Google & Twitter & Facebook, Oh my: Does the organization of content change?

by drm on April 4, 2009

The net’s self-referring neural networks have been flashing white hot over the past few days about talks between Google and Twitter. These talks recall the Twitter/Facebook talks, and Facebook/Google talks. Observers and participants buzz with excitement at the prospect of the Net’s ecosystem undergoing a climate change.

Why would any of these combinations mark a such a change?

Google’s search hegemony had been based on the marriage of a sophisticated way looking at activity on the web  to a simple consumer interface, that has generated dominant usage and created cost-effective opportunities for marketers to intersect with consumers looking for specific things.

Picture 3.pngThe economic benefit of this combination is intimidating: $117 billon of market cap, nearly $9 billion of cash and a culture that mandates innovation and experimentation.

By contrast Twitter and Facebook are smaller and under-developed businesses, both in terms of realization of a business model and, objectively, consumer reach. Twitter particularly has a fraction of the traffic of Facebook, never mind Google.

Looking at Quantcast‘s analysis of usage engagement also shows that Twitter attracts a core group of passionate users and a high number of passers-by, who must be checking out this buzz-ridden phenomenon and then moving on.

The attention being paid by Google to Twitter speaks to the soft underbelly of Google’s fierce domination of the information world on the web.

Picture 1.png

Google established a search engine that could create value and scale more easily than any other existing engine by focusing on relevance and usage. They built algorithms that looked at the content on each page for consistency and relevance, and ranked the value of pages in terms of linkage. Who links to what pages is critical to a page’s value in Google. The underlying assumption is that if credible, high-quality sites link to another site, then that site must have good content, pushing it up higher in Google’s organic search.

Paid links drive Google’s revenue, but organic search drives Google’s usage.

In assessing Google’s long-term prospects, observers typically consider the probability of another search engine emerging that presents more intelligent and better-organized search results. There is no real barrier to entry in developing a search engine: all of the content on the web is freely available. Analysts generally believe however that the embedded advantages of Google’s vast database of content, their culture of innovation and their overwhelming financial resources will allow the company to both stave off any emergent threats and to find other profitable applications of their underlying technology.

That assessment assumes that the way that people search for relevant content online will evolve linearly.

Human action, not technology interpretation

Google’s analysis of content relies on using the power of data processing to discern patterns in content creation and consumption that we aren’t able to see individually.

Ultimately, however, people bring more context, insight and expertise to their interaction with content, whether online or offline, than even the most powerful technology can create. That’s the brilliance of  fuzzy logic; we are parsing, selecting and connecting the things we experience and making leaps of logic that search can’t replicate.

That’s also the genius of the editor. Some people have unique skills in terms of organizing and connecting content that is interesting and engaging to other people.

The best properties in media rely on the genius of editors and content creators: 60 Minutes, The Wall Street Journal, Vogue, the HBO lineup.

What has been missing from the web has been the primacy of the gifted information organizers, the individuals who have such a feel for a topic, the ability to tell a story, the patience and practice to gather and filter information in such a way that a topic becomes accessible and exciting.

Google’s economic power has been in using its role as an intermediary — because that is all that it is, an intermediary between content creators and content consumers — to disintermediate the content creators. Google offers marketers a way to intercept consumers in their search — offer them a sponsored information solution — before the consumer moves on to the actual sources of content that are relevant to what they are looking for.

Over the past five years or so, a number of different services have developed to leverage the human element in filtering content. Bookmarking services like delicious, voting services like Digg, ranking services like Technorati all give users resources to directories of human-ranked content.

These are supplemental and marginal services in contrast to Google. And, they don’t restate the basic interaction of consumers and content. Nor do they appear to scale in a manageable and efficient way. They are valuable, but they are complementary.

Twitter, Facebook and the organization of content

People create networks on Twitter, Facebook and other social networks. These are networks of interest, permission and sociability. The interest is established by the request of one person to link to another. The permission is implicit in the act of connecting: I am giving you permission to share your things — photos, links, updates, music, likes, dislikes — with me. I am sociable with you because I’ve expressed interest and given you permission to interact with me. It is an extension of the social compact that is established in interactions in the offline world.

When people create networks on a service like Twitter, they share content. And people are sharing more and more content.

This is different than finding a link on a search engine. Someone who you have a social contract with is presenting content that they find interesting, often with an annotation to make it relevant.

Every piece of content that gets shared suddenly has a rich set of discrete indicators about the content. First, it was selected to be shared. Second, some number of users will click through and interact with the content. And third, some number of users will share that content with their personal network. The more re-sharing that occurs, the bigger the network effect of the sharing.

This is an utterly different kind of way of organizing and interacting with information than is available in the traditional search engine paradigm. It is human, it is dynamic and it is current. It is the way that people use content in the real world.

Look at this series of Tweets from Business Week’s editor John Byrne. He’s using Twitter to transfer the collective  intelligence of the editors at Business Week to a group of interested followers. These followers will assess, organize and act on the information in their own personal way, and potentially distribute this content with authority to their own spheres of influence.

