Simpatico Rankings

 
The Real Search Engine Watch

Our pages may never attract millions of visitors like Yahoo, Google, eBay, or Hotmail. Nevertheless, we will periodically evaluate metrics that measure the impact of Web sites, starting with the one thing most people do when they get on the Internet.

The Google Conundrum

What began as a simple Web search exercise has turned into a sobering look at search engines. You see, we’ve been using Google to gauge the popularity of our Web pages since early 2003. Since we perform the same Google searches each month, we can tell when something is amiss with its software. October/November 2003 was the first time when Google inexplicably stopped listing some of our pages. We also noticed a dramatic decrease in the number of hits found across the board. For example, one query’s results went from 15 to 2, another from 82,000 to 58,000.

How do we know when a page is no longer in Google’s index—or at least not displayed by Google? Well, if the number of hits is relatively small—say, under 200—we can simply look through them all. And we can also search for a specific phrase (using quotes) that we know is on a particular page. We suspected at the time that Google’s filtering software might have skipped those pages that contained strong language (and perhaps the pages that linked to those “offending” pages as well).

When some people questioned how Google could return a hate site as the top result of a search, Google’s response in April 2004 was that it would never deliberately omit a Web page unless required to do so by authorities. So maybe Google doesn’t censor hate sites, but what about Google’s misguided filtering option, which we believe was introduced around September 2003? It seemed quite a coincidence that all of our missing pages happened to contain some “bad” words. In the meantime, search marketers were blindsided in November 2003 when Google made an unannounced search algorithm change (known as the Florida update). Could this have been the reason for the sudden de-listing of some of our pages?

We did contact Google by e-mail about our concerns during the second week of November 2003. Google specifically stated there was no change in indexing techniques or algorithms and attributed what we reported to normal “fluctuations.” A bug by any other name is still a bug.

From 2004 to Present

Before 2004, the top three search engines—Google, Yahoo, and Microsoft—were not dramatically different from each other (despite some of the Google quirks documented above). By the end of 2004, we started noticing substantial differences among the three contenders.

Google: Throughout 2004, our Web pages, which were usually ranked in the top 10, would disappear from Google’s database now and then. Things didn’t improve much in 2005. Though they didn’t get worse in 2006, Google continued to produce disappointing results. For example, when we entered a general search for local radio, a mortgage company was returned. We were also surprised to see a nine-year-old newspaper article on the firings of two DJs.

Yahoo: Yahoo’s search engine is the only one of the Big Three that’s been working consistently. We shudder to think what would happen if and when Yahoo decides to “improve” its engine. July 2006 was the only time when several of our Web pages became MIA. Yahoo got back on track right away (we did report the problem to the company).

Microsoft: It made sense that Microsoft and Yahoo would return similar results in the past because Microsoft used the same engine Yahoo was based on. Then Microsoft released its own search engine in January 2005—after only two months of beta testing. We’d warned Microsoft after we tested the beta version with our queries and reported the same problems again with the new release. But for the first half of 2005, Microsoft managed to skip most, if not all, of our Web pages. The company finally fixed the problem in July 2005—so its search engine was basically useless for six months. Since mid-2005, our pages have been ranked consistently in the top 10. Despite a small hiccup in May 2006, Microsoft is now as reliable as Yahoo.

[See the latest Search Engine Scorecard]

We don’t expect our Web pages to have the same rankings across different search engines. Indeed, a 2005 study shows there is only a 3 percent overlap among the first 10 results returned by three of the top engines (Google, Yahoo, and Ask Jeeves). But we do expect all our pages to be indexed at the very least. When search engines fail to do that, it makes one wonder what other Web pages are being ignored. And when we see dozens—and in some cases hundreds—of less relevant pages returned ahead of ours, it’s a little disconcerting.

The press tends to accept business claims without challenge. When a company declares its product has a large technical lead over the competition, most journalists are too lazy to do their own investigations. Don’t believe the hype.

Have you heard of searchenginewatch.com? We’d hoped it’d be the Consumer Reports for search engines. Well, the last time we checked, we didn’t see any concrete and meaningful comparisons like what we’ve published here. What a shame. We urge Web surfers (especially those who publish their own pages) to monitor search engines as we do. Let’s compare our findings.

Caveat Emptor

As the amount of information on the Internet increases exponentially, search engines play a major role as gatekeepers. When they don’t get it right, we all lose.

Cool Web Surfers Don't Cut and Paste

Would you like to share this Web page with friends? Don't cut and paste. Provide a Web link to this page or refer to its Web address. We invite all content providers to join our "Don't Cut and Paste" campaign.

Copyright © 2002-2018 Calba Media LLC. All rights reserved.

Labels: google, internet search, microsoft, popularity, simpatico, yahoo

posted by Blog Masters at Permalink