How JCPenney possibly rigged Google's search algorithm

The Google search page appears on a computer screen in Washington on Aug. 30, 2010.


Kai Ryssdal: We are all, no doubt, familiar with Google and its search engine. The company says it gets a billion searches a day. And generally speaking, the Google methodology tends to work pretty well.

The thing is, the system can be gamed. New York Times business reporter David Segal pointed that out this weekend, when he noticed that until very recently, when you searched for all sorts of things from dresses to jeans to tablecloths, the first retailer that popped up was JCPenney. And it happened again and again and again.

So we called him up to talk about it. David, good to have you with us.

David Segal: It's nice to be here.

Ryssdal: Seems to me the place to start is with a definition of a how a good, normal Google search works once you type something into that box -- what happens?

Segal: Well what Google does is they measure the popularity of your website by essentially counting the links to your website from other websites that are typically associated with the same topics. So if you have a website on, say, folding paper airplanes, the more people who are aficionados in the art of folding paper airplanes that link to your website, the higher your website about folding paper airplanes will rise in Google's rankings. That's the way it's supposed to work.

Ryssdal: And in fact, what happened with this JCPenney incident is this thing that you call "black hat optimization." Tell us how this works.

Segal: If you think of the way it's supposed to work, as a form of voting, then what happens with black hat is essentially vote-rigging. JCPenney or someone working on its behalf established thousands of links related to dozens of different search terms and set up those links on websites around the world. So there are, for instance, 2,015 different pages that had links to JCPenney's dresses page. In fact, JCPenney ranked number one for several months for the search word "dresses."

Ryssdal: We should point out that JCPenney has denied this to you; they said, basically, we didn't do it. But you found some websites out there with these JCPenney dresses links on them that were pretty obscure and fairly random.

Segal: Totally random. This was the strangest thing about this, as I sort of waded into it. You know, there was a casual dresses link on nuclearengineeringaddict.com. There was casinofocus.com; dentist websites. Tons of them seemed to be completely abandoned.

Ryssdal: This is not kosher in the land of search engine optimization, but did Penney get in trouble with Google because of this?

Segal: They did. So I presented all the evidence of this campaign last week to Google, and they quickly determined that this violates their guidelines. What they did was something called "manual action," which means instead of changing their algorithm, which would change results for everybody using the same strategy, someone in Google went through and by hand, essentially, demoted JCPenney in dozens and dozens of searches. So it had just been buried. And it was kind of humbling to see just how much power Google wields in the world of e-commerce.

Ryssdal: Not that this was any of my business, but were you just randomly searching for dresses and bed linens on Google one day and noticed that everything was coming up JCPenney?

Segal: No, a reader noticed. A reader who is in the retail business, although not in a field that is a competitor to JCPenney. He actually makes signs, of all things, and just had taken an interest in the world of search engine optimization and had noticed there were three to four of these anomalous results. I thought it was very peculiar and then found a guy, Doug Pierce at Blue Fountain Media, who did an incredible job of just sort of sleuthing through exactly what had happened.

Ryssdal: David Segal, he's a business reporter at the New York Times. David, thanks a lot.

Segal: Thank you.

Log in to post5 Comments

While Mr. Martinez's comment that google places little value in links is completely incorrect, in the case of JC Penney, there was also a lot of on page webspam going on, such as hidden, spammy keyword rich links.

But make no mistake, the completely irrelevant, low quality exact anchor text off page links garnered for JC Penney, played a huge part in their ability to rank do highly for so many coveted and lucrative keywords.

The problem, in my opinion, is that giogle puts way too much weight on anchor text, I.e. the words that are in the clickable part of a link.

Sergey Brin and Larry Page spoke with the media on October 7, 2009. Several Websites attributed the following remarks to Sergey Brin:

Schonfeld: Is PageRank long in the tooth, are links the still the best metric?

Sergey: No they are not and we decided that in 1999. We use various link algorithms, including what pagerank has evolved to, links are 1%, 100 other factors we look at. Yes,there is spam, and the web changes. We are able to do better and better, can do a much better job ranking than we could a decade ago, if we had rested on our laurels and just stayed with what was in our paper we published in 1998 we would be in pretty bad shape right now.

SOURCE: techcrunch.com/2009/10/07/a-conversation-with-sergey-brin/

1. Google makes enough money to create a crawler program to look for this sort of thing. Since they're making big bucks fr/ search, I believe they have an OBLIGATION to do so. It would make knowledgeable search users, at least, more confident in the results. 2.Could someone knowledgeable please comment on Mr. Martinez's comment? Is this factual?

way to go - i guess the largest and most powerful internet company in the world really needs the help of the NYT and public radio to investigate this fake "fraud". google has SO MUCH power and influence and yet they just randomly make "guidelines" that everyone must follow? just ignore the 30,000 people a day that starve to death every day on this planet and focus on helping google "nab" jc penney. yeah! journalism in america!

Mr. Segal's information about Google's algorithm is very misleading. Links rarely play the role in search rankings that he says they do. Commercial search strategies simply cannot capture enough link anchor text to outperform the natural search rankings that are determined by the content on Web pages.

The NYT story and this particular interview will probably do more damage to Websearch by perpetuating these link myths than anything a spammer might have done on behalf of JC Penney or other companies.

With Generous Support From...