On Bubbles and Magic Words….

As law librarians many of us teach our users the best approaches to using the internet, which includes how to best construct a search. It may seem pedantic to talk about operators and specificity, but it does sometimes help to narrow the results to a manageable number. We know that most people do not look beyond the arbitrary 10 results on the first page of a Google result screen, so it is pretty important that those ten are the right ones. We teach this stuff on the assumption that we are working on a level playing field with the various search engines. We also know about algorithms and that they can make a difference. But it is the ever expanding manipulation of the search results that raises alarm bells.

In May, Omar Ha-Redeye wrote in Slaw about the ‘Anti-Google search engine’, Million Short, a search engine that strips up to a million results from a search on the premise that there are many spammy sites out there gaming the Google algorithms. This was a good heads-up to the changing world of search, and the efforts by many small entities to counter the work of the giants.

A recent blog by Gabriel Weinberg, creator of search engine DuckDuckGo, discussed the way Google tailors the search results you receive, based on who you are. From their own research, they write that when searching for raw information on Google, the results it returns appear to be constrained by the ‘filter bubble’ that is created by Google personalisation, and by the creation of ‘magic words’. Because of the timing of the article, and the use of Obama as one of the test words, the idea of ‘magic words’ was further investigated and written up in the Wall Street Journal on November 4 . Subsequent comments on that article then veered off into the political sphere of left versus right.

It becomes a concern when what is still purported to be a search site is as dominant as Google. The results of survey site Stat Owl gives Google 81% of the search market, followed by Bing and Yahoo. Others say that Google’s share of the search market is 66%. Either way, that is dominant by any stretch of the imagination. A paid search study in Search Engine Land claims that Yahoo/ Bing growth rate is greater than Google, but that the click traffic on Google is outpacing the Yahoo /Bing growth.

I may be a Pollyanna, but I thought the Google Settings page in Chrome page allowed the user quite a lot of control over the searching, adverts and cookies. But because I am not mathematical, and I do not understand the nuances of algorithms and the advertising side of Google, I read articles such as Weinberg’s and become a touch paranoid about what I am using to search with, what results I will retrieve, and how the results appear.

I remember the heady days of 1998 when Google first appeared as a new search engine, eschewing the onscreen advertising that was starting to creep into the 90’s versions of search engines AltaVista, Excite, Yahoo, Lycos, Netscape and Ask Jeeves. Librarians rejoiced because we are all a bit hippie at heart, and loved the idea of a pure search world untainted by commercialism. It seems ironic now that the underlying raison d’être of Google is no longer plain old ‘Do no evil’, but monetisation of people’s basic curiosity. In fact the mantra has evolved into item 6 of Google’s Ten things we know to be true : You can make money without doing evil.

Am I naïve to still think it is important that search engines should not distort, twist or engineer information? I like my facts or results unadorned, unfiltered in the first instance, based on my query. I do not want to see the most popular or most highly ranked result for my search based on a mathematical analysis of my previous search behaviours. If I want crowd sourcing or majority rules results, make that an option for me, not the default result.

The most insidious thing is that it seems to exploit the many people who just use Google to search for anything and everything. They don’t care about all the background shenanigans that exercise the mind of nerds and librarians. Quite simply, it is not fair on the public, not because it is secret – it isn’t – but because what is being done to retrieve the results to your simple query is so complex, the majority of us can’t understand it. I know there are millions of things we don’t all understand, from jet propulsion to the workings of our own brains, but until recently the tools we had to find out information were not engineered, they were overt and open. We knew the owners of newspapers, of publishing houses, their politics and biases, and if we wanted to, we could avoid or ignore those influences by choosing equal alternatives.

It will be some time before there is another phenomenon such as Google that comes along and takes up the mantle of being the global search engine of choice, providing search results that are not in the filtered bubble. This is despite the ongoing creation of many small search engines doing just this, but only available those who know to seek them out.

Now, unless we are savvy, we are hostage to result filtering without necessarily being aware of it. I’m not convinced this is a good thing.

Retweet information »

Comments are closed.