Journal: News & Comment

Signal and noise

Permalinks to this entry: individual page or in monthly context. For more material from my journal, visit my home page or the archive.

I find it annoying that so many articles at PC Magazine and other ZDNet/C|NET sites (not to mention other sites around the Web) start with a "summary" page, from which you have to click something else to read the full article.

Here's an example: an article about the Sierra Wireless Voq smartphone. It gives you some specs and the first chunk of the first paragraph, and then you have to click through to the actual article. If you want something you can print that doesn't include all the ads, you have to click yet again.

Now, in this case, the actual article is 4,215 characters long—about 4 KB, including title and byline. The picture of the phone is 7.8 KB. So, 12 KB total for article and photo. How much stuff did my computer have to download to get to that printable version? Something like 1 MB overall, once all three pages (the first two of roughly 300 KB and the "stripped down" one of some 100 KB, plus ads)—with all their associated stylesheets, graphics, and ads—have come down to my browser.

So, the ratio of useful text and graphics to presentation, markup, and advertising is 1:86. In other words, only 1.2% of the bandwidth used by those pages (though more of the visual space, I admit) is actually what I want to look at. Even if I'm generous and take both the basic HTML code and the large stylesheet for the articles as useful, we're still looking at only 54 KB of real quality content—5.2% signal to 94.8% noise.

Compare that to a weblog like my home page. I have about 22 KB of core content there right now, counting only the actual journal postings and none of the graphics, links, or other information. The overall page is something like 200 KB all told. You get everything on the first page, with no need to click through to the full articles. The ratio here is 1:9 signal to noise, or 11% useful content. If you count my HTML and stylesheet too, that's about 67 KB, or 33.5% useful (vs. 66.5% non-useful).

So it's little surprise that, from a pure information-density perspective, I generally find weblogs a more useful source of informative reading than ZDNet.

Of course, I have played fast and loose with the statistics here, and the Voq piece is a short article, but it's pretty clear regardless that the vast majority of stuff on many "big site" web pages is not actually what people are there to read.

 |