Saying No: When to Trim the Market

Jack Zeal, Lead Developer

Considering the technical politics involved, and over a decade of development, the Web has remained remarkably universal. Most sites work for most users. Much of that success can be linked to standardization and a general attitude of “information is most useful when least encumbered.” However, there are situations where maximum accessibility represents becomes a tradeoff with other business concerns. When you have to define the audience as less than “everyone”, who should you consider excluding, and why?

Outdated Browser Users

You probably have a friend or relative somewhere who still runs Internet Explorer 4.0 on Windows 98. Her computer came with it, and she’s not going to upgrade until the computer is replaced. However, she is in the vast minority now. Although every site is slightly different, the combined market share of older browsers (Internet Explorer versions below 6, Firefox below 1.5, almost any version of Netscape) is likely to be less than 5% on a general-interest website.

Certain specific markets can make even more focused judgements, based on a pre-targeted client base. Your corporate intranet is unlikely to be viewed by any browser the IT department won’t support. A Linux support site will rarely be accessed from Internet Explorer.

Why is this so important? Outdated browsers tend to require much greater levels of “workaround” development and testing to ensure compatibility. By releasing resources from satisfying less than 5% of the market, you can devote them to adding generally desirable features, or ensuring popularity in browsers which are growing in popularity, like Safari and Firefox.

An important thing to remember about this type of exclusionary behavior is that it’s a ‘soft’ exclusion. Just because you don’t explicitly test and develop for Netscape 4 compatibility doesn’t mean that the the site will automatically become unusable in it. It might work fine, or degrade to an ugly, but still usable, state. You may therefore be able to say “We’ll launch with a site that works for 98% of the users, and if budget comes available, fix the rendering oddities for the last 2%”

Mobile Users

The availability of Web-capable phones and ways to present specialized content to them—from .mobi domain websites to WAP/WML “mobile web” offerings, has made many firms eager to find ways to put their site in your pocket. Unfortunately, the message often gets left behind in the process.

If you’re developing a site specifically for a mobile device, you must ask yourself “what do you expect people to do with the site on the go”. Remember that mobile phones vary wildly in interface and functionality, and that data plans are often expensive, tightly metered, and slow or erratic in performance. This leads to two basic concerns:

Can your sales pitch or message be successfully delivered in those constraints? If you’re selling an expensive, complicated product that benefits from huge images and detailed specifications, many users aren’t going to want to scroll through them.
Even if it’s technically feasible, is it customer-approved? There are many mobile markets which are not yet mature, in part because of customer hesitation. Will non-technical customers that have barely learned to trust buying online at their PC be ready to make a $500 purchase via their mobile phone? Will they want to type in enough data to complete an order or contact form on a telephone keypad? Will they even be willing to visit your mobile site out of fears that it will result in a $82 data tariff?

In many cases, it makes more sense to develop mobile-specific content which appeals to users’ desires for quick-hit information, but NOT to struggle to adapt the meat of your site to fit on a two-inch screen. Let the users buy their airline tickets at their PC, but check if the flight’s delayed from their phone. This not only spares you the hassle of trying to cram a desktop experience into the customer’s pockets, but it demonstrates you understand their needs and aren’t just trying to get a check-box feature.

Search Engine Robots

Yes. I said it. There are times you may not want the search engines all over your site. Yes, it’s heresy coming from a search-engine-optimization firm. Or is it? The goal of getting indexed is to connect you with customers. Letting search engines blindly index your site, in combination with specific design problems, can lead to listings that just strand or disappoint customers.

Dynamically-calculated content in general is a bit dangerous. The search spider saw something that looked relevant two months ago, but now the page is completely different. This problem can be largely remedied by ensuring that a URL stays on the same general topic over time, but some pages—like recent news listings or direct links to search results, are always subject to losing their meaning.
If your site offers multiple URL paths that point to the same content, it’s worthwhile to try to limit search engines to the least changeable one. For example, a forum system might display the same information as “Thread 123456” and “Forum 21, 8th thread from the top”. “Thread 123456” is likely to always resolve to the same place, but a link to “Forum 21, 8th thread from the top” might refer to another thread by the time the customer arrives.
Session-specific URLs often inspires extremely poor search indexing. When the customer clicks on the link, they may not be able to continue the session the search engine spider started, thus getting different data, or in the worst cases, obnoxious “session expired” error messages that kick them back to a starting point with little clue how to get to where they wanted to go.

Notice that I’m not advocating banning search engine spiders from the meat of your site— but rather, limiting access to make sure they stuff their index with useful models of your site, not the pages that won’t be useful by the time a customer arrives.

Overseas Users

Again, I’m presenting a controversial approach. Many website owners might be tempted by the promise of large untapped markets, just across the border. However, in specific, well-defined conditions, there may be a benefit in excluding overseas customers.

If your site is “expensive” to present to customers—for example, if the server is easily bogged down, or the hosting company charges you a fortune for bandwidth—you might want to try to drop visitors who you can expect won’t convert into sales. In many cases, overseas visitors are the easiest ones to write-off. Some businesses are inherently bound to small regions—nobody from Ottawa or Beijing is going to call a firm in Tucson, Arizona to send out a dishwasher repairman. Other sites may sell worldwide, but are of clear disinterest to a region. If your business plan is to buy goods easily available in one country, mark them up, and resell them to the world, nobody in that country will pay you to be their middleman.

Furthermore, there are times when tailoring your offerings to an international clientele can be the source of the cost. For example, investment laws vary wildly from country to country, so you might not want to show someone a product which cannot be legally sold where he lives, or spend thousands of dollars developing a compliant version. Other similar expenses could arise from laws requiring business- and thus your site—to be offered in the local language, or through local affiliates.

Finally, even when you COULD service international customers, some firms have come to the conclusion it’s not worth it. Some countries are known for high tarrifs, unreliable shippers, or simply unusual rates of payment fraud. You might lose a few legitimate sales, but if it saves you the frustration and expense of chasing down these problems, you may end up ahead.

There are several different approaches to excluding foreign customers, and the approach you choose depends significantly on the reasons you want to keep these customers out:

Outright banning of access to the site based on the visitor’s IP address can be useful if you’re absolutely certain any customer from that country will be a waste of bandwidth, you’ve experienced wide-scale abuse (like a Denial-of-Service attack or non-stop spam), or if you’re seeking to ham-handedly comply with local regulations that prohibit your offerings in that region.
Well-designed forms and check-out processes can simply reject customers from overseas, or divert them for manual review, if you’re worried about fraud, shipping, or customs concerns.
Simple “underdesign” of your site—not supplying every possible translation or providing localized details—may be enough if you’re primarily trying to dissuade casual inquiries you know you can’t service.

Correctly limiting access to your site is a matter of surgical precision and the “90/10 rule”— 10 percent of the visitors represent 90% of the hassles you’ll have to deal with. In spite of the benefits which can appear if done well, it’s something to be approached with extreme caution and skepticism—block the wrong part of your site from the search engines, or ignore the rendering faults in a browser used by too much of your market, and you start throwing out serious customers.