5 common e-commerce site mistakes

Your overall site design and infrastructure count, too.

How fast do your pages load? How navigable is your site? What about server capacity and browser compatibility?

Brent Melson is a senior technical designer for NTSL (National Software Testing Labs), a Philadelphia-area company, who makes his living testing Web site architectures and related Internet technologies. Here are his five most-common e-commerce site boo-boos.

Overestimating concurrent traffic — and spending too much as a result.

Most tech writers like me tell you to plan your site for peak traffic periods, especially those that occur in the last-minute buying frenzy before the December holidays.

Melson, however, says the term “concurrent” is frequently misunderstood by Web operators. It doesn’t mean number of customers overall, or even site traffic in a day, but rather the number of users converging on a site simultaneously. Too often, he says, Web operators pay thousands of dollars more for server space and related software because they have been too generous in their estimates on concurrent traffic.

“If you aim high and have an unlimited hardware budget, it’s not going to be a big deal. But if you don’t have an unlimited budget, my advice is to think realistically about the number of people you will have using the site at any one time.”

Alexis Gutzman, an e-business author and consultant, cautions that spikes do occur for many e-tailers around Dec. 12 or so (one of the last days you can ship packages long distances to arrive before Dec. 25). “In my experience, many people don’t overestimate. They fail to plan for the peak,” she says.

No consideration of resolving performance issues with software rather than buying new hardware.

When performance bottlenecks occur, many Web operators quickly conclude it’s a hardware problem — and rush to buy a new server or two to add more capacity. That’s often a mistake, Melson says.

“In our experience, about 70% of the time it’s a software issue and 30% of the time it’s a hardware problem. But instead of thinking about how they can fix or redesign their software, they throw more hardware at it.”

It may be a case of simply rewriting the software or adding new application software, he says. Perhaps they need to buy more memory for their database server, rather than buying a new server. Often the software solution is cheaper and quicker to implement, he says (although hardware currently is as cheap as it has been in some time).

The difficulty here, Melson acknowledges, is that small businesses don’t have information technology (IT) staffs or the time and money to diagnose infrastructure problems and/or rewrite software. Most businesses should have an IT consultant or trusted value-added reseller (VAR) to advise them, however.

Not making the site compatible with more than one Web browser.

If you had to choose one browser to support, it would be Microsoft Internet Explorer, the dominant browser with more than 80% of the market.

“But what about the Apple Macintosh customer — do you want to turn him away?” Melson asks. “How about the Netscape user?”

If not, you need to test your site and system with the other browsers. Some tweaking of the user interface is likely to ensure tables, charts, graphics and functionality work well on the different browsers.

“Often these are very simple or minor fixes. You might need to only change some colors, or add new elements. If you don’t ever look at your site on other Web browsers, you won’t ever know that you need to fix it,” he says.

Supporting more than a single browser is more important for an online retailer than a B2B company, Melson says, because a retailer’s customers are more random. But if you don’t feel it is important, he says, you should put a disclaimer on your site noting that it supports — or works best with — only one particular browser.

Failure to get outside feedback on usability.

“Usability” is now more than a buzzword. It has emerged as a significant metric for how Web sites are viewed today. Usability surveys, usability tests, usability scores and usability focus groups are all part of the research and development of most large Web sites.

Melson’s finding is that many smaller e-commerce operators don’t get usability feedback from anyone beyond those on their development team. But those developers and others are too close to the process and biased toward the chosen design and infrastructure. “You get used to your site and used to any foibles.”

For small businesses, organizing a focus group to evaluate your Web site is beyond your time and resources. But getting some sort of outside perspective — be it employees not involved in the design, or your spouses or friends — is crucial to the site’s development and performance. “You need to hear from people who aren’t working on it.”