Wednesday, April 3. 2013Guardian Goggles need Biased Assimilation, not Homophily
On April 1st we noted the Guardian April Fools joke (that people put on Guardian Goggles to not see things/opinions they don't like) wasn't a joke. People bias their views without needing Guardian Goggles to do it for them. Since writing that, a paper published in Science Daily on March 27 (a few short days before April Fools day) came into Broadstuff Towers, explaining that old Sociological thesis of happy Homophily is bunk. Essentially:
The prevailing sociological theory, known as homophily, is that like seeks like. Those who have similar opinions tend to aggregate together and reinforce opinions that grow more divergent from the center over time. This is the echo chamber model that would seem to gain validation in the era of talk radio, cable news and the Internet. According to this theory, we are polarized precisely because we have greater ability to choose our social networks and news sources. We narrowly tailor our information sources by selecting them based on how closely they mirror our own tastes. The paper piints out that mathematical models that try to use homophily to explain polarization have come up short as most are based on something known as De Groot's model, which assumes that people form opinions in a way that minimizes overall disagreement within their network of friends and relations. In theory, an individual's opinion gradually converges to an average of those in his or her network. The flaw in these models is that they predict that opinions in society as a whole can only become more uniform over time, resulting in depolarization rather than polarization. Which clearly is not the case. What does explain things better is Biased Assimilation:
Apperently the research is being done to create better recommendation engines and online collaboration tools to help people find common ground on difficult and divisive societal issues, the front-runner approach seems to be systems that show how little difference there is between X and Y, rather than just highlight differences. An interesting fact is worth noting was, when looking at current recommendation systems: "The system that recommends the most relevant item to a user turns out to be always polarizing. The other two systems, which chose a random item liked by the user and recommends an item most similar to it, were polarizing only if the user was biased to begin with. It was surprising to find that biased assimilation provides a useful framework to analyze the polarizing effects of recommender systems. So, maybe Guardian Goggles would be better off laying out their Right-On texts alongside the most hateful Torygraph equivalent, and emphasizing the similarities instead, for all died-in-the-wool liberalistas to get the warm and fuzzies from maximium hand-wringing ![]() Wednesday, March 27. 2013The eclipsing of Google search by...Twitter?
I wrote about the Summly acquisition by Yahoo yesterday, and wanted to catch up on where it had got to today. Now, there are essentially 2 stories about Summly:
1. Yay! The made-for-media one: Smart 17 year old forms company in bedroom, 18 months later sells smart app to Yahoo for $30m If you search Google, you get 7 pages of Story 1 in various guises, only on page 8 is the first questioning article shown. If you search Twitter, within 7 or 8 tweets you get a far more nuanced view - it reflects the puff pieces Google has, but also views on what Summly actually is, it's background, why Yahoo may have done the deal etc etc - the sort of stuff investigative journalists once did. I think its a very interesting comparison between social curation and what is essentially now a "popularity" algorithm. Google's algorithm essentially treats quantity of links as a proxy for quality, whereas increasingly the link quantity is actually a measure of common popularity, not the same thing at all. Read Google, and you'd barely know anything about Summly because the first 7 pages comprise of press regurgitation and it has utterly failed at telling you anything useful about it. To summarise Google, you would think the story is:
(Hat Tip to Marco.com for that). But search Twitter, and you get a totally different story. Twitter, despite a reputation for being celebrity and inanity obsessed, is in fact - on the basis of my search anyway, far less so than Google. What is certain is that Twitter gave me a far fuller picture, within the first page I got, and, in this case anyway was the better search engine by far. Also, an interesting thought to ponder is where exactly the accurate measure of the sentiment and influence lies. Reading Google, you would think its a "Yay, Boy Done Good" world out there, and the Yahoo PR department has Done Good and deserves a slap on the back and a good raise. Reading Twitter, however, you get a totally different picture and you can see a far higher high proportion think Yahoo has not been the sharpest knife in the block, and qute a few people qurestioning what Summly actually had to offer. Totally different to Google. And which is more influential? Google records 2.24 million searches on Summly, Twitter is still turning over like a Geiger counter 36 hours later, if you average that at say a Twt every 2 seconds (say 30 per minute) and assume the average Twitterer has 50 users that's c 3.5 million people reached in that time. And the sentiment split on Twitter is very different, its still running at about 50:50 "Yay" vs. "WTF?". So according to Twitter, the Yahoo PR Dept shouldn't be reaching for the Champagne just yet. (Actually, reading the comments to the "Yay" pieces in the Mainstream media gives you a more balanced view too) But if I were hazarding an educated guess, I'd say that the Twitter audience were likely to be more influential, and probably a more accurate measure of overall sentiment out there. Which, going forward, makes the question of who has the better Search engine very, very interesting. Because that drives the Ad revenue. Update - was mulling this overnight, another factor is probably the amount of Search Engine "Optimisation" that Google suffers from, ie the entire industry dedicated to spamming Google search barey exists on Twitter today, the most we see so far (apart from teh pretty girl/link to url scammers) is "enthusiastic supporters" of various posters/topics/etc muddying the twittersteam sentiment with endless retweets apporobations etc of their hero (or endless knockdowns if they are opposed). No dount if Twitter became the default search mode this industry would go into overdrive on Twitter too. Wednesday, January 16. 2013Searching for the value in Facebook Graph Search
