.
Despite such setbacks, the routine experimentation advocated by Edison is now widely practised. It feels much safer than 20 per cent time or upside-down management; it’s less anarchic, less of a threat to the existing power structure in a company, and less of a threat to the status quo. When experiments become routine, a company such as Wal-Mart or Capital One can crunch the numbers from head office without upsetting the corporate hierarchy. By contrast, creating the space for employees in your medical products division to get together and create the market-leading brand of guitar string is exhilarating in retrospect but profoundly disquieting to most corporations at the time. There is surely a reason why so few companies have actually emulated W.L. Gore over the past half-century. And yet some business scholars wonder if even the approach of a W.L. Gore or a Google is really radical enough to cope with truly disruptive business ideas.
Guppies breed so quickly that John Endler was able to produce guppy evolution within months. When Clayton Christensen of Harvard Business School wanted to understand why some apparently capable companies find themselves wiped out by a sudden shift in the competitive landscape, he looked for the economic equivalent of a greenhouse full of guppies. The disk-drive industry was his first port of call: a market in which upstarts frequently seem to usurp the market leaders. As with John Endler’s guppies, what Christensen discovered points to a much wider truth.
Christensen’s initial explanation for the brief life-span of a disk-drive manufacturer was the ‘technology mudslide’: the pace of technological change is so frenetic that companies frantically scramble to reach the technological summit as the ground keeps slipping away beneath their feet. No wonder one decade’s dominant manufacturer is the next decade’s corporate basket case. But this plausible-sounding theory makes little sense on close inspection. Top disk-drive manufacturers have the cash flow to fund further innovation, and are constantly refining their procedures and responding to a constant stream of feedback from customers. They are farther up the mudslide than new entrants, and win purely technological races against upstarts again and again, whether this is disk-drive manufacturers coming first to market with faster drive speeds and denser storage, or camera makers with the latest, sharpest lenses, or sports-shoe companies with new styles and better-designed soles.
Christensen found that it isn’t cutting-edge technology that tends to undo the market leaders. It is the totally new approach, often with quite primitive technology and invariably of little value to the best customers of the leading industry players. In the late 1970s, leading disk-drive manufacturers were making their products better and better for their main customer base of large corporations and banks with room-sized mainframe computers. To these customers, a new generation of physically smaller drives – with much less storage – was of no interest. But these new drives tapped into a new market for the desk-side computers then being pioneered by the likes of Wang and Hewlett-Packard. Eventually the smaller drives became more technologically advanced and even the mainframe customers began to buy them, and by that stage the traditional manufacturers were hopelessly far behind.
The more familiar example of digital photography offers much the same lessons. The first digital cameras were expensive, performed poorly and offered little storage. They were of little use to either the amateur snapper, who wanted something cheap, or the professional photographer who wanted a sharpness of image that such cameras could not provide. Leading manufacturers of film cameras, which had been the only game in town since the invention of photography, might have worried but would have seen little cause for concern from the market itself.
But early digital cameras appealed to some niche users who wouldn’t otherwise have bothered with a film camera at all: in the late 1990s, for example, I used one to photograph flip charts at corporate meetings so they could be stored on a floppy disk and transcribed later. Neither the price tag nor the poor image quality was a problem: what mattered was that it was easy to get the pictures onto a computer and emailed to an assistant back at head office. These niche markets gave the technology a foothold to improve – and very quickly – until only a few nostalgic holdouts were sticking with film. By then, the photocopier company Canon had a powerful position in the market, and many established names such as Fuji, Kodak, Olympus and Leica were scrambling to catch up in a landscape that had dramatically changed.
The battle between desktop email software and webmail is in some ways even more telling. Back in the 1990s, the desktop program Microsoft Outlook was clearly the superior product: webmail services had limited storage, were cumbersome, and were extremely slow over a dial-up connection. Outlook handled most corporate email and Outlook Express served the still-small home market in a way that most users found far superior to webmail. But webmail did have a niche as a backup account for the web-savvy, or for students who had free internet access and wanted the ability to hop from computer to computer across campus. Only later did connection speeds, storage costs and the sophistication of browsers improve sufficiently to demonstrate the true potential of webmail: it could be used to archive and store every email you ever received; used as a document backup; used as a primary email account, packed with features; and used offline. What is striking is the difficulty Microsoft had with this transition, even though it had bought the leading webmail service, Hotmail, quite early in the game, and even though webmail wasn’t a complicated technology for Microsoft software engineers to master. Yet Hotmail’s features were eclipsed by those of Google’s Gmail.
Disruptive innovations are disruptive precisely because the new technology doesn’t appeal to the traditional customers: it is different and for their purposes, it’s inferior. But for a small niche of new customers the new disruptive product is exactly what is needed. They want smaller, cheaper hard drives, or cameras that produce digital files, or email that you can access on any computer – and they are willing to tolerate the fact that the new product is inferior to the old one along all the traditional dimensions. That foothold in the niche market gives the new technology an opportunity to develop into a true threat to the old way of doing things.
The problem for a market leader in the old technology is not necessarily that it lacks the capacity to innovate, but that it lacks the will. When a disruptive technology appears, it may confound an existing player because the technology itself is so radically different (that was true of digital cameras, but not of webmail or smaller disk drives, which were assembled using off-the-shelf technology). More often, Christensen found, the problem was not technological but psychological and organisational: it is hard for a major organisation to pay much attention to a piddling new idea that makes little money and invites a yawn or a blank stare from important customers. Microsoft bought Hotmail, yes – but it was always going to be hard for Microsoft to pay more attention to Hotmail than to Outlook. Microsoft’s core corporate customers regarded webmail as an irrelevance. Google’s users did not. Google only made web applications, and Gmail was a natural fit.
