Mark Pesce wrote a column in El Reg this week relating the story of how a friend of his had stumbled upon an example of the phenomenon noted by social media researcher Danah Boyd and Microsoft’s Michael Golebiewski as a “data void”. The particular void that Pesce’s friend had entered was around details of the accuracy of the polymerase chain reaction (PCR) test used for Covid 19. When searching for definitive, and authoritative, information about this test, Pesce says instead he got back a “tsunami of conspiracy and antivax propaganda.”
Pesce goes on to say:
That a data void exists on a topic as important as the accuracy of PCR tests highlights the weakness of algorithmic approaches to indexing the Web, something we’ve been doing in one form or another since DEC launched AltaVista twenty-six years ago.
Pesce’s article links to a NiemanLab post by Josha Benton discussing the dangers that data voids pose for manipulation by those seeking to spread disinformation. Benton’s post makes for interesting reading. He tells the story of the time in 2004 that an SEO consultancy posed a challenge to SEO manipulators to get a (then non-existent) phrase to the top of the search engine rankings. Benton notes:
The idea was that, at a set time, the organizers, a company called DarkBlue, would announce a phrase that didn’t appear anywhere on the web. From that point and over a few weeks, competitors would try their best to make one of their webpages be the top Google search result for that phrase. To the winner went glory and the aforementioned iPod.
When the phrase was announced — “nigritude ultramarine,” a play on “DarkBlue” — swarms of SEO types filled the web with it and tried their own techniques to get their version to the top.
Benton goes on to note however that:
But the winner of the SEO Challenge wasn’t a black-hat consultant or a stack of Macedonian children in a trench coat. It was a blogger. Anil Dash, who at the time worked at the blog company Six Apart.
Dash was not, apparently, a fan of SEO manipulators (I like him already) considering them to be “barely above spammers”. So, according to Benton, Dash’s approach to the challenge was to:
put up a single blogpost, headlined “Nigritude Ultramarine,” and ask his readers to link to it. Many did. And that combination — the Google cred Anil had already built up as an active blogger plus all the PageRank-boosting “votes” created by those links — was enough to beat the most dark-arts-savvy SEO consultants in the world.
Benton concludes his article by saying:
There’s lots more interesting stuff in the paper, including a section on manipulating recommendation engines (*cough* YouTube) that I didn’t touch on here. It’s all worth a read. Not least because these behaviors can have impact for a looooong time. Fifteen years later, Anil Dash’s post is still the No. 1 result in Google for “nigritude ultramarine.”
That last sentence may have been true in 2019 when Benton wrote the article, but it does not appear to be so now. However, I particularly like this story because nearly twelve years ago, in June 2009, I posted a nonsense phrase as an experiment to see how long it would take before it was indexed. The phrase was “wild xenomanic yiddish zebu”. To this day, my post with that phrase is still number one on google.