I had written quite a long post on Facebook Graph Search and managed to accidentally delete it before saving. Frustrating, but in restrospect I realised my whole post boiled down to this summary:
1. It searches your social graph and their likes to get a (hopefully) more relevant set of results than Google's "Whole World Wide Web" link heirarchy does (good summary here on Search Engine land) Now to be sure, the Google method is showing its age, what with continual SEO activity subverting its link values, and popular culture ensuring that the most popular link is seldom best, most valuable or even correct (see Gresham's Law of Bad Information Driving out Good), and using likes instead of links is clever - but is this going to be any better? Not really, or to be more accurate, not as yet. To use the tired old use case that hucksters of new technology for the last 20 years always come up with, i.e. finding a great restaurant in a new town like, oh, say San Francisco, and to outperform Google (or just any old recommendation website), it has to go far farther than my own social graph (as only a few people on my own graph live there in SF), to pick up people who don't know me, but who do live in SF. It also has to have these people record the great restaurants they all like in SF, as well as tell FB lots about themselves so it can persuade me that a search via those people is more relevant than me doing a quick Google or TimeOut or Zagat or whatever. And that means opening up people's data to people they may not want to open up to, and also getting them to collect data they may not want/be arsed to do. And that is a very big ask.....(In fact the relatively low takeup of Foursquare et al suggests few will willingly volunter lots of location data - cue breaking open the Smartphone location sensor data...) (Update - I see Slate has said some roughly similar sceptical things) (Update 2 - ditto over here) *Not just the standard Tech Hype Pack, many IPO investors are desperate for a way to surface from being underwater too! Friday, February 25. 2011Google and the Amazing Algorithm Alteration
On January 2nd we wrote a post on "The Increasing Uselessness of Google Search" based on our experience on searching for commercial items over the holiday period. Matt Cutts of Google told us (and others who had noticed the same) that we'd never had its so good with Spam and later that month on TWIG a number of A Lister US Tech luminaries told us we didn't know what we were talking about. Roll forward to the 24 February, and Google announces the most brutal change to its search algorithm ever - Search Engine Land:
New Change Impacts 12% Of US Results What can we say except Told Ya So ![]() Actually, I do give Google a lot of credit for moving so far, so fast, and Matt Cutts did what he had to as a Corporate Droid - but it really is so frigging transparent, why no just 'fess up guys?. Anyway, Kudos! As for those A Listers, we know who you are....... and we're saying nowt more 'cos we know we'll mess up one day as well ![]() There are some who suggest this is PR as much as real action - which may well be true, but I think Google now really understands it let things rot a bit too far...... Sunday, February 20. 2011Techmeme, Twitter, Quora and mini link-farms.
Ryan Spoon wrote something I've been increasinglty thinking for a week or so about the New Techmeme, and it's pointing to Twitter and Quora:
Techmeme has been making an increased effort to move beyond blog posts by integrating Twitter and Quora conversations. Conceptually it is attractive, but figuring out how to cohesively merge the different conversation types is quite difficult. Techmeme will test their way into the right solution… and I give them credit for integrating Twitter beyond a side-bar widget (most attempts)… but I am not sure examples like this add value to the experience (other than getting to headlines very quickly): He goes on to show a conversation on the board that was a bit circular and unhelpful (the inevitable impact of pointing to a microblog and an aggregator if one is not carefu)l. Curiously enough, he then looks at yesterday's gerfuffle with various Uber-reaching Twitterclients, and felt they they had kinda got it right... And as it broke on Friday, there was a mixture of real-time commentary, news and updates from the companies themselves (namely Twitter / @Support and Bill Gross). ...which was interesting because just before I read his article I was putting pen to paper to observe similar frustrations to him, but I was writing about the frustration of navigating around the news item he felt was "getting it right". Having compared the two experiences I think I know why - when he was writing it was a breaking item on a sidebar (see below) Great, pithy early warnings. However, by the time I woke up in the UK, it had mutated to this: That's Techmeme the same as it's ever been, I hear you say.....amd on the surface, you are right - but now click on "expand discussion" on that big story from TechCrunch....this is about 1/4 of the expansion ( I have highlighted the twitter input in yellow, its nigh on half the input - btw I missed one, fourth from bottom): The issues I had were the following, from a user experience point of view - firstly, Twitter:
