“WHAT are you on? The ‘fuck Windows’ strategy?” Back in the late 1990s, when Bill Gates was still Microsoft’s boss, any employee who had the temerity to suggest something that could possibly weaken the firm’s flagship operating system was sure to earn his wrath. Even after Steve Ballmer took over from Mr Gates in 2000, that remained the incontestable law at the company’s headquarters in Redmond, in Washington state. Everything Microsoft did had to strengthen Windows, to make it ever more crushingly dominant. Many of the company’s best innovations were killed because of this “strategy tax”, as it was known internally.  

Today the rules are different in Redmond. The new boss who took over last year, Satya Nadella (pictured, centre, with Mr Gates to the left and Mr Ballmer on the right), recoils when he hears the term “strategy tax” and says he now tells his staff simply to “build stuff that people like”. Some of the things he has done would surely have been seen by his two predecessors as “fuck Windows” strategies. Office, the company’s popular suite of word-processing, spreadsheet and other applications, now runs on mobile devices that use competing operating systems. The company is embracing free, “open-source” software, which it used to regard as anathema. At an event in San Francisco last October Mr Nadella showed a slide that read: “Microsoft loves Linux”. In contrast, Mr Ballmer once called the open-source operating system a “cancer”.

As Microsoft celebrates its 40th birthday on April 4th, its executives and shareholders will be looking back wistfully to their company’s lost youth. Born in the year that Captain & Tennille topped the American charts with “Love Will Keep Us Together”, by its 20s it had leapt ahead of the stumbling behemoth of information technology, IBM, only to slow in its 30s and be overtaken by its eternal arch-rival, Apple, a company barely a year younger than itself (see charts).

Mr Nadella’s formula for reinvigorating Microsoft is to move as quickly and as far as possible away from being a Windows-only company to be a global network of giant data centres that provide a broad range of online services for companies and individuals. So far he has done well in beginning to turn round a supertanker of a company, with 123,000 employees and $87 billion in annual revenues.

Microsoft’s transition is being watched by other tech giants, old and new—because they are either going through similarly wrenching changes or worry that they may have to reinvent themselves in the future. Cisco, EMC, Hewlett-Packard, Oracle, IBM and SAP—all these have to navigate the shift from a world where computers lived on people’s desks or in companies’ basements to one in which number-crunching mostly resides in the “cloud”, meaning remote data centres, and in people’s hands, in the form of mobile devices.

Amazon, Apple, Facebook, Google and others like them are constantly on guard for the emergence of a new technology “platform” that allows rival businesses to build applications that lure away their customers. That is why the 11-year-old Facebook recently spent $22 billion on WhatsApp, a fast-growing messaging service, and $2 billion on Oculus VR, a maker of virtual-reality headsets.

That Microsoft is just about Windows has always been somewhat of a simplification. A more accurate image is that the operating system was the core of a tight bundle of programs the firm built over the years. After Windows had achieved dominance of desktop personal computers in the 1980s, it was packaged with Office, which became just as ubiquitous. As high-powered PCs, called servers, became standard equipment in companies’ internal data centres, Microsoft repeated the trick: it created a suite of server software, including e-mail systems, databases and all manner of other business applications, that were tightly integrated with Windows. It was this bundling that helped Microsoft overtake IBM as the most valuable tech company in the mid-1990s.

All the programs can be made to work with rival products, but work best together. That boosted Microsoft’s business (and got it into antitrust troubles)—but has also been a benefit for customers. Yet as computing is moving into the cloud, this model is breaking down. Software is becoming a service delivered over the internet and mostly based on open standards. “It is easier than it has ever been to use programs from multiple vendors,” explains George Gilbert of TechAlpha Partners, a consulting firm. “And the idea of an operating system is much less relevant.”

Ironically, Microsoft was among the first software firms to recognise the potential of cloud computing. But its vision was distorted by its obsession with protecting its existing product line. So from 2006, when it began working on Azure, its cloud-computing service, it was mainly an online replication of Microsoft’s bundle of proprietary programs. That left open a market for other cloud providers—in particular Amazon—to sell businesses raw computing power, on which they could run whatever mix of programs they wished. Likewise, Microsoft was among the first to see the promise of hand-held computers, as smartphones were once called. But it tried to force users to run Windows on them, rather than developing a new operating system better suited to mobiles, as Apple did.

