The iPad launch went rather well this past weekend. There are some mixed reviews and a few initial problems (overheating and connectivity,) but this was generally a good launch. This brings in a new class of device that initially is more like a vacation home than a place you'd want to live for most of us. It will evolve, but it will never be as capable as a PC. Having said that and looking at how increasingly employees are using either very old PCs, netbooks, smartphones and products like the iPad to get work done, I wonder if the benchmarks we use represent the real work being done. Granted the iPad, and most of these devices, are more about content consumption than creation, but that is changing.
As you look at your PC refresh cycle, if you don't look at your existing performance requirements, you likely will buy equipment that either under- or overperforms what you actually need. In short, increasingly I'm seeing people buy hardware well ahead of their needs or that is wrong for their users.
Windows 7 Deployment Lesson
A few weeks ago when we had our IT Business Edge midmarket show, the feedback from early deployment and testing sites for Windows 7 indicated the vast majority, unlike prior cycles, were deploying the new OS on existing hardware. Since this hardware seemed to average about three years old, this suggested that the performance gained with the new OS and getting rid of all the junk that was slowing down the machine, was adequate. I'd still be a little worried about hardware failure, given that most hardware was designed for a three-year service cycle, but reports from these sites clearly indicated that users were satisfied with this older hardware.
Switch to Mobile
Also over the past decade, most sites have switched aggressively to mobile products that typically underperform their desktop counterparts. This hasn't proven to be a problem for most users, supporting the contention that performance isn't a gating factor right now and feedback from users that they would favor a product that is more portable over one that is more powerful for most work tasks if they have to carry the laptop a lot. If they don't, the question of whether they actually need a laptop comes up.
Flash Drives vs. Processors
This is also anecdotal, but in listening to users who have received PCs with current-generation flash drives, they appear to appreciate the benefits of higher reliability, faster boot and application load times, and weight over application performance. This would suggest they prefer storage speed over application performance.
Increasingly users appear to be using graphically intensive applications and, according to the executives I speak with, are requesting systems with stronger graphics as a result. This suggests that even for those focused on performance, that type of performance might be an important metric to use when specifying a particular piece of hardware.
VMware, Microsoft and Citrix are all improving their thin client solutions, and Google is attempting to roll to market with a cloud-based enterprise strategy. At the Midmarket CIO Summit Google actually won the award for best midmarket strategy. This all supports the contention that, at least for basic task workers (data entry, clerical, point of sale, reception, etc.), very little desktop power is really needed when the applications these folks use are hosted or cloud-based. Here performance is measured on the server and not the desktop, and the critical factor is likely the reliability and speed of the network, not the desktop PC.
The iPad drives this home for, while it isn't yet ready to displace a PC, future versions might be, and we will likely shortly see employees use the device to supplement their PCs. Once again, this would showcase that, for these users, something other than just performance should be the key selection criteria.
Time to Assess Your Performance Needs
In general practice, companies set a relatively high performance requirement for most of the company, then have vendors bid to meet it. But what if you could accept a much lower level of performance, supply hardware that was both less expensive to buy and less expensive to power and still maintain employee satisfaction?
I think this practice of using a flat benchmark like SySmark to cover the entire population of a company with the same score (except maybe the workstation users) is likely obsolete. It might be time to categorize users in specific groups and match the performance needs of the group to the bid specification to best optimize the related PC purchase budget. You may find that many can continue to use hardware with a refreshed and cleaned system, that you can buy less-expensive hardware for many, and may need to buy more expensive hardware for some with a net reduction of your hardware purchase cost.
I see too many sites that don't ask what their internal clients' needs are and then get surprised that the employees and managers aren't happy with the result. I think that with any tool, including PCs, you need to start with the individual needs of the group or user, then build the spec around that. As new devices emerge, you may find your money better spent on hardware that has a better technology mix by simply refreshing your existing hardware and redeploying it than buying generically. Regardless, the first step is to assess what your users actually need to do their job and then revise the bid specification for that.