PC prices continue to sink even as their power increases, leading some experts to think a two-year replacement cycle makes sense for compute-intensive jobs.
A recently published study, Replacing Enterprise PCs: The Fallacy of the 3-4 Year Upgrade Cycle, revealed productivity gains and cost savings based on tests that map computing performance with typical job roles in the enterprise.
Machines that are more than 2 years old make people less productive than workers who use newer machines, said Jack Gold, president of J.Gold Associates, an IT consulting company in Northborough, Mass.
PCs should not be viewed as a pure cost, he said. "It's a tool you can provide your users to make them more productive," Gold said.
The study bucks today's conventional wisdom, in which enterprises typically refresh PCs every three to four years. And it's not surprising that some IT directors are not buying the study's logic wholesale.
"Unless you move to leasing, it's a tougher buy for enterprises to move from three years to two years," said Brian Katz, director of mobile innovation at a large pharmaceutical company based in New Jersey.
Katz did say he believes the study makes a compelling case for enterprises to move to a two-year PC replacement cycle in situations where employees need more powerful devices for their jobs.
Ira GrossmanCTO, MCPc Inc.
The study, which was published earlier this month, used the SYSmark 2012 application benchmark for six job test scenarios, including Web development, office productivity, media creation, data/financial analysis, 3-D modeling and systems management. Each test ran on a 2-year-old notebook, the HP EliteBook 8570P, and the newer HP ProBook 650, with the main difference being the faster processor.
The research weighted the time workers spend on the computer in a typical workday, mapped the benchmark tests to job roles and accounted for the average salary of those employees. The roles included office worker, engineering, administrator, business analyst, Web developer and IT staff.
Results showed productivity gains ranging from 3.6% for data/financial analysis to 20.1% for 3-D modeling. In jobs that used the new computers versus a 2-year-old machine, engineering saved enterprises $13,103 -- the most annually -- followed by a Web programmer at $11,151.
With the annual savings, the study also calculated the average number of equivalent workdays gained per job role, with the lowest being 11.12 days for a business analyst and the highest being 27.3 days for an engineer.
Will the study's results become reality?
IT pros doubted that a two-year replacement cycle would fly in the enterprise.
"The test results are interesting but the conclusion is not applicable to most companies," said Mike Drips, a longtime IT consultant in Houston. Two years ago, Drips upgraded a global company consisting of 100,000 employees with added RAM and video cards rather than new PCs.
Bring your own desktop (BYOD) programs also play a factor in whether companies choose to adhere to a two-year replacement cycle.
Many companies support BYOD to save on hardware costs. However, enterprises still must control the assets, Drips said.
In some cases, a two-year PC replacement cycle might work, depending upon the job function and the company's IT environment.
Keeping devices more than three years past the manufacturer’s warranty encourages the company to be in the break-fix business, said Ira Grossman, CTO of end user and mobile computing at MCPc Inc., a large system integrator in Cleveland.
While businesses may be able to extend the life of a PC for a task or data-entry worker where downtime is not critical, power users will need updated systems, and IT needs an accelerated refresh schedule for them. Engineers and developers might require high-performance computers compared with office workers, but cost is a factor before enterprises replace PCs or laptops every two years.
Other PC replacement factors
As enterprises add cloud computing, IT will need to spend less time upgrading machines and applications, the study said.
Not all IT professionals agree.
"The cloud will not have an impact," Drips said. He stressed that the two-year PC replacement cycle is more dependent upon a worker's function.
Meanwhile, capital costs and how long it takes enterprises to transition from one PC to another also weigh into the debate over a two- to three-year PC refresh rate.