The History of the GoogleBomb
And Google's Response To It, Part 1 of 3
June 16, 2011
Posted in Ken's Blog
Leave a Comment
The Googlebomb has a fascinating 10+ year history to it. Google's reaction to the ones that "go public" is perhaps even more fascinating.
I think you'll find this to be a surprising look at the inner thinking of Google. What may be seem reasonable and believable at one particular moment in time turns out to be quite the opposite with the luxury of reviewing a longer sweep of history.
We tend to think of Google as an open, transparent, webmaster-friendly company.
That's what we are supposed to think. The reality, it turns out, is far from the perception.
The next 3 posts cover the 10-year history of the Googlebomb and Google's attempted explanations (AKA "coverups") to their public "bombarrassments."
A "Googlebomb" is characterized by many web pages ("linking pages") that link to one particular page (the "target page").
The linking pages each place the same keyword (or closely related terms) in the text of the links to the target page.
The goal? To push the target to the top of the Search Engine Results Pages (SERPs) for that keyword.
It only works for less competitive search terms because its success depends on "overwhelming" one part of Google's search ranking algorithm... inbound links.
This makes it possible to manufacture a Googlebomb for a term like "site build it scam" relatively easily. But not for the word "website," which is too competitive to exploit any single factor (or narrow range of factors) in a ranking algorithm.
How To Detect and Deal With a Googlebomb
As you can see, a Googlebomb has very particular characteristics. It is, therefore, quite easy to recognize...
The 5-Minute Googlebomb Algorithm
Here is a quickly composed set of requirements for what should be a relatively simple and effective algorithm to execute (for Google)...
Find a (target) page that has the following characteristics...
1) It receives a disproportionately high number of inbound links (i.e., high relative to the search demand of the keywords in the link text of the inbound links).
2) i) Keywords in the link text of inbound links have low search demand.
ii) A higher-than normal percentage of the keywords in the text across all inbound links are either identical or highly related.
iii) And the target page's on-page keywords MIGHT have little or nothing to do with those in the link text of inbound links (NOT a necessary condition, but adds to the level of certainty if this condition is present).
3) Inbound links are from a higher-than-usual percentage of pages on sites/blogs with little or no "authority."
4) Inbound links accumulate significantly faster than normal (given the subject matter) over a period of time, often shortly after creation of the target page.
What should be done when this algorithm detects a Googlebomb? Depending on the level of certainty, either...
1) De-index the target page permanently.
OR (for lower levels of certainty)...
2) Lower the target page's search ranking for the involved keyword(s) (instead of de-indexing it) according to the level of certainty.
3) Penalize the linking pages that participate in the bomb, too.
I could improve this algorithm if I took another 5 minutes. ;-) But the point is that this algorithm is simple for Google (obvious in fact). As you read the following history, as you encounter each example bomb, ask yourself the following question...
"If The 5-Minute Googlebomb Algorithm existed, would this one have ever seen the light of day?" You will see that the algorithm would have detected and defused Googlebombs from 10 years ago, all the way up to today, equally well. Combine that conclusion with the following fact...
Googlebombs are as obvious as many "Black Hat" SEO techniques that have long since been defeated by all engines. That leads us to a surprising question (one that I would never have given any thought before)...
How Strong Is Google's Grasp of the
Link Structure of the Web... Really?
The nature of the Googlebomb has not changed. Nor has Google's attitude to it. The sophistication of their communication has evolved, as we shall see.
Google's inability to detect Googlebombs today calls its quality (and hence its users' experience) into question. The fact is that Google is no better at finding Googlebombs today than it was 10 years ago.
The Googlebomb is an extreme example of link abuse, so it should be easily found and defused. If you can't detect and dismantle Googlebombs, how good is your grasp of the link structure of the Web?
And What About "Pap?"In my last post (about the upcoming Panda 2), we explored how Google is unable to recognize "pap" (regurgitated content) more accurately.
The Panda release did succeed in eliminating much Content Farm (and other) pap. It came at a heavy cost, though... an unacceptably high rate of false-positives, causing much harm to good sites.
Sites that create pap exploit Google's weakness for relatively "long tail" keywords through crappy content.
Sites that create bombs do the same through link abuse.
Isn't the fundamental role of Google to not only find relevant search results, but to find the best?
Content... Links. Why aren't more people asking how good a handle Google has on the basics?
As we review Google's 10-year response to "The Bomb," you will see a remarkable parallel to how Google manipulated the Panda non-conversation (and doubtless will continue to do, with Panda 2.2 coming).
You will see "Evolving Opaque Transparency" in action, reaching its pinnacle just recently. By the end of Part 3 of this series, you will better understand their Panda (and future) communications.
Let's get started...
The Bomb Begins
This historical review covers the highlights and focuses specifically on link-bombing as it relates to Google. This is not an exhaustive listing, rather a representative sampling to illustrate Google's reaction to H-bombs at various times. (Note: The "Hydrogen Googlebomb" is one that breaks out of "Net marketing blogs" and into the mass media.)
We begin with the first widely known Googlebomb...
Over 10 Years Ago
The First "George Bush" Googlebomb (for "dumb m_f_r", article from WIRED.com)
Google's response? It plays dumb...
"We think that this appears to be an anomaly."
"It's difficult to see which factors contribute to this result."