The same thing happens on Facebook. Not as dynamically, and not with as much independent content, but with more organization than the currently fast-flowing Twitterstream.

Content organized by people in other places is why Google cares

What matters to Google is that content is getting organized in other places with a higher degree of personal intervention and intelligence.

If Twitter and Facebook make a choice to further organize and lPicture 6.pngeverage this content they can potentially disintermediate Google.

I know it seems unlikely now. But you can see how this parallel universe of personally recommended content gets created.

And, I don’t think it can get gamed the same way that Google search can get gamed. Right now there is an entire industry of experts who devote their life to figuring out how to create just the right mix of words and phrases and links to improve the organic search results of their site.

This is not an activity that adds any value to content. It is an activity designed to figure out how to make content more meaningful to Google. That is the wrong focus, long-term.

With the content recommendations that flow through places like Twitter and Facebook, users are constantly assessing them for interest and relevance. If someone on Twitter starts abusing Twitter, you just “unfollow” them. You can’t do that with a Google result in organic search.

Michael Arrington at Techcrunch had a great post about Twitter as a search engine last month. I recommend looking at it. Here’s an excerpt:

I told [my friend] what I thought of Twitter as a micro-blogging service: it’s a collection of emotional grunts. But it’s wonderful nonetheless. And enough people are hooked on it that Twitter has reached critical mass. If something big is going on in the world, you can get information about it from Twitter.

Twitter also gathers other information, like people’s experiences with products and services as they interact with them. A couple of months ago, for example, I was stuck in the airport and received extremely poor service from Lufthansa. I twittered my displeasure, which made me feel better – at least I was doing something besides wait in an endless line. I’ve also Twittered complaints about the W Hotel (no Internet, cold room) and Comcast (the usual Internet gripes).

More and more people are starting to use Twitter to talk about brands in real time as they interact with them. And those brands want to know all about it, whether to respond individually (The W Hotel pestered me until I told them to just leave me alone), or simply gather the information to see what they’re doing right and what they’re doing wrong.

That’s why Google is there talking with Twitter. A group of about 50 innovators have built a really dynamic,  simple platform that is creating a new paradigm for organizing people and content.

Sounds like Google once upon a time, doesn’t it?

Related Posts with Thumbnails
Share
  • http://Www.Nstein.com OlegR

    Dan, great post!
    It got me curious (and sorry for a cheezy one) – according to your understanding, where does semantic search fit in between google and Twitter/facebook paradigms?

    Thanks

    Oleg

  • http://Www.Nstein.com OlegR

    Dan, great post!
    It got me curious (and sorry for a cheezy one) – according to your understanding, where does semantic search fit in between google and Twitter/facebook paradigms?

    Thanks

    Oleg

  • http://www.nci.com drm

    Oleg,

    Sorry for the tardy response. When I think of semantic search, I think of objects (lemma) that generate a higher level of semantic understanding in the analysis of the content on a page. In that paradigm, the user/redistributor himself can become an object associated with the content, increasing the ability to generate relevance around the content.

    And…that’s obscure.

    Thanks.

    Dan

  • http://www.nci.com drm

    Oleg,

    Sorry for the tardy response. When I think of semantic search, I think of objects (lemma) that generate a higher level of semantic understanding in the analysis of the content on a page. In that paradigm, the user/redistributor himself can become an object associated with the content, increasing the ability to generate relevance around the content.

    And…that’s obscure.

    Thanks.

    Dan

  • http://www.nstein.com Martin Brousseau

    To follow up on your question Oleg, I think that semantic search is a layer of “searchability” that can leverage both search paradigms – any search paradigm in fact – and this is what Dan means by “higher level of semantic understanding”. It’s not “something” you can put between Google, Twitter or Facebook searching capabilities, but mostly something to add over – or around- them.
    If we add this layer to Twitter for example, a query like “How many people (and who) has a bad opinion about my product since a month” can be answered.
    My question to you Dan is what do you find ‘obscure’ in semantic search capabilities, is this the concept itself or its feasibility, the means/technology that can be used to achieve them?

    -martin

  • http://www.nstein.com Martin Brousseau

    To follow up on your question Oleg, I think that semantic search is a layer of “searchability” that can leverage both search paradigms – any search paradigm in fact – and this is what Dan means by “higher level of semantic understanding”. It’s not “something” you can put between Google, Twitter or Facebook searching capabilities, but mostly something to add over – or around- them.
    If we add this layer to Twitter for example, a query like “How many people (and who) has a bad opinion about my product since a month” can be answered.
    My question to you Dan is what do you find ‘obscure’ in semantic search capabilities, is this the concept itself or its feasibility, the means/technology that can be used to achieve them?

    -martin

  • http://www.nci.com drm

    Martin,

    Great clarification/expansion on the thought.

    As far as the obscure comment goes, I was blithely poking fun at myself. When a concept ends up being too circular and self-referential, I get suspicious of it, because my brain gets tangled.

