web analytics

Businesses think long and hard before buying computer hardware and software. Some work from finely tuned technology plans. Others might take an ad hoc approach, making up rules as they go. This may sound unwise, but it has the advantage of flexibility.

Some managers consult with their end users and technology specialists before drawing up specification lists. Often business buyers prefer the security of a known brand, some seek comfort in long-standing ‘technology relationships’, and others like the prices or local service offered by smaller outlets.

Whatever the details, business people tend to put a lot of research and effort into their system and software purchases. So should you. Just make sure you worry about the right things.

While it makes sense to plan software buying in detail, this might not be the case for hardware. After all, modern computer hardware is a commodity. There isn’t much difference between one brand and another. Inside they are pretty much the same. After years of conducting in-depth benchmarks (among other jobs I edited the Australian edition of PC Magazine), I can confirm that the difference between the top-performing brands and the average is rarely more than a couple of percent.

Whisper it quietly, but this difference is well under the margin of error. Indeed, often, when comparing classes of computers, the performance spread between the highest and the lowest is often less than the margin of error. And even if it isn’t, I challenge anyone to sit at any two similarly configured PCs and tell me which one is running five percent faster. You won’t notice any difference running Microsoft Word or working online with Firefox or any other browser.

For people in business performance is an issue, but the performance that matters is that between different classes of machine, not different models within a class.

If raw processing speed worries you, most of the time you can boost it simply by adding more Ram. Spending a $100 on extra memory chips is the best IT investment you’ll ever make. Not only will this kick-start sluggish systems, but you’ll be able to do more work and work more productively.

Of course, benchmarking does show up poorly performing products. But these are as likely to come from the most prestigious stables as from the cheaper no-brand operators. By all means use benchmarking information to avoid the dogs, but in the long run, average performing machines are as good a purchase as the fastest.

While the performance spread of similarly specified PCs is minimal, prices tend to have more variation. Both follow the well-known bell curve. But the top and bottom performers in any class might deviate three or four percent from the average, while prices can vary by up to 20 percent; and even more if we include Apple‘s expensive hardware in the list.

You might expect that prices vary with performance. They may, but only up to a point. Statisticians and economists call the way two variables interact; ‘correlation’. So, if price and performance ratings match, they would be highly correlated, if cheaper machines performed best, then they would have negative correlation. In reality, there is merely a weak correlation between price and performance.

If you draw a graph and plot performance against price, there would be a pattern, but a number of points on the graph would sit a long way from any trend line or cluster. These are the machines to watch. Those that are nearest to the corner where performance is sluggish and prices are high represent the worst value. Those in the opposite corner represent the best.

It might seem like a lot of work, but this can be a worthwhile process if you need to buy a lot of hardware. However, it is worth remembering that differences in performance rarely matter, dollars in your pocket do.

So, what PC purchasing lessons are there for individuals and small business owners? The key is to get your IT spending into perspective. When shopping for hardware, you should pay more attention to the features included in the package than to any benchmarking details.

Remember warranties and reliability are more important than performance. And above all else, remind yourself that a low-price, average performing system plus $100 spent on Ram will almost always give you a better return than a pricey speed demon.