Secondly, Quora - I am not sure of the wisdom of an Aggregator linking to another Aggregator, because - in my experience anyway: (i) You sort of just get the same news review you have just read on Techmeme again (especially as many of the Quora participants have typically made a Twitter post that has already been picked up on Techmeme), and then you have to go scanning the Quora page to see something new and its just sloppy seconds of the same stuff (in this case anyway, maybe other discussions are better). But to take Ryan's point, clearly there is a need to bring the real time into Techmeme - my take on it is this:
But it is an interesting UI/UX question, and like Ryan, I think if anyone can get it right its the Techmeme team, but the fundamental value Techmeme has had to me is that it reduces, not increases the stuff I have to wade through...until now. . I shall watch with interest. Monday, February 14. 2011Google - an Akismet Almanac arms race
Normally when we predict things will happen its over months or years, not weeks - so when we predicted the inevitabiity of search spam filters early this year and also noted the things Google should to to get themselves back on track we noted that they should, as a matter of Class A Fundamentals:
Fundamentals - secure the next 3 years 3 weeks later and Lo, they announce the exact idea I highlighted above:
Kudos, though it is also more evidence that the will to act on spam has only really been there since a fairly strong user driven backlash since the New Year's superspamming of search results. First thing to cull - Demand Media, and as we noted, HuffPo had to sell, we think they saw this coming. I hope we can also share our culls with friends, sort of like an Akismet Almanac ![]() Personally I'm going to wait for Mozilla to do theirs because I don't like toolbars/apps that phone home to public corporations who are already playing fast and loose with user data. But net-net, this is a first move in a trend is going to create a major balance of power shift in the SEO spameconomy, because now: (i) Users will be able to cull sites they know are cr*p, rather than waitfor Google et al to (not) act.. Cue arms race in frequency-hopping URLS..... Now, lets see some movement on Item 2...... Sunday, February 13. 2011More evidence of the Google Slippery Search Slope
It's interesting - ever since we wrote about the "Increasing Uselessness of Google Search" after the 2010 Christmas holidays, it looks like a whole lot of Googleflaws are coming out. Today there is a report in the NYT of how JC Penney gamed them over several months in the holiday period. Essentially JC got some black-hat SEO guys to build a link-farm to boost their rankings across a whole lot of consumer product terms, and Google didn't spot it.
There are links to JC Penney.com’s dresses page on sites about diseases, cameras, cars, dogs, aluminum sheets, travel, snoring, diamond drills, bathroom tiles, hotel furniture, online games, commodities, fishing, Adobe Flash, glass shower doors, jokes and dentists — and the list goes on. Some of these sites seem all but abandoned, except for the links. The greeting at myflhomebuyer.com sounds like the saddest fortune cookie ever: “Sorry, but you are looking for something that isn’t here.” When you read the enormous list of sites with Penney links, the landscape of the Internet acquires a whole new topography. It starts to seem like a city with a few familiar, well-kept buildings, surrounded by millions of hovels kept upright for no purpose other than the ads that are painted on their walls. The thing is, these things are not new - anybody in the whole 'Net space knew about all this 5 years ago at least, those in Search knew about it for even longer. So the question you have to ask yourself is "how come it takes such a bright company like Google so long to not solve this problem?" . The Google Hypothesis is that this is all very hard, and there is a continual war with SEO guys, and the War On Spam is a long term slog, and besides their metrics tell them there is less spam and better results than ever. I think they are being somewhat disingenuous, or (more scarily) their metrics are plain wrong as they don't understand what "spam" really is. It's not that all of Google is corrupted, but our experience is that typically anything of mass market consumer value is gamed to the nines - and it's not as if one has to put any deep study in, its visible to the naked eye, as it were. It is blatantly obvious what is being gamed, and how it is being gamed, and has been for several years, and I for one cannot believe that it is not possible to develop approached to solve this for the 5% of the Googleverse that is taking 95% of this spam if one had the will (I believe they have the know-how). Non trivial, sure - but do-able, even if it requires human intelligence - especially over 5 years. So we have to look for the "why" there was not been the will - one option is that they have been so busy with all the other projects that the eye has been taken off the ball. Possible, and I suspect that's what they will tell us in about 3 months or so when more comes out the Googlewoodwork (There will be more, I am sure....). Another hypothesis is that Google, by a combination of intellectual comfort (no Search challengers till recently) and reliance on Big Algorithms, wasn't able to see/ build effective defences. But heck, competitors and decent New Search techniques have been around for the last 2 years now. But here’s another hypothesis, that also explains all the facts and complexities, the almost unchallenged arrival of mass Content Farm businesses, and win's the Patrick's Razor award - we will let the NYT run with it: Last year, Advertising Age obtained a Google document that listed some of its largest advertisers, including AT&T, eBay and yes, J. C. Penney. The company, this document said, spent $2.46 million a month on paid Google search ads — the kind you see next to organic results. The hypothesis that Google is colluding in all this as eventually a lot of the Ad money flows to them is clearly a heinous suggestion - but it does explain things rather well...... *Patrick's Razor - when it comes to money, the most cynical explanation is usually right Friday, February 4. 2011Google blinks...
