Google’s results no longer in denial over “Did the Holocaust happen?”

google-truth-lies-real-fake-ss-1920

Google’s been under intense pressure to alter its results after it was found a week ago to be listing a Holocaust-denial site first for a search on “Did the Holocaust happen.” Now, that’s finally changing.

It’s not clear whether this is due to a change by Google to its ranking algorithms or by the efforts of external parties to influence the results.

I’m checking with Google, but my bet is on the latter. Google has previously said that it wanted to address this search and similarly egregious results for some other searches but that the process would take time. Still, Google might have made a quicker fix.

On iPad, denial site bumped from first place

When searching on my iPad this evening around 1am ET, I found that a page from the United States Holocaust Memorial Museum had moved to the top free listing, knocking a denial discussion forum page from Stormfront that had been there into second place.

c0gsdb7ucaa5reh

I was not logged into Google and using incognito mode within Chrome, so my personal search history shouldn’t have been influencing this result. I also saw these results on desktop, when I simulated using an iPad.

On desktop, denial site also moves

Initially on desktop, the older results were showing still showing for me, as was the case with mobile results on my iPhone (again, logged out of Google and in incognito mode):

did_the_holocaust_happen_-_google_search

But by 2:45am ET, my desktop results had changed to move the USHMM page first (my iPhone ones had not):

google holocaust happen

I suspect we’ll see what’s happening for me begin to spread to all devices and for more people. It can often take time for ranking changes to move across the whole of Google, because of the many data centers it is using. But I’m not alone in seeing them:

New sites aim to alter the results

Also as part of the change, a new site is showing up called Did The Holocaust Happen. It was specifically designed to get into the rankings for this search and take on the denial sites. For me, it was appearing in position seven on iPad and desktop:

did_the_holocaust_happen_-_google_search

Later after initially spotting it, it has bounced around. Sometimes I’d see it only on iPad.  One time, I saw it rank as high as second in desktop search results.

Another site I learned about tonight hasn’t yet seemed to have gained ground into the listings.

Why the change?

As I said earlier, I don’t think this was a deliberate change by Google. In particular, the denial site remains in the listings, which is a sign it hasn’t been penalized. That would have been problematic for Google, because it really doesn’t have a good policy to pull or penalize sites for this type of reason (except say in Germany, where there are laws against Holocaust denial).

If it sounds crazy that Google couldn’t just pull a site pushing the horrible fiction that the Holocaust never happened, I strongly encourage you to read my story from last week on this topic. Google pulling sites for not being able to prove things could lead to it having to do stuff like banning religious sites, which are built on the foundation of faith.

Google needs a comprehensive, defendable policy. It needs time to develop an algorithm that can better cope with the amount of “post-truth” content that’s been growing. It also needs to work across a multitude of queries, not just whatever happens to get spotted.

Still, it could be something Google has done through a short-term fix. After I initially wrote this article, sometimes the results would bounce around somewhat dramatically, which is often indicative of some type of general Google ranking change. Again, I’m checking with the company on this.

How third-parties change Google’s results

If Google didn’t make the change — and even if it did — what external third-parties do has an impact on its results.

For one, as people have reported on this story, starting with the Guardian that first wrote about it and followed by others, those news stories have gained ground. That’s natural for both Google and Bing. Fresh content, especially from news sites, is often rewarded on a short term basis and ranks well.

The writer for the Guardian who has been tracking this issue also did a follow-up story about buying an ad against these results, to jump above the denial site through paid listings. However, she’s not the only one. The aforementioned US Holocaust Memorial Museum has also either been buying ads or someone is buying them for the museum, as you can see in my first example above.

That ad is almost certainly NOT causing USHMM’s free listing to rank better. Google strongly denies that ads influence rankings this way, nor has there ever been any serious evidence that it has an impact.

However, having the ad there could be causing more people to realize that the museum should be listed first naturally, for free, and begin linking to it. A rise in links, especially from authority sites and with the right type of textual context, could have an impact. The right links, the right way, can effectively act like votes to improve ranking.

As for that new site that’s appearing, it was built by an SEO — someone who knows search engine optimization — specifically to do well for this search. It’s like hiring a PR firm to deal with bad publicity. PR people know how to push for good press. SEOs know how to push for good search results.

