-----------------------------------------------------------------
==> Topic: Personal Interest Graph
-----------------------------------------------------------------
Originally posted in I-Search #153
Last Friday at World Economic forum in Davos Switzerland, CEO Marissa Mayer of Yahoo! (formerly of Google), spoke at length about her vision of the future of Web technology. As a noteworthy search influencer and considering social search, she spoke in terms of waves of technology innovation and described a future of discovery from an 'Interest Graph' in response to traditional and social search that we're seeing the beginnings of today.
Interest Graph differs from Facebook's Graph Search and Google+ in important ways. With her mention of Twitter, it seemed to me like her vision might be informed by Twitter Discover and her awareness of the media advantage that Yahoo! enjoys. She spoke of catering to 'daily habits' and how well Yahoo! is suited to provide for it in a mobilized digital age. Daily habits can, in turn, inform a collection of interests mapped by hierarchy to populate a map or graph: Interest Graph.
Google Now compartmentalizes this with its 'cards' while Yahoo! has a much more mature hierarchy of information buckets with user history and its legacy services (which are to be refreshed). The discovery process takes on push services like that of old but in an entirely new way. It is intriguing to me for the future of information retrieval. Where "you are the query" a query process might begin to seem largely unnecessary with Interest Graph.
The way I picture it, historical queries of yours can be collected in a portable format for use with any open service. Imagine if you could export your search history from Google and map it onto the Interest Graph at Yahoo!, with all its media strength and partnerships. You'd get better information. The Yahoo! ontology of "you" and your topic interests are informed by everything from your personal preference settings to tracked interaction with media on the network.
The brilliance of this approach is its anticipation of perfecting behavior targeting beyond what Facebook and Google could do any near as quickly as Yahoo! could. Privacy will always remain a serious concern until such a time when standards are enforceable by law that automatically comfort users. It may be a long time off but data as sensitive as your search history should be portable, transferrable as your phone number in today's US telecommunications market.
Search is already viewed by smart engineers as a simple commodity, available at a moment's notice and applied with an easily coded search text field. This assumption gets furthered when looking at Facebook's Graph Search, a product that is merely a collection of utilitarian search features specifically for Facebook's walled-garden. It is the data underneath any search service that makes up the value of a search engine.
The algorithm is a necessary utility where the discovery magic can happen. The limits of relevancy by way of encoding an elaborate algorithm may be at hand. Google has not solved the search puzzle and WolframAlpha is too esoteric for wide commercial use although it handles computations for certain Siri queries (I like their chances in the long run. Alternatively, it is possible for information to bubble up without user queries.
We are definitely moving in this direction with Google Now on Android and Twitter Discover. Google Now provides topical push 'cards' that populate information based on the context of where you are and what you were last doing. Twitter Discover shows Tweets based on the context of who you follow. Twitter is its own Interest Graph with constant updating user-propelled media pushed to your Discover page. I look forward to seeing implementation of something further enhanced than these services at Yahoo!.
-----------------------------------------------------------------
==> Topic: Facebook Search Engine
-----------------------------------------------------------------
Originally posted in I-Search #149
[an error occurred while processing this directive]Facebook announced today what some, including I-Search, predicted would be a social search engine. People predicted other things like car integration, improved gaming and gifts, to a wholly Facebook branded phone. Really? It was search that won the day. There might be some things Facebook left on the table, however it is clearly a powerful step to try and disrupt search. For now, it only consists of a series of filters and keyword search.
https://www.facebook.com/about/graphsearch
The Facebook Graph Search is a social search product that is limited to what you can get from Facebook itself. That means it is both a fix for the notoriously terrible pre-existing Facebook search experience and it can reach out to more of the data from your whole Facebook Social Graph, including Web page shares. That makes it just a glorified site search. The power is the data.
The irony in the newly minted "Facebook Graph Search," against a backdrop of Google Web search is the notion of them selling it as a search engine that respects privacy. Totally off the mark. It's better to address privacy directly, than to try and spin what amounts to a limitation. Facebook Graph Search only returns results from material shared with you. That's very limiting.
Facebook Search does not index the Web. That's actually quite awesome with one small exception. While it is noise diffusion device for the unwieldy Web, it makes Facebook Graph Search too much like a defanged Google+ Search, only one click away from Google Web search. Facebook is betting they have better data, different data in their social ecosystem than Google.
That's not going to remain true for long. Finding Web shares from a Facebook Graph-filtered search product is a search experience that might have made a huge leap in relevance, a leap already glimpsed at with Twitter's Discovery tab. Facebook is proving to be very carefully treading on this type of turf where Twitter is innovating. Facebook essentially introduced filter tools and that is it.