Before becoming chief executive Mr Nadella had been in charge of Azure, and he revived its flagging fortunes by cutting the price of its cloud services and letting customers use their own choice of software. Since assuming the top job, besides making Office available on Apple and Android mobiles, he has moved it towards a “freemium” model: for personal users (though not businesses) the basic version costs nothing, but you pay for extras. Under Mr Nadella the company is showing flexibility in other ways, such as by cutting deals with rivals. For instance, users of the online version of Office, called Office 365, can now save their files on the servers of Box, an enterprise-software firm. “They used to treat us like arch-enemies,” says Aaron Levie, Box’s boss.

Mr Nadella is also seeking to rejuvenate Microsoft’s internal culture, to make it more like the startup it was back in the days of flared trousers and outsized lapels. Far from having them stamped on, Microsofties are now encouraged to test their “wild ideas” on a public website called Garage. It rushes early versions of products out of the door, so customers can try them out, and help spot any bugs. Microsoft has also bought some hot young startups. One such is Mojang, the maker of Minecraft, a popular online game. Another is Acompli, whose e-mail app has been renamed Outlook and is now Microsoft’s e-mail service for Apple and Android mobiles.

Yet Mr Nadella’s biggest achievement so far is that he has given Microsoft a coherent purpose in life, as it enters its fifth decade. He sums it up in two mottos. One is “mobile first, cloud first”: since these are where the growth is going to come from, all new products need to be developed for them. The other is “platforms and productivity”. Windows is still an important platform, and Office still an important set of productivity tools. But Azure is an increasingly significant, and more flexible, platform. And among a new range of productivity applications Microsoft offers is Cortana, an intelligent personal assistant much like Apple’s Siri and Google Now. It already recognises natural language, answers questions and issues reminders; in future it will increasingly anticipate what a user needs, for instance by pulling together the documents for a meeting.

Talent competition

Although industry observers mostly praise what Mr Nadella has done so far, they have several nagging questions. One is about talent. “Microsoft has lost a lot of great people,” explains Marco Iansiti of Harvard Business School. Although it has also gained fresh blood by buying startups, it will have to work hard to keep up in the battle to hire the best programmers. Another problem may be the quality of its software. In the past Microsoft has worked hard to prove to commercial customers that its programs are reliable, says Rob Helm of Directions on Microsoft, a research firm. As it tries to function like a startup, releasing products early and often, they may become less “trustworthy”, to quote another of Microsoft’s mottos.

What should happen to Microsoft’s other businesses is also an open question. Analysts have called for them to be hived off. Bing, the firm’s search engine, seems safe: after losing billions, it will soon be profitable; and it provides important data inputs for Cortana and other services. The Xbox, a gaming console, is recovering from misguided attempts to turn it into a hub for the digital living room. But Nokia’s handset business, which Microsoft bought last year to save the last big maker of Windows-powered phones, now looks like a mistake after the changes in the firm’s strategy.

The biggest question is about Microsoft’s finances. As people and businesses buy ever fewer, and cheaper, PCs, the revenues from Windows are falling. In the quarter to December 31st they were down by 13% on a year earlier. Sales in the commercial cloud business, including Office 365, more than doubled year-on-year, to reach an annual run-rate of $5.5 billion. But this is only a small fraction of total revenues, and the division is still thought to be losing money. Furthermore, says Rick Sherlund of Nomura, an investment bank, it is unclear how newer applications, such as Cortana, will make a living. “The business models have yet to be invented.”

They are unlikely to be as much of a cash cow as Windows and Office, which still generate 44% of revenues and 58% of profits. The gross margin of classic Office is 90%; that of Office 365 is currently only 53%. In other cloud businesses profits will be elusive, given the intense competition from Amazon, Google and increasingly IBM. Citigroup, another bank, recently estimated that Amazon’s margin in the cloud was between -2% and -3%. In other businesses Amazon has shown a high tolerance for losses as it strives for growth.

As other tech firms, young and not so young, seek to learn from Microsoft’s middle-aged decline, the main lesson they will draw is that it was too protective, for too long, of its main franchise, which led it to ignore threats that eventually became unignorable. Mr Gates was most to blame for this; Mr Ballmer was adept at continuing to milk the Windows cash-cow for longer than might have been expected, but he too was slow to respond to the rising threat from the moves to mobile and the cloud. Now Mr Nadella is having to move rapidly to dispel the impression, among customers, analysts and tech talent, that his company’s better days are behind it.