That SEO was John Doherty, who’s well followed by other SEOs on Twitter. They, in turn, likely found ways to push links and promote the site for the good cause. Remember that the next time you hear that SEOs are all scum-sucking evil doers.

Doherty and SEOs in general have no superpowers. They can’t guarantee a favorable change in rankings, any more than a PR person can promise a good story in the press. But they have great knowledge that can improve the odds, and that’s what I think is happening here.

The challenge of the infrequent search

It’s also easier to happen because frankly, this isn’t a popular search. Very few people do it. The Guardian reporter who started running those ads, tapping into Google’s own data, found it happened about 10,000 times per month or about 300 times per day. In contrast, people search for “holocaust” worldwide around 15,000 times per day. Google itself handles 5.5 billion searches per day. In short, this search rarely happens.

Since it doesn’t happen that often, it’s easier to impact the results. Rare searches often have less content relevant for those top spots. In fact, one of the reasons the denial site has probably ranked well and for so long is because practically no one who would be concerned about this happening has done the search to even notice, nor notice to the degree of creating content to combat it.

Why does Bing get it right & Google get it wrong?

Of course, it is weird, disappointing and disheartening that on this search, Google wasn’t getting one of the good, authoritative anti-denial sites that were listed second and third into the top position. That’s especially so given the oft-maligned Bing managed to do it and still does, showing Wikipedia’s “Holocaust denial” page first among the non-news web listings:

bing holocaust

As my story last week explained, there’s some speculation that Google’s results are different because it’s rewarding clickthrough behavior more heavily. IE, if people who do this search click a lot on a particular page in the results, that could move the page higher in rankings IF Google operates that way.

With this particular search, perhaps most doing it are already in a Holocaust denial frame of mind. If they see a denial page in the listings, they might click on that more than anti-denial ones. And with the clickthrough theory, this causes the denial page to move higher.

The problem with this theory is that Google has been steadfast in saying that clickthrough does not directly impact its rankings like this. But, it could be that some of the clickthrough behavior is being indirectly mined by Google’s machine learning RankBrain system in a way that is causing these results and others to move some pages higher than with its old system that depended more on links.

That can cause some people to wonder why Google might not shutdown RankBrain or shift back to depending more on links. But links have their own problems and can be gamed, to the degree that some exploits even became known as Google bombs.

In addition, RankBrain probably helps improve many results that are far more popular than this one. Google’s challenge is to fix its system so that whatever is working well for more popular searches isn’t causing results that are one-sided, far-fetched or biased for less popular ones.

About that supposed right-wing bias

By the way, you might think that Google’s problem has been with right-wing bias. That’s what another Guardian article wrote, that Google was promoting information with “an extreme rightwing bias.”

The only bias really was in that article itself, which didn’t do searches to see if Google was perhaps also showing an extreme left-wing bias. And you could make that argument with things like this “white people are stupid” search:

google white people are stupid

Or with this set of “white people are inbred” results:

white_people_are_inbred_-_google_search

Sure, people on the right-wing aren’t all white. But these certainly don’t give the impression that the alt-right is somehow in control of Google.

Nor did that article bother to note that if you want to cherry-pick infrequent searches, it’s not hard to find Bing suggesting objectionable searches and then delivering actual objectionable results in response. For example, “was the holocaust a hoax” on Bing lists in the first position a page that says it was:

was_the_holocaust_a_hoax_-_bing

I point this stuff out not to excuse Google. It’s the leading search engine, proudly touts the high relevancy of its results and fails in that goal when listing a Holocaust denial site first for the “Did the Holocaust happen” search.

But it’s not just a challenge for Google. It’s not just something that has a right-wing bias. It’s an overall search challenge, one that really few have noticed until our attention has been focused on it, as people grapple with the growing concerns of “fake news” and a “post-truth world.”

Solving for the long-term

Coming up with an overall fix for this is something I expect will take weeks, and when it arrives, will likely be in the form of how Google tackled the challenge of “content farms.” There, it took on filtering out low-quality content from its results through what became known as the Panda Update.

Having met with people at Google last week, as I wrote, they definitely want to make changes. It’s as top-of-mind as anything I’ve ever seen Google concerned about. But they want to make it in a deliberate and comprehensive fashion.

In the meantime, the results — like any search results — will change on their own based on external factors. Now that awareness has been raised here, that seems to be sparking change.

The post Google’s results no longer in denial over “Did the Holocaust happen?” appeared first on Search Engine Land.