It is incredible to apply the Facebook 'Social Graph' experience onto a website that you are browsing. Find out what your Facebook connections shared, what is the most popularly shared content and find new connections this way. Facebook is bringing this capability home with a new series of filter tools, throwing in more than Web pages. The most useful features can be explored further in experimental builds.
What is amazing about the Facebook demonstration today, is its total lack of search sophistication. Filters aren't very exciting and had to be presented in a way that demonstrated what one can accomplish with the structured data of Facebook Social Graph, rather than set pretty higher at launch. Hopefully this will go beyond a search experience that amounts to little more than a series of filters at Facebook, and reach Bing.
While it's not sophisticated, it still points to potential future sophistication. Keywords are an area of concern. Facebook will use Bing Search for backfill. Will Bing get access itself? It is not only Facebook that is trying to innovate and disrupt search. Bing might play a more significant role with their first-hand search integration with the Facebook Graph.
Matches for search intent is clearly something Facebook and Bing want to improve. Facebook Search already asks for 'people, places and things' (in keeping with how people search by nouns.) People have been trained by Google to search for Web pages. Among elite search professionals, search is a task of finding corresponding Web pages. Twitter, Facebook, Bing and WolframAlpha all aim to get back to a core search experience instead.
There are gaps to address in Facebook's Social Graph, as Danny points out in his observations. He doesn't know if his dentist and plumber are even on Facebook, let alone if they qualify for a share to his SEO graph connections. Since SEO is an area of primary concern for I-Search, Facebook Graph Search SEO would mean increasing the popularity of targeted connections for Facebook pages, just like what we do with Google+ for SEO.
The more things change, the more they stay the same.
[an error occurred while processing this directive]-----------------------------------------------------------------
==> TOPIC: Mysterious Facebook
-----------------------------------------------------------------
Originally posted in I-Search #148
[an error occurred while processing this directive]Facebook is holding a mysterious event next week. There is plenty of speculation about what they could announce, everything from a new smartphone or tablet device. That is pretty far out and certainly is not likely. The buzz in news tends to be more about how to speculate in advance, with lists as headlines such as "Five Things Facebook Could Announce Next Week" (Forbes), in order to generate views for advertising. While this is nothing new, it's pervasively what makes up today's commercial noise.
I think in 2013 we'll turn a corner in online noise pollution. The sheer volume of inane whimsy people crank out for rankings and hits has reached a fever pitch that feels like it's responsible for global warming as much as anything. That may be a fun coincidence, through it all there are good pieces by people we still enjoy reading. It's the back-slapping herd of friends inmarketing who help spread buzz and everyone's happy.
Maybe a re-invented Threadwatch(.org) by Internet Marketing Ninjas will pull of "Less Noise, More Signal" in earnest. Who Knows? I-Search happened to get started up again at practically the same time. The year 2013 is starting off with a bang. The only caveat we can think of is, remember, every new year starts off with a bang. The road to hell is paved with good intentions. That's as true in online news as anything.
Facebook's Mark Zuckerberg has mentioned that brand-specific hardware is a terrible idea. There is more buzz about Facebook announcing a smartphone or tablet device, enhancing existing features or integrating Facebook with a car. These are bets placed by the writer who contributed the Forbes article. Nothing mentioned makes any sense at all. They are either safe bets on features that already exist (Instagram, gaming, gifts) or ideas that have a low likelihood (integrated devices or cars).
Our bet, though we'll make no prediction, is Facebook will announce a new search engine. The pieces about Facebook search are due to make their appearances on sites such as Search Engine Land soon enough. There is reason to believe Facebook could announce a search engine given that Mark Zuckerberg has in fact said a Facebook search engine is "inevitable." Couple that fact with the existing Bing integrated search experience and the need for Microsoft to stay relevant, it would not be surprising that they could be mentioned.
Time will tell. In the mean time, we're looking forward to a future of less noise, more signal. If Threadwatch and I-Search can pull it off, at least for social search marketing news we'll do our part in green technology as far as noise pollution is concerned. There probably is a very real carbon emission impact from not selling so much advertising fuel for buzz bull. Forbes is a source that may only get worse in this sense. For that, it may not be able to stem the tide of declining ad sales, as noise pollution burn out may arrive in 2013.
[an error occurred while processing this directive]-----------------------------------------------------------------
==> TOPIC: ChaCha Search Guides
-----------------------------------------------------------------
Originally posted in I-Search #088
From: SearchReturn <digest@searchreturn.com>
http://blog.searchenginewatch.com/blog/060904-044533
Essentials: In an approach to real-time social search, a new search engine called ChaCha has roughly 3,000 paid guides to assist in the search process. Guides basically open a chat conversation with you while helping you through the search process earning anywhere from $5 up to $20 per hour. The pay is based on a sliding scale according to user feedback rating each guide with a scoring system.
