Do Numbers Help?

By Bob McElwain
Posted Wednesday, April 23, 2003

Demographics about the Web abound. You may find such data helpful. In general, when I look at the research available, I get a feeling it's incomplete. The Web is so vast, I don't think it's possible to find a small representative sample from which significant results can be obtained that reflect the whole. In the end, what is reported with numbers may not matter at all to you, even though the source is impeccable.

Numbers Don't Always Work

For example, it has been reported that 330 million people are "on the Web." I have not read closely enough to know if this means daily, occasionally, or somewhere in between. To me it doesn't matter.

Even if this number were doubled, it would still mean nothing to me. I am interested in reaching an extremely small fraction of web users. The implication I've been seeing in spam messages of late is that I can reach all 330 million people. This is a lie. But there would be no gain in trying to do so in any case.

Honest Numbers Can Be Wrong

I recently read a report that of nearly 100,000 spam messages received by one firm, about a third were promoting po-rn sites. (I used a hyphen in hopes of ducking blocking software.) What does this mean?

Numbers are funny. I never doubt such reports from respectable firms or people. But I am always skeptical about the numbers themselves. Sure, those were the results obtained. I will accept this without hesitation. But they often do not seem in accord with my experiences.

I get lots and lots of spam. Less than 3% is po-rn related. Do I thus conclude the report was wrong? That they were lying for some devious purpose?

Not at all. It only means their sample of email received was not representative of what I receive. In like fashion, it is doubtful my email is typical of yours.

100,000 spams messages is a very small percentage of what is mailed each day. It is so small, results from this sample have very little, if any significance. These results were obtained, that's true. But they may have no meaning relative to you.

Leave the particulars of demographics to those keen on the topic. Your best plan is to ignore such numbers and focus on interactivity with readers and visitors. In every way you can, seek input, then derive your own demographics from it.

Your Log Files Can Mislead

Recently I was chatting with a fellow who was having trouble getting a page to load under a specific condition in Netscape. Since he uses Internet Explorer, which handled this case correctly, he hadn't noticed the problem until I pointed it out.

When I did, he commented, "Hey, I don't need to worry. Only 5% of my visitors are using Netscape." This fellow is dead wrong in two ways.

Of visitors to my site, over 40% are using Netscape. So have I got it wrong? Or is the fellow reporting 5% wrong? Neither of us is. We are both reporting accurately.

Why Are There Such Great Differences?

The apparent dilemma stems from the fact that we all have our own set of visitors. Each comes to us from a vast pool of many millions of Web users. Those who show up on my site may never even hear about yours, let alone visit.

Thus my visitors are not representative of yours, except as to the fundamentals. For example, all site visitors ask first, "What's in it for me?" Such basics relate to every site. The specifics do not.

Even if a massive, well respected study reported only 1% of surfers use 640 x 480 monitors, it still might not apply to your site. For as suggested above, the pool is so vast, hoping to draw a truly random sample from it is impossible.

Further, things change rapidly on the Web. Not long ago, Netscape was the browser leader. As Microsoft continued to demand Internet Explorer be installed on all new systems delivered, the dominance of Netscape began to fade. Even after being acquired by AOL, market share continued to drop.

Can you assume it will continue to do so? That would leave us with only one major browser. A Microsoft product. A company already at odds with the Justice department in anti-trust actions. It may prove to be in their best interest to assure that Netscape regains a significant share of the market.

What seems so today is suspect, for it may not be so tomorrow. Rather than making assumptions which may prove false tomorrow, the better plan is to accommodate all possible options today, and be prepared to make changes tomorrow.

The Mistake That Matters Most

But the second mistake made by the fellow mentioned above is in ignoring Netscape users however small their numbers be. Suppose only 5% of my visitors use Netscape. To toss away this many potential customers is foolish at least. I take the time to make it work for them.

Hasten Slowly

JavaScript has been available for some time. Is it wise to use it if N% of systems can not deal with it? The better plan is to offer an alternate way to access your site for those who can not.

Plug ins are popular of late. Will users take the time to download and install one so as to see your site in all its glory? I doubt it. What's best is to offer the option to do so, but be sure your site functions effectively without it.

One of my systems uses a Pentium II with awesome supporting resources. However, it doesn't have a sound card. A site that requires I have one, will hold my attention only so long as it takes to hit the Back button or enter another URL.

Killer Assumptions

If we make assumptions about the power and tools our visitors have readily available, to the extent we are wrong, we are driving them off our sites.

When you consider how hard it is to draw a new visitor, driving even one away seems a pretty silly thing to do.