By now all of us will have taken for granted (while quietly marvelling at) at Google’s new predictive search facility. Admittedly I find predictive text a pain on my phone, but extremely useful when I’m trying locate files on my overcrowded desktop.
Now every time we use Google’e search box it tries to second guess what we are about to type. You may also have noticed that the suggestions are usually pretty close to what we 
are looking for. In fact these searches process on the page asynchronously i.e. updating the search page results as we are typing our search terms.
All this mind blowing wizardry is made possible by Google’s sophisticated algorithms and indexing sytems. But there is another intrinsic element in the personalisation algorithm that we often overlook, and that is US – our past choices, location, the browser we use etc etc.
But how much information is Google secreting away from our own online browsing habits?
According to Rene Pickard there may be a whole gambit of ‘signals’ (50+) Google is discreetly storing and then asynchronously returning back to our browsers when we search.
This might not be just confined to our online Google searching. More that a quarter of us are now using Google Chrome as our preferred browser, probably because of the way it seemlessly combines searching and web URLs in the same address bar space. So Google’s ‘personalisation’ has become part and parcel of our overall browsing experience.
It certainly appears to make life easier when it comes to navigating around the web – especially when page URLs are not easy to remember and when searching by a word is much quicker and easier to do.
Bui it does raise some quite important questions in terms of user subjectivity and Google’s own objectivity:
- How ‘personalised’ is the Google predicting/interpreting the algorithmic process?
- Does Google reserve a covert commercial stake in the way it prioritises its results ?
Before we get too paranoid about the Google Zeitgeist it might be worth a quick refresher on how search engines operate.
To try and illustrate (and running the risk of massive over simplification) Google’s original probablility algorithm has the browser’s actions at its heart of the calculation (see wikipedia article).e., it tries to predict ahead the liklihood of what you and me browsers on the web are likely to look at and in what sequence.
Over the years by increasingly fine tuning the algorithm Google has maintained its competitveness as the number one search engine market leader. The holy grail has been (or as we would hope) to return ‘clean’ search results faster than anyone else.
Undoubtedly the latest algorithms reflect this sophistication to the nth degree and to the point we have arrived at now i.e., personalisation (rather than from an assumption of user randomness – see original algorithm),.
But has the central philosophy changed? And have the lines between ‘clean searches’ and monetisation of results blurred beyond recognition?
I was recently discussing differences in search results based largely on another Google’s search innovation, Localisation – i.e. when we type in a product or service Google switches into commercial mode- paid listings are the top three results on the first page. Local businesses may well feature on the first page results, and these results will differ depending where we are located.
I have got used to discounting the top 3 paid listings in Google although I inadvertently counted these in recently because they are now almost indistinguishable from the unpaid listings (following a recent colour scheme change on Google’s search pages.) How much of this was a deliberate blurring of the distinction between paid (monetized or referred to as sponsored by Google) and unpaid (actual objective search results*) remains another of Google’s trade secrets.
(*objectiveness is also debabtable when we take into account optimisation methods by website owners)
Over recent years Google as the undisputed Search engine King does seem to have a adopted a carte blanche approach to collect and process vast amounts of our data world wide. The key to success has been monetising this information through the efficiency and speed of its algorithimic process, i.e. providing search results to you and me – the quicker this happens on the page the more likely we are to continue using Google services. It follows that if Google can turn more of this process into profit, the better its chances in staying ahead in an increasingly search orientated online environment.
Maybe we are happy to overlook and accept the subtlteties of the switch into ‘commercial mode’ when we are looking to buy online, but we may not be so comfortable with the concept of manipulating search results based on what Google predicts for us.
We all would like to think of a search engine as a service is free in every sense of the word i.e. searches clearly distinct from its advertising revenue stream, but as with everything else on the web perhaps we have become too accepting and trusting in the seemingly innocuous search engine, and to start to rethink of ourselves- our online relationships, conversations, wherabouts etc etc as central in the stored repository of online search data. There is no doubting thefact that the the web user has always been at the centre of the search algorithm. Only now perhaps the focus is increasingly on me rather than us.
And maybe we should not be expecting Google to sit on its ethical laurels while trying to hold on to its market position as the dominant search engine.
Further Info:
http://www.further.co.uk/blog/Google-uses-57-signals-to-personalise-search-when-you-re-logged-out-332
http://www.networkworld.com/news/2011/072911-how-cloud-computing-will-change.html?hpg1=bn
http://www.guardian.co.uk/technology/2008/jul/04/youtube.google
http://www.rene-pickhardt.de/google-uses-57-signals-to-filter/
For some time the idea of cloud computing has been floating around (scuse the pun) on the web as the next emerging technology set to transform the way we all work, play think and breathe in cyberspace.