The approach with ChaCha is interesting because it can help new users that either go without answers to their query, or spend in inordinate amount of time with the learning curve. According to Inventor Scott Jones, users take an average of 11 minutes to find information using traditional algorithmic-based Web search engines. The real-time chat nature of ChaCha shortens the cycle of search and response for new users.
Questions of scale and competition from the major search engines remain. Although Yahoo! and Google answer services already exist, ChaCha was started by Indiana inventor Scott Jones who pioneered voice mail technology and created Gracenote, a technology used in iTunes. There is plenty of room for ChaCha to succeed and prosper in the marketplace.
Comment? mailto:digest@searchreturn.com?subject=PersonalSearch
-----------------------------------------------------------------
==> TOPIC: Yahoo! Upgrades MyWeb
-----------------------------------------------------------------
Originally posted in SearchReturn #069
New Discussion
From: SearchReturn <digest@searchreturn.com>
http://www.ysearchblog.com/archives/000310.html
Also:
http://searchenginewatch.com/searchday/article.php/3611306
Essentials: Several new features to Yahoo!'s MyWeb that appear to be significantly influenced by the strong direction the company took towards personalized search and social bookmarking, especially considering the purchase of Del.icio.us. Note the compelling commentary about the strength of tagging and search results within community bookmarks of saved pages, even with interests that are not as popular as CSS or XML.
Comment? mailto:digest@searchreturn.com?subject=PersonalSearch
-----------------------------------------------------------------
==> TOPIC: Personal Search Dead End?
-----------------------------------------------------------------
Originally posted in SearchReturn #019
New Discussion
From: SearchReturn <digest@searchreturn.com>
http://vivisimo.com/docs/personalization.pdf
Essentials: Due to pitfalls regarding how tracking search behavior for personal search intelligence is flawed, Raul Valdes-Perez makes a case for clustered results a la Vivisimo. A sobering read for those who've drunk the personalized search potion. When users want an Amazon effect for search though, it is no doubt starting to get offered by major engines.
Comment? mailto:digest@searchreturn.com?subject=PersonalSearch
-----------------------------------------------------------------
==> TOPIC: GOOGLE PERSONAL SEARCH
-----------------------------------------------------------------
Originally posted in SearchReturn #014
"I find it hard to believe that the search engines will go to these lengths for personalization but use it only on paid search. If they think they can improve the organic results through personalization, I think they will. Now what will *that* mean to"
From: SearchReturn <digest@searchreturn.com>
With Yahoo!'s MyWeb personalized search results, and Google search personalization coming out of beta today, this is exactly how things are transpiring. Nick from Threadwatch noticed this message from a Google search today:
http://www.threadwatch.org/images/googlepersonalizationui.jpg
Stay tuned.
-SR
Comment? mailto:digest@searchreturn.com?subject=PersonalSearch
-----------------------------------------------------------------
==> TOPIC: PERSONALIZED SEARCH
-----------------------------------------------------------------
Originally posted in SearchReturn #013
"Pushing the privacy line with post search behavior-targeting heating up."
From: Mike Moran <mikmoran@us.ibm.com>
One area to watch in personalized search (which I guess they are now calling behavioral targeting) is on the organic side. Everyone is focused on how much advertisers will pay for clicks that are demographically or behaviorally chosen, and that is an important topic, but what about the organic results?
I find it hard to believe that the search engines will go to these lengths for personalization but use it only on paid search. If they think they can improve the organic results through personalization, I think they will. Now what will *that* mean to search marketing? Goodbye rank checkers--there is no such thing as the #1 result anymore. Maybe now you want to know for what percentage of searches for that keyword you were #1, or what your average rank was across all searches for that keyword in a period, or your average within a demographic group, or... the mind boggles.
But even more interesting to me, is who will have that data. It won't be public information, the way rank checking is today. Only the search engines will know that information and I doubt they will give it away. So be prepared to have to pay the search engines for organic search--not to rank well, but to find out where you did rank. I don't know that this revenue stream will be anywhere near what the paid search money is, but right now (except for Yahoo! Paid Inclusion), the revenue from organic search is zero.
Mike Moran
IBM Distinguished Engineer & Manager ibm.com of Site Architecture
Co-author of IBM Press book, "Search Engine Marketing, Inc."
http://www.mikemoran.com
"I also like I-Search, with its old-school look and feel, and the intelligent posts."
-Andrew Goodman