Google and Spin 2

Two years ago, Prince Charles was subject to some unlikely allegations of sexual misconduct. Instead of turning a blind eye, he issued strenuous denials to all news media. At the time, some people thought this unwise - it merely drew attention to the allegations.

But the effect of the denials was to swamp the Internet. It became almost impossible to find the original allegations, because a Google search merely came up with hundreds and hundreds of denials, repeated by every news channel in the world. Even if you eventually reached the allegations, you would almost certainly have seen (and been conditioned by) repeated denials, together with many statements undermining the credibility of those making the allegations. In my post Google and Spin, I suggested that this was in fact a clever bit of news management by the Prince's staff.

I also noted that Google search wasn't always even-handed, but could be strongly biased towards those advocating a particular technology with a marketing budget to support their efforts, making it hard to find sites that were sceptical or critical. The example I cited then was biometrics. Laurie Sullivan of TechWeb News suggests that a similar thing is now happening with RFID. RFID Firm Turns To Google For Pro-RFID Campaign (November 1st, 2005)

It seems that the pro-RFID campaign has a large budget explicitly for information warfare, with Google as the battlefield, and is prepared to spend heavily on sponsored links, which presumably bias the search rankings in its favour. The pro-RFID campaign has made no secret of its hostility towards an anti-RFID book called SpyChips, and would doubtless be happy if this book received very poor search rankings. Nobody has suggested that Google would ever take money for suppressing an inconvenient website; the search algorithms are not disclosed, so we must trust that Google would never abuse their position of power. But the effect can probably be achieved anyway without any moral hazard for Google - simply as an emergent effect of positive sponsorship and marketing spend.

Trusting Commons

Can we trust Wikipedia? Does Wikipedia give us the truth? James Governor applies the Invisible Hand argument to this question. He points out that there are perhaps just as many errors in traditional encyclopaedias, and mentions both Grove and the Encyclopaedia Britannica.

In our writings, we identify four types of trust. The Encyclopaedia Britannica pretends to offer a form of trust based on authority, but it is in fact a commercial product and relies ultimately on commodity trust. In contrast, Wikipedia offers a form of trust based on the network. Is that good enough?

Instead of asking the binary question: can we trust, yes or no. The key is to ask the discriminating question: for what kinds of question does Wikipedia offer good/better/best answers? Clearly if you want uptodate information about an important but relatively nonpolitical event (such as the Tsunami). the Wikipedia is a pretty good resource. But if you want impartial information about a controversial topic, you may need to exercise caution.

The question of trust/truth in relation to Wikipedia is similar to the question of trust/truth in relation to Google. Can we trust Google? What I'm asking here is not a question about the commercial ethics of the company, but about the use of any Internet search engine as a method of obtaining a true and fair picture of a given topic.

A Google enquiry suffers from some of the same problems as a Wikipedia enquiry. If there is a consensus, you will get the majority view. If there is no consensus, you will get a confused view. You often find the same stupid ideas and ill-informed opinions replicating from one website to another. If there is a coherent minority view, you may never find it. And if you want an original view, you may need to disconnect from the Internet and do some real thinking or discussion instead.

More on Google

Google and Spin

News management and opinion management requires careful google-spinning

originally posted November 13th, 2003

---

The marketing value of search engines has been obvious to many for ages. If you are selling spangled wangles, you want your website spangleplanet.com to appear near the top of any google search. Webmasters are constantly looking for new ways to trick google into giving them a higher ranking than they deserve.

As people (including journalists) rely increasingly on googling to find out what is going on, then news management and opinion management also depend on google spinning. There is often a systematic bias in the material that is presented to the googler. This bias may be contrived, or it may be an inevitable consequence of the economics of truth.

Here are some examples from my own googling over the past few days.

The British home secretary recently made an announcement about biometrics. When I searched for biometrics on Google, I found hundreds of pro-biometric websites, from vendors, industry consortia and government bodies. There are undoubtedly also websites that present a more critical stance on biometrics, but these don't rank highly on Google, and I didn't have the patience to page down and find them. (As it happens, I had other ways of reaching the material I wanted.)

A member of the British Royal family recently denied some allegations that had been circulating on the Internet. If you search for this story now, you are presented with a lot of news items reporting the denial and discrediting the source - before you get to the allegation itself. This creates a different context for the allegation, and many people will read the allegation differently as a result.

A few days later, the British minister for children unwisely and perhaps inadvertently attempted a similar tactic, when she described one of her critics as "extremely disturbed". Many years previously he had been a victim of child abuse when living in a children's home for which she had had overall responsibility.