Just after Christmas we wrote a post on the "Increasing Uselessness of Google search" which (if we may say so ourselves) helped kick off a furore which still continues, the latest episode this week being a debate between Google, Bing and newcomer Blekko - over to Vivek Wahwa, who emceed' it:
I had the opportunity to moderate a panel discussion this week between Google, Microsoft, and Blekko. The event, which I emceed, was called Farsight 2011: Beyond the Search Box, and was organized by BigThink and Microsoft. As I joked, it seemed odd that Google was playing the role of “evil” monopolist; Microsoft, the “good” contender, whilst Blekko was a fly on the wall. Cue media frenzy - but when Vivek says that Google has "changed the argument" away from spammy search and how Google's money is made, I think he is wrong - at least outside the Valley where maybe more bloggers are in the Googleshadow - here in the Uk it seems to be seen quite clearly. Here is the Torygraph: Google’s Search Fellow Amit Singhal complained via Twitter that Bing was copying Google’s results. That should tell you everything you need to know about the scale of this dispute: serious accusations of theft of intellectual property, even in Silicon Valley, are made in court. Here is e-Consultancy:
Puremango: When Google set up their experiment, they created special pages on their site which contained words which didn’t exist anywhere else on the web, then they installed Bing toolbar and clicked on links from those ‘synthetic’ pages. Bing toolbar sent MS that data, just like it would for any page, and their system incorporated that clickstream data into the other signals just as normal. Slashdot wasn't fooled either, as this exchange early on in the comments shows:
Conclusion - This is a Googleblink! (Update - Kara Swisher reckons it's got Larry Page written all over it - a blink indeed) Sunday, January 23. 2011What should Google Do?
Lots of stuff on the webz about what Google (i.e. Larry Page) should do to get Google's mojo back - much of it mutually contradictory of course, or requiring too many things at once. From a strategic perspective however, the oldies are still goodies - Fundamentals, Focus and Future, via the simple action of following the money:
Fundamentals - secure the next 3 years 1. Get the core search back to the quality it was a few years ago - this is the golden goose, the bread and butter, etc etc - if this drops away then everything else is moot. Sorry, Matt Cutts, if you believe your algorithms tell you it is better than ever before, then your algorithms are wrong. If I were looking to make Google search social, a good start would be letting people sharing spam-site killing data for filters. Crap search = fewer users, fewer users = fewer Ads seen, fewer Ads = less revenues, less revenues = misery. Those are the big things, the "A" class "Must Do's". Then there are the "B" Class things - are we going to throw more resource at them or cut our losses. The main ones are: Focus - stop doing all the things you do, especially look at rationalising, and - er - firing:
Future - where are the next things?
Simples ![]() Friday, January 21. 2011Not so In-Demand Media
Given Google's recent announcement that they are cleaning up the spam sites (see our article below) probably means that the "cr*p content" sites like Demand Media's ones will probably also not get top billing either, which willl significantly impact their IPO - and future profitability. Search Engine Land:
This announcement comes just as Demand Media gets set for an IPO. Demand owns eHow, LiveStrong.com, and several other properties that often get labeled as “content farms,” and is reportedly going to go public next week. AOL, with Seed.com and Yahoo, with its Associated Content purchase last year, are also in the content farm business. Those two, in fact, will be speaking on a panel called "Content Farms" Or The Smartest SEOs In the World? at our SMX West conference in March. By cr*p I mean unhelpful given its prominence - eHow's advice is muchly of the level of the Feynman Problem Solving system 1. Write down the problem. Still, live by the SEO, die by the SEO. It would have happened eventually. But timing, as they say, is everything..... Update - fascinating post by eHow's original owner Josh Hannah on why Google may rate it so highly: Google seems to weight domain-level credibility very heavily, and to not be very good at understanding if individual pieces of content are any good. And historical credibility, once earned, seems to decay very slowly. Tut - wash yore mouth out on that last one, you're wrong sez Google and its Fanbois
(Page 1 of 8, totaling 73 entries)
» next page
|
QuicksearchMore Broad StuffFor More Information about Broadsight:
Contact us Broadsight website Articles To sign up for Broadstuff on other services: Broadstuff - the Twitter edition Broadstuff - the Jaiku edition Broadstuff - the FriendFeed edition Subscribe to Broadstuff via email Books we are reading: Poll of the WeekWill Augmented reality just be a flash in the pan?
Archives Popular Entries
Categories
Creative Commons LicenceBlog Administration |