When Hewlett-Packard spent roughly $10 billion on the software company Autonomy, it thought it was buying a slice of the future -- investing in the hot trend of big data. But the deal turned out to be a debacle, and not only because H.P. wrote down $5 billion of the purchase.
The ill-fated marriage of the companies is a lesson for H.P. and other older technology giants as they throw billions at supposedly game-changing acquisitions, trying to gain a foothold in the future.
In that future, smartphones and tablets, connected to cloud-computing data centers, are the essential tools of work and play. Companies rent software over the air, rather than buying it with expensive maintenance contracts.
And vast streams of data are continually analyzed to find new patterns and make predictions about consumer behavior and product design. Autonomy, for instance, makes software that can analyze marketing patterns and advise a company on matters like where it should increase marketing resources.
These forces threaten older businesses, like H.P.'s traditional personal computer and data storage products. Other companies, like Oracle, Microsoft and Cisco, also face pressure. They are all trying to buy the future -- and have the cash to do it.
In July, Microsoft decided to pay $1.2 billion for Yammer, which makes a Facebook-like social media product for the office. Oracle recently paid over $3.4 billion for two small cloud computing companies that provide software for human resources and sales management. Last Sunday, the computer networking giant Cisco agreed to pay $1.2 billion in cash for Meraki, a company that manages the free wireless service at Starbucks and other businesses.
There are lots more such deals, from these companies as well as I.B.M., SAP and others.
The pace of change is so fast that Google, so recently seen as an upstart and a giant killer, in 2011 paid $12.5 billion for Motorola Mobility to supercharge its Android smartphone business, which competes with Apple. Earlier this year, Facebook spent $750 million on Instagram so it wouldn't miss the next thing in social media.
But identifying the next big thing can be difficult, said Jeffrey Sonnenfeld, a professor of management at Yale University. Likely as not, he said, deals like the one for Autonomy have "maybe a 40 percent success, 60 percent failure rate."
He added, "The odds are against you succeeding, but the odds are also worth taking."
The real hazard, he said, is in the way companies describe these acquisitions as "natural, inevitable victories." They should be seen, he said, as "an investment, like in research and development."
The pace of acquisitions hardly matches the level seen during the late '90s Internet bubble, when Cisco paid $9.6 billion for three networking companies that former officials said yielded almost no benefit. In 2002, AOL Time Warner wrote down $54 billion in good will related to its troubled merger. Several telecommunications equipment makers and service providers also had multibillion-dollar write-downs.
But there are also notable successes.
EMC, a maker of data storage equipment, paid $625 million in 2003 for VMWare, which makes some of the critical technology in cloud computing. Today, after taking some of the company public, EMC's stake is worth about $30 billion. VMWare recently took a gamble of its own, buying a next-generation networking company, Nicira, that had virtually no revenue, for $1.26 billion.
Microsoft, which also paid $8.5 billion for Skype in 2011, said Yammer was being integrated into SharePoint, Microsoft's successful collaboration software, and will be included in its premium version of Office communications and productivity software at no cost. It did not say whether, or how, Yammer would be profitable on its own, however.
The real issue with Autonomy may be less its questionable sales than its core technology, said Leslie Owens, who follows data analysis software at Forrester Research.
"H.P. thought it was an entirely new platform, but Autonomy's clients said it wasn't as good as Google's corporate search product," she said.
Autonomy, which was founded 16 years ago, "was based on using powerful algorithms," she said, but the software was not attuned "to new kinds of search signals, like what your friends are doing, what people like you are doing, what other data you might draw off on." Over time those new methods attracted customers.
In a demonstration of the Autonomy product last month, H.P. appeared to have addressed some of those issues, but others remained. An application to help figure out what to pay for Web ads initially failed to work, then delivered results that appeared similar to those of several other search products.
"We remain 100 percent committed to Autonomy and its industry-leading technology," H.P. said in a statement. "The company's products are cutting-edge and provide many customers with unique solutions."
In an interview Tuesday, Meg Whitman, the chief executive of H.P., indicated that the company also had strong hopes for its security software, most of which it got from other acquisitions.
H.P. has accused Autonomy of improperly accounting for much of its sales in the years before H.P. bought the company last year. It took a $5 billion noncash charge related to the purchase price.
Ms. Whitman gave Autonomy tepid support.
"It's very disappointing," she said of the write-down. "We've integrated the technology into a couple of places; we will integrate it into more."
This article originally appeared in The New York Times.