Sebastian's Pamphlets

If you've read my articles somewhere on the Internet, expect something different here.

MOVED TO SEBASTIANS-PAMPHLETS.COM

Please click the link above to read actual posts, this archive will disappear soon!

Stay tuned...

Thursday, July 05, 2007

Why eBay and Wikipedia rule Google's SERPs

It's hard to find an obscure search query like [artificial link] which doesn't deliver eBay spam or a Wikipedia stub within the first few results at Google. Although both Wikipedia and eBay are large sites, the Web is huge, so two that different sites shouldn't dominate the SERPs for that many topics. Hence it's safe to say that many nicely ranked search results at Googledia, pulled from eBaydia, are plain artificial positioned non-results.

Curious why my beloved search engine fails so badly, I borrowed a Google-savvy spy from GHN and sent him to Mountain View to uncover the eBaydia ranking secrets. He came back with lots of pay-dirt scraped from DVDs in the safe of building 43. Before I sold Google's ranking algo to Ask (the price Yahoo! and MSN offered was laughable), I figured out why Googledia prefers eBaydia from comments in the source code. Here is the unbelievable story of a miserable failure:

When Yahoo! launched Mindset, Larry Page and Sergey Brin threw chairs out of anger because Google wasn't able to accomplish such a simple task. The engineers, eager to fulfill their founder's wishes asap, tried to integrate mindset-functionality without changing Google's fascinating simple search interface (that means without a shopping/research slider). Personalized search still lived in the labs, but provided a somewhat suitable API (mega beta): scanSearchersBrainForContext([search query]). Not knowing that this function of personalized search polls a nano-bugging-device (pre alpha) which Google had not yet released nor implemented into any searcher's brain at this time, they made use of that piece of experimental code to evaluate the search query's context. Since the method always returned "false", though they had to deliver results quickly, they made up some return values to test their algo tweaks:

/* debug - praying S&L don't throw more chairs */
if (scanSearchersBrainForContext($searchQuery) === false) then {
$contextShopping = "%ebay%";
$contextResearch = "%wikipedia%";
$context = both($contextShopping, $contextResearch);
}
else {[pretty complex algo])


This worked fine and found its way into the ranking algo under time pressure. The result is that with each and every search query where a page from eBay and/or Wikipedia is in the raw result set, those get a ranking boost. Sergey was happy because eBay is generally listed on page #1, and Larry likes the Wikipedia results on the first SERP. Tell me why the heck should the engineers comment out these made up return values? No engineer on this planet likes flying chairs, especially not in his office.


PS: Some SEOs push Wikipedia stubs too.

Labels: , , ,

Share this post at StumbleUpon
Stumble It!
    Share this post at del.icio.us
Post it to
del.icio.us
 


-->

0 Comments:

Post a Comment

Links to this post:

Create a Link

<< Home