@T: ALWAYS THE TWAIN

Today I read this in the paper version of the ‘London Evening Standard’:

There’s no interest on my part about whether Google’s European PR chief is / is not close to any candidate for the Labor leadership. However, I am interested in finding out how Google’s strategic alliances and partnerships in Europe are working out and what their perspectives are on the technology horizon and business models. This is what the Google Zeitgeist series of conferences are all about: showcasing how well Google’s partnerships (including strategic investments) are playing out.

Ergo…..once online, I decided to see whether there was any more information available about Google’s Zeitgeist 2010 conference so I googled “Google Zeitgeist 2010 Eric Schmidt” and “google zeitgeist europe 2010 eric schmidt”. The automated options feature offered the possibility of viewing the results on standard basis (Wonder wheel, Timeline and Page previews) like so:

Since Wonder wheel was unfamiliar as a tool to me, I decided to take a look and this is what it produced: a Mind Map-esque visualization of clustered topic classifications:

Here’s a YouTube video from last May 2009 of a Google engineer explaining how the Wonder wheel works for the search term “stir fry”:

Ok, so that looks pretty cool and neat, right? A Mind Map-esque topic clustering approach to search. Cool except for three FUNCTIONAL limitations:

(1.) In no view did it actually surface any links to “Google Zeitgeist 2010 Eric Schmidt” which was the term searched for.

Under the usual list format, the top PageRanked link was ‘Google chief hints at partnership with Twitter’ by the Daily Telegraph from………May 2009!

Under the Timeline view, listings were provided in order of 2008, 2009 and 2010 with the entries in 2010 actually from 2008 and 2009 ===> date relevance and data quality control issues in the Google algorithms are apparent here.

Under the Wonder wheel view, THERE’S NO DATE REFERENCE AT ALL! So it clusters terms like “eric schmidt twitter”, “eric schmidt larry page sergey brin” and “ceo eric schmidt” without simultaneously time ordering these clusters to enable us to discern which ones relevant.

(2.) In no view is there any indication of the quality of the linked to content.

(3.) In no view is there any ability for the user to treat the Mind Map-esque visualization tool like an editable wiki, so that we can actually topic cluster terms according to our OWN PERCEPTIONS and contextual order of terms.

For example, when I google for “Google Zeitgeist 2010 Eric Schmidt” I want this information to either appear in the Mind Map-esque Wonder wheel or be editable according to my needs:

* where the Google Zeitgeist 2010 is taking place (including interesting landmarks in the vicinity);

* who of Google’s strategic partners will be attending, providing keynotes and panels and some bios of these people;

* when each session will commence and how long they’ll last;

* content of each session and implications for the technology horizon;

* how much the event will cost (or even information about it being free to attend but by invitation only); and

* what are the key take-away points likely to be from Google Zeitgeist 2010?

At the moment all that the Wonder wheel gives are these clustered terms (in a clockwise direction from the top):

* google ceo

* google zeitgeist europe 2009 eric schmidt

* john mickelthwait

* nikesh arora

* sergey brin

* sir richard branson

* fireside chat

* larry page

The conclusion would be that the Wonder wheel is an interesting visualization of search terms but the same limitations that persist in Google’s standard lists persists in this snazzy visualization and attempt at topic clustering. These limitations are:

(1.) Relevance — including date.

(2.) Acuity / accuracy.

(3.) Context.

(4.) Personalization.

(5.) QUALITY metrics for the content — not simply the binary-based frequency count or probability proximity of content associations (Bayes, group sets and cellular automata, in “maths speak”).

So…….as I noted in an email exchange with someone at MIT Collaboratorium…..topic clustering is an evolution on from time stamping content for search term surfacing, but there’s still some way to go before machine code works like our natural organic brains and the way we classify, orientate, prioritize, contextualize and synergize information on a personalized quantitative and importantly QUALIFYING basis for added value.

There is a wheel tool in 360-2020′s plans but it looks and works nothing like a Mind Map or Google’s Wonder wheel. Today was the first time I’d happened upon the latter and it’s just lucky that I have a completely different approach to Google’s Wonder wheel so that, hopefully, those 5 persistent limitations I note (which are applicable to most search engines) become……..increasingly no more.

Ok, so……….onwards with the Twain approach……….(since the tools I want don’t yet exist and it doesn’t matter how many MIT PhD Google engineers are working on the Wonder wheel because they’re still producing those persistent limitations I can spot — regardless of how “pretty” the visualization tool makes their search lists look).

LOL, :*).

I just read these comments from Bob Iger, CEO of Walt Disney:

· http://blogs.wsj.com/digits/2009/07/07/sun-valley-paid-content-on-the-agenda/

That’s consistent with Rupert Murdoch’s plans at Newscorp:

· http://www.hollywoodreporter.com/hr/content_display/technology/news/e3i22db2de2f279e316e965eaca55ef769a

· http://www.readwriteweb.com/archives/rupert_murdoch_lets_charge_for_online_content_again.php

Now this may not be obvious but paid CONTENT is different from paid SEARCH LINKS where Google dominates almost 70 percent of the global market. There’s some concern that the public will not pay subscription charges to read articles, view videos or the right to participate on discussion forums.

Personally, I think ordinary consumers would pay if there’s a PURPOSE, CHANNEL FILTER and CONTEXT to that content. As a consumer I wouldn’t pay for content that was simply mindless noise or dross. I’d pay for the content if I needed it to complete a project, though. For example, I’m prepared to pay for substantial quality McKinsey / DataMonitor / Jupiter Media Matrix reports; I’m not prepared to pay for tweets by Ashton Kutcher on Demi Moore’s derrière, although others would pay. That type of content is harmless and purposeless banter. Now contrast this with Demi Moore showing women how to apply make-up in a YouTube video. THAT is content some women would probably pay for if they wanted to look like a Hollywood A-lister (and there’s your context — LOL).

PAID CONTENT = PURPOSE X CHANNEL FILTER X CONTEXT