    I am fascinated by the dynamics around developing and propogating semantic search. It feels a little like Linux 8 years ago: an more powerful and elegant language that did not get as widespread adoption as it might have because of the overwhelming dominance of other languages. Semantic search will require a reworking of the mark-up language creating web pages, right? What will the catalyst be for that kind of development?

  • http://www.nci.com drm

    Martin,

    Great clarification/expansion on the thought.

    As far as the obscure comment goes, I was blithely poking fun at myself. When a concept ends up being too circular and self-referential, I get suspicious of it, because my brain gets tangled.

    I am fascinated by the dynamics around developing and propogating semantic search. It feels a little like Linux 8 years ago: an more powerful and elegant language that did not get as widespread adoption as it might have because of the overwhelming dominance of other languages. Semantic search will require a reworking of the mark-up language creating web pages, right? What will the catalyst be for that kind of development?

  • http://twitter.com/alltoute alltoute

    Sorry for the long comment :-)
    That kind of markup surely requires some help from automatic semantic annotations technologies (named entities extraction, automatic categorization, sentiment analysis, fact extraction -> text analytics in general). But I think the real catalyst to Web level adoption will be a question of SEO. If someday websites owners are faced with a major SEO problem because of bad (or absence of) semantic markup, then they will have “no choice” to join the parade. And when we talk about SEO, we talk about search engines. That’s the main test for linked data: major search engines have to leverage semantic annotations. Yahoo! is already trying interesting things right now (Yahoo! Boss * Search Monkey) and it’s not because Google is not aggressive on that field that they can’t do it. Right now it makes more sense for a company to invest in semantic technologies at the enterprise level or inside a news web site for example because they have an immediate significative ROI. Talking about Google, the thing a lot of people forget is that semantic search can also be done without linked data. Semantic Search and Linked Data are 2 different things. Google is not completely against the Linked Data and Semantic Web idea. They seems to just believe that unstructured information and semantic annotations themself are not sufficient to resolve all the problems. And again, to be able to play on that level of disambiguation, text analytics and related technologies are the key. Context analysis will always be necessary and I think that’s one of the Twitter disadvantage: context is in a lot of cases very poor or contained in a link. They will need to go on the Web and the Web level is a very serious technical challenge. It’s also difficult to imagine how deep semantic annotations can be included in a Twitter post right now.

  • http://twitter.com/alltoute alltoute

    Sorry for the long comment :-)
    That kind of markup surely requires some help from automatic semantic annotations technologies (named entities extraction, automatic categorization, sentiment analysis, fact extraction -> text analytics in general). But I think the real catalyst to Web level adoption will be a question of SEO. If someday websites owners are faced with a major SEO problem because of bad (or absence of) semantic markup, then they will have “no choice” to join the parade. And when we talk about SEO, we talk about search engines. That’s the main test for linked data: major search engines have to leverage semantic annotations. Yahoo! is already trying interesting things right now (Yahoo! Boss * Search Monkey) and it’s not because Google is not aggressive on that field that they can’t do it. Right now it makes more sense for a company to invest in semantic technologies at the enterprise level or inside a news web site for example because they have an immediate significative ROI. Talking about Google, the thing a lot of people forget is that semantic search can also be done without linked data. Semantic Search and Linked Data are 2 different things. Google is not completely against the Linked Data and Semantic Web idea. They seems to just believe that unstructured information and semantic annotations themself are not sufficient to resolve all the problems. And again, to be able to play on that level of disambiguation, text analytics and related technologies are the key. Context analysis will always be necessary and I think that’s one of the Twitter disadvantage: context is in a lot of cases very poor or contained in a link. They will need to go on the Web and the Web level is a very serious technical challenge. It’s also difficult to imagine how deep semantic annotations can be included in a Twitter post right now.

  • http://www.nci.com drm

    Not long at all, and thank you for adding an informed perspective. The comment was very helpful. If I understand your analysis correctly, I wonder why news web sites, for instance, wouldn’t standardize on a semantic search structure? Would that give them an opportunity to create a differentiated search environment that could begin to drive it’s own usage? Are there substantial technology impediments to such an industry-wide standardization?

  • http://www.nci.com drm

    Not long at all, and thank you for adding an informed perspective. The comment was very helpful. If I understand your analysis correctly, I wonder why news web sites, for instance, wouldn’t standardize on a semantic search structure? Would that give them an opportunity to create a differentiated search environment that could begin to drive it’s own usage? Are there substantial technology impediments to such an industry-wide standardization?

  • http://twitter.com/alltoute alltoute

    I think more and more news web sites are going in that directions, we also have to remember that semantic web is still in progression. I like “drive it’s own usage” :-) I think it is one of the most interesting application that could be build on top of a semantic search platform.

  • http://twitter.com/alltoute alltoute

    I think more and more news web sites are going in that directions, we also have to remember that semantic web is still in progression. I like “drive it’s own usage” :-) I think it is one of the most interesting application that could be build on top of a semantic search platform.