Consumers delay their hardware purchases because they are confused by PC terminology, according a survey conducted by Advanced Micro Devices.
AMD surveyed 1,535 consumers around the world about their comfort level with various products, and quizzed respondents on the definitions of some commonly used PC terms, such as megahertz and web browser.
The respondents had trouble identifying the correct definition of many of those terms, even those who had owned a PC for several years. Because of this confusion, many consumers become frustrated when researching a new PC, and postpone their technology purchases as a result.
For consumers to fully understand what a PC can do for them, the industry needs to develop a system-level performance metric that can compare one PC with another, regardless of whether it uses a processor from AMD or Intel, a graphics card from Nvidia or ATI Technologies, or any number of other components from different suppliers, said Patrick Moorhead, vice president of consumer advocacy and corporate marketing at AMD.
"The key is making it relevant to the end user, taking 'What can I do differently with my new PC than my old PC?' and converting that into plain English," he added.
Insight 64 principal analyst Nathan Brookwood believed that many system vendors would rather keep the confusion alive.
"This lets them put together configurations with a couple of strong components, and several weak components, and get away with it," Brookwood said. Skimping on certain components can reduce the overall cost to build a PC, but the strong components can justify a higher price, he added.
Right now, the only way to compare hardware is with benchmarks, which are tests of processor performance in completing different application tasks, and are generally sponsored by independent media organisations or industry consortiums.
Over the past year, AMD, Intel, Nvidia, ATI, and Apple Computer have either lodged, or had lodged against them, allegations of cheating on benchmark tests based on the common practice of optimising code. Some analysts and users believed that sometimes this optimisation goes too far, and casts doubt on the validity of benchmarks in general.
Developing a system-level benchmark or metric that measures general performance, but also conveys the benefits of a new technology to those specific markets, would be an extremely difficult undertaking, said John Peddie, president of John Peddie Research. The PC industry would need to take all those different markets into account and come up with a way to represent them all, which would take a lot of time and effort, he added.
Tom Krazit writes for IDG News Service