We already know one possible solution for corporations faced with a potentially disruptive innovation: a skunk works, a sort of corporate version of Lübeck, in which the regular culture and priorities and politics of the old corporation do not apply. Lockheed’s Skunk Works got its name (originally ‘skonk works’) because it began life inside a circus tent pitched next to a foul-smelling plastics factory. Its engineers – who dressed down even in the 1950s – let off steam from their high-pressure, top-secret projects by playing pranks on each other. Lockheed’s basic corporate culture, whatever its strengths and weaknesses, had little influence on how the skunks behaved.
The skunk works can be a quasi-independent division, or even an entirely new organisation. It can focus on core business in a new way, as the original skunk works did, or it can branch into totally new lines of business.
This idea can work far beyond the arms industry. Target, the discount retailer, was a largely separate entity within the more traditional department store chain Dayton Hudson. It adapted more naturally to the big-box, out-of-town format and grew to overshadow its parent – an outcome far preferable to the obvious alternative of Dayton Hudson letting some other upstart overshadow it. Charles Schwab, the stockbroker, decided to get into the internet brokerage business by setting up a completely independent organisation to run a discount share-trading service online. The online organisation grew so quickly that it swallowed up the parent within eighteen months. If Schwab had taken a more cautious approach, the online service might well have been smothered by vested interests, and Schwab itself would likely have been marginalised by some other online player within a couple of years.
Another example is Richard Branson’s Virgin Group. Branson started in music distribution before setting up a record label, Virgin Records. His other projects have included transatlantic airlines, no-frills airlines, mobile phone services, passenger trains, bridal wear, cola, vodka, high-end tourism (including space tourism), radio stations and financial services. Each of these enterprises has been attempted within a separate, standalone company – sometimes several stand-alone companies in different countries. Some of the ideas flopped: Virgin Cola’s main achievement was in provoking a crushing response from Coca-Cola. Others, such as the Virgin Megastore music shops, had years of success before the business model eventually waned and Branson moved on. But the whole structure of the Virgin Group has always been to maintain a high degree of separation between different lines of business: this allows different organisations to focus on their own priorities, and it also allows the failures to fail in isolation.
When the US Army faced the ‘disruptive innovation’ of guerrilla warfare in Vietnam, there was great reluctance to accept that it had changed the nature of the game, making obsolete the Army’s hard-won expertise in industrial warfare. As one senior officer said, ‘I’ll be damned if I permit the United States army, its institutions, its doctrine, and its traditions to be destroyed just to win this lousy war.’ That is exactly how senior executives must feel when their cutting-edge, market-leading business finds itself being disrupted by a foolish-looking new technology. A sufficiently disruptive innovation bypasses almost everybody who matters at a company: the Rolodex full of key customers becomes useless; the old skills are no longer called for; decades of industry experience count for nothing. In short, everyone who counts in a company will lose status if the disruptive innovation catches on inside that company – and whether consciously or unconsciously, they will often make sure that it doesn’t. As a result, the company may find itself in serious trouble. It may even die. And remembering the ‘Who’s excellent now?’ experience of Tom Peters we discovered in chapter 1, such a fate is highly likely to befall many companies – even those praised in this chapter.
But when companies die, does that matter?
Corporations have become such a fixture of life that they seem more permanent to us than they were ever intended to be. A central point of the corporation, as a legal structure, is that it’s supposed to be a safe space in which to fail. Limited liability companies were developed to encourage people to experiment, to innovate, to adapt – safe in the knowledge that if their venture collapsed, it would merely be the abstract legal entity that was ruined, not them personally.
I spent a few years working for an oil company, Shell, which – with an eye on keeping up with potentially disruptive innovations in its field – made various sallies into solar energy, wind farms and other renewable energy technology. Nothing much seems to have come of this yet. Conspiracy theorists may believe that this is because Shell has an evil plan to dominate and disrupt the threat from renewable technologies. I doubt this. If there really is a cost-effective renewable alternative to the eons of energy concentrated into crude oil, it would be very much in Shell’s interests to commercialise it. The explanation is simpler: following Clayton Christensen’s logic, there is simply no reason to expect an oil company to be particularly good at inventing, manufacturing or distributing photovoltaic solar panels. Oil companies are good at very different things: negotiating with African and Middle Eastern governments, complex drilling operations, building and operating refineries and chemical engineering plants, and selling liquid fuels in roadside forecourts. When renewable energy takes off, there is no more reason to expect Shell, Exxon and BP to prosper from it than there is to be surprised that the leading internet company is Google rather than some giant of former technologies such as Texas Instruments or Univac.
Even a skunk works operation is no guarantee of success in the face of disruptive innovations. Skunk works are, by their nature, isolated from the parent company. That gives them latitude to innovate and freedom to fail without bringing down the parent. But that may not be enough. Good ideas may simply stay trapped in the skunk works because the parent company does not understand them. If so, the company may well be doomed.
Fine. There is nothing that says a business must live for ever – and, as we saw in chapter 1, the entire success of the market system is predicated on the fact that they don’t. Suppose that in some start-up, right now, a breakthrough renewable form of energy far cheaper than oil or gas has been discovered and it is about to hit the market. The likes of Shell, Exxon and BP could then, quite plausibly, quickly die. They would not be much missed. It would be inconvenient for their employees and costly for their shareholders, but the employees would in most cases find other uses for their talents. Shareholders accept risks and, if they are wise, put their eggs in more than one basket. Former employees and shareholders alike would, meanwhile, feel the benefits of cheaper and cleaner power just like everyone else.
Corporations exist precisely because we don’t – and shouldn’t – care when abstract legal entities fail. We care about individuals. And it is to individuals, struggling to adapt and learn and grow, that we finally turn.