Since I published last week’s story on Hillary Clinton & Google’s autocomplete, Google emailed the Washington Post about the matter. Here is part of the statement as published
“Google Autocomplete does not favor any candidate or cause, claims to the contrary simply misunderstand how Autocomplete works. Our Autocomplete algorithm will not show a predicted query that is offensive or disparaging when displayed in conjunction with a person’s name. More generally, our autocomplete predictions are produced based on a number of factors including the popularity of search terms,”
I have two issues with that statement.
1. It’s simply untrue
2. They are showing less popular terms
If the phrase is so disparaging why does it appear in pretty much every English speaking version of Google, well the ones I’ve checked anyway, except Google.com?
And then there is the problem of the search phrase popularity. It is clearly more popular than at least one of the phrases they are choosing to show in Google.com. Here is Google’s own data to prove it. I have deliberately excluded June data as it has been tainted by the Sourcefed story.
Google’s statement makes no sense, as there is clearly a more popular phrase been searched for in the USA that Google is not showing in autocomplete. Why it is not being shown I have no idea but Google’s explanation simply does not cut it.
Now have a look at those Australian search volumes. Google is choosing to show phrases that have zero search volume to users of Google.com.au and the popular ones as well. So even if the “disparaging” phrases had no search volume in the US, which they clearly do, Google may still choose to show them.
I have no idea. I think the institutional claims of Google bias are a little far fetched, I suspect something a little more subtle. I can’t believe Google would be so dumb as to manipulate .com results but not other ccTLDs like .com.au . I’m not even sure how you could “SEO hack” that result.
Jim’s been here for a while, you know who he is.