Google goes on to say...
"It has to do with Google's ranking algorithm ... and we like to find out about these things. They let us know how to make Google better."
"... strictly some kind of mistake, one that would be fixed in Google's next index of the Web."
Read the rest of the article. The journalist is genuinely puzzled, as is the webmaster of the victimized site, perturbed to the point of suspecting sabotage. He speculates that maybe "someone hacked our site."
The journalist and the webmaster were confused.
And Google? It faked confusion...
But the SEO community knew of Googlebombs as far back as 1999, according to this article in Search Engine Watch, discussing how a search for "more evil than Satan himself" delivered "Microsoft" at the top ranking at Google in 1999.
Two years "post-Satan," Google must have known exactly what was happening, instead of pleading that it was "difficult to see which factors contribute to this result."
Moving on to "GB II"...
7.5 Years Ago
Second "George Bush" Google bomb (for "miserable failure", article from New York Times)
The article quotes Craig Silverstein, Google's director for technology...
"the company sees nothing wrong with the public using its search engine this way. No user is hurt... because there is no clearly legitimate site for 'miserable failure' being pushed aside."
No user is "hurt" since there is no better site to deliver for "miserable failure" than President Bush's biography?
If George Bush himself was to give such a disingenuous and insensitive answer to a question on national policy, the press would be all over him.
"We just reflect the opinion on the Web, for better or worse."
The opinion of "the Web" (i.e., the collectivity of all opinion on "the Web") was that George Bush's biography was the best result for "miserable failure"?
The actual truth, and Google must have known it at the time, is that Google was being manipulated, by a relatively small number of people, into giving a bad search result. This was neither the will of the Web nor was it the best result for that particular search.
In 3 years, the Googlebomb's characteristics had not changed. Google's reaction, though, did. It had evolved the message from "fake confusion" to "it's OK, nothing's wrong." Google had entered the "dismissal phase" in their "Evolving Opaque Transparency."
Basically, their reply boils down to, "I'm OK, You're OK, But He Looks Like An Idiot." So far, this is true. But this was not the will of the Web. It was the manipulation of the few for political gain.
The mass-media was not (and is not, still) savvy enough to figure it out. Or perhaps it does not care, happy to cover the eyeball-pulling sensational aspects. At least it does some good. It forces Google into the light.
5.75 Years Ago
Marissa Mayer, still on George Bush as "miserable failure" (from post in Official Google Blog)
Ms. Mayer basically explains that Googlebombs exist...
"Google's search results are generated by computer programs that rank web pages in large part by examining the number and relative popularity of the sites that link to them. By using a practice called googlebombing, however, determined pranksters can occasionally produce odd results."
She then explains how it's done...
"a number of webmasters use the phrases [failure] and [miserable failure] to describe and link to President Bush's website, thus pushing it to the top of searches for those phrases."
She adds that "we don't condone the practice," but that they are...
"reluctant to alter our results by hand in order to prevent such items from showing up."
They finally admit it. Google will admit further truths in the future, such as whitelisting and blacklisting. Why admit what they had earlier denied?
They had no choice. When ones starts to look worse than silly, one might as well admit what everyone already knows.
Ms. Mayer closes by trivializing...
"Pranks like this may be distracting to some, but they don't affect the overall quality of our search service, whose objectivity, as always, remains the core of our mission."
We are in the "admit -- trivialize -- not OK -- no manual change" phase of Google's evolution. We have arrived here after passing through "fake confusion" and "dismissal -- it's OK."
There is nothing here that was not known almost 5 years earlier. We now have a "prank" that can't be fixed in order to "stay objective." However...
Undermining the authority of the presidency is not a prank. It speaks to Bush's presidency that there was no national outrage over "miserable failure."
The lack of uproar is why there was no manual fix. President Bush was just a prank. Later in this review, we shall see that President Obama (and his wife, in a separate bomb) would not be called "pranks." As we shall see, they will receive manual fixes.
Does anyone have any doubt about the following statement?...
If Bush was a beloved president, if there had been a loud public uproar, that bomb would have disappeared, with Google exclaiming its deep respect for the highest office in the land (we shall see this "concerned good citizen" behavior in the future).
Why the Obsession Over "No Hand Changes?"
The manual removal of a malicious manipulation does NOT diminish the "core objectivity" of any search service. How much "objectivity" does any search engine have, really?...
An algorithm only has the illusion of "objectivity." It merely performs and delivers what has been programmed by humans, who are subjective.
Yes, Google is data-based. But at any given moment, the algorithm is not perfect. It is "beatable."
If Google wants a principle to stand upon, how about defending flagrant violations of Google's core guidelines, and in particular those guidelines specifically concerning links schemes?
If you won't defend your guidelines (those are "the core"!) from "determined pranksters who [can] occasionally produce odd results," who exactly will you ever defend them against?
The fact is that Google does not like to make manual changes. Ms. Mayer did not explain why. We'll learn the reason in the next installment.
Also in the next post, we'll see that they certainly do make hand changes... when a bomb's fallout hurts them.
Time Marches On
We are a little over half-way through our 10-year history. We have passed through 3 stages of Google's "Evolving Opaque Transparency" that have managed to avoid the big question...
"When will this easily identified manipulation be definitively fixed?"
All the best.