Computing, fast and slow

SOME ingredients are missing: the ping-pong table, cheap furniture, and inappropriate T-shirts. But otherwise this could be a shared workspace for internet startups: people sit around long tables and in front of large screens; others lounge on bright orange couches; the walls are full of sticky notes saying things like “I’m happy” or “be yourself”—the products of a brainstorming session.

However, this office in a building in central London belongs to IBM, an information-technology giant long known for its buttoned-down culture and blue business suits. The new “interactive experience lab” is one of 12 such places around the world where teams of employees from IBM and its customers jointly think up new online services and apps. Such projects, the firm hopes, will help it grow again.

IBM’s revenue has declined, year-on-year, for ten straight quarters. Its recent third-quarter figures were particularly disappointing: sales were $22.4 billion and earnings per share $3.68, both well below analysts’ expectations. One reason was that IBM had decided to pile on the bad news: it also scrapped its long-held goal of reaching earnings per share of $20 for 2015. When the firm reports fourth-quarter earnings on January 20th, analysts expect the numbers to look better, since global demand for IT appears to have strengthened.

It will take more than a quarter to get IBM back on track, however. The company is not about to go belly up, as it seemed about to in the early 1990s after corporate IT buyers had shifted from mainframes to more distributed forms of computing, such as PCs and servers. But the industry is going through another wrenching change, which is in some ways more challenging: instead of owning computers, businesses are increasingly renting computing services in the cloud. Not just the technology is on the move, but everything around it.

To understand the depth of this shift, consider mainframe computers. These, combined with related software and services still generate 24% of IBM’s revenues and 35% of its profits, according to Toni Sacconaghi of Sanford C. Bernstein, a research outfit. Mainframes are a prime example of what consultants call “systems of record”. Banks, for instance, use them to manage customer accounts. But to keep the mainframe relevant, on January 13th IBM launched a new model, which is more of a “system of engagement”. One machine can process 30,000 transactions a second, and analyse and encrypt data in an instant. Such features come in handy in the age of the cloud and its corollary, mobile computing: people increasingly do their banking and many other things on smartphones.

As corporate IT opens up to the outside world, the way in which it is developed, sold and used is changing too. Power is flowing from hardware engineers to software developers, the best of whom are in great demand. It is no longer mostly chief information officers (CIOs) who hold businesses’ purse strings; by 2017 chief marketing officers will spend more on IT than their CIOs, reckons Gartner, a market-research firm. Customers are no longer willing just to buy the latest technology; they want to pay for specific results, for instance for sales increases achieved by using analytics software.

All this means that longer-established IT firms such as IBM, HP, Oracle and SAP must rethink how they do business. They will have to offer a working environment that attracts younger people (such as IBM’s design studios in London and elsewhere), develop new products together with their customers and become nimbler to keep up with more focused startups, says Ralf Dreischmeier of the Boston Consulting Group.

IBM had a late start mainly because it concentrated for far too long on its financial goals. Rather than getting ready for the new world, the firm continued to cut costs, buy back shares and shed lower-margin business, such as low-end servers and chipmaking. Things began to change in 2013 when it acquired SoftLayer, a cloud-computing provider. Last year it teamed up with two “new-tech” firms, Apple and Twitter, to develop mobile business apps and mine social-media data. IBM has also started to tweak its organisation. Earlier this month it created separate units for its fastest-growing businesses, such as data analytics, to make them more focused, the better to compete with upstart rivals.

Will these changes be enough? Steven Milunovich of UBS, a bank, thinks it will be another three to four years before IBM’s new businesses, such as its cloud offerings, bring in more money than old ones, such as IT services. But he is optimistic that IBM will make it through this transition intact. And the firm’s longer-term bets, such as Watson, an artificial-intelligence computer system, could still turn into something big. It continues to pump money into research: last year, for the 22nd year in a row, it was America’s largest recipient of patents.

Other analysts imagine a more pessimistic scenario: IBM is increasingly held back by its legacy business; it becomes less profitable because cloud services offer lower margins than the more customised IT services IBM is used to providing; and it steadily loses customers to newcomers, such as Amazon’s cloud-computing arm.

If this starts to look like it is happening, pressure will mount for Ginni Rometty, IBM’s boss, to take more radical action. It is unlikely that the firm will break itself up, as its rival HP has decided to do: IBM’s businesses feed too much on each other. But it could be forced to separate its old from its new businesses more clearly. Many of its clients are going for “two-speed IT”, in which they separate their faster-moving and more innovative IT needs (data-crunching, say) from more basic services (payroll processing, for instance). To keep them, IBM may need to do the same.