Home » , , , , , ,

What is Facebook upto?

Share on Facebook0Tweet about this on TwitterShare on LinkedIn0Email this to someone

Sandeep Amar, Head (Marketing, Audience and Pre-Sales) at Indiatimes writes about Facebook’s open graph, meta tags and the potential for search. He blogs at sandeepamar.blogspot.com. The views expressed below are his personal views.

“I would rather be hated for who I am, than to be loved for who I am not” is a famous quote of Kurt Cobain, one of my favorite rock musicians. We all know what Facebook is loved for, and now they want to be loved for something else: web search. I am not the first one writing on it. Since the start of open graph and introduction of the “like” button, plenty of reports and blogs have hinted at this move.

Today, Google is the undisputed king of search, and Facebook leads the pack for social networking, with access to over 500 million unique visitors a month. Now with open graph, search functionality in the Facebook interface, and metadata additions in web pages, it looks like Facebook is seriously looking at search. Although the current web results on Facebook are powered by Microsoft’s bing, the interest of Facebook in search cannot be taken lightly.

Today Google runs on the Page Rank algorithm, with bots picking up data, and results driven by the complex ever changing algorithms, I had recently written a piece on the same, how it is getting affected by SEO, the increase of data, and local algorithm integration.


Facebook’s plan appears to be to introduce the open graph, push it to portals under the name of social on site search, and with metadata, and offer the ability to show a part of page as rich information which can be picked and showcased on Facebook. Now generally metadata is put up for search engines like Google, to be placed well in search engine results, and this new metadata makes way for the sementic web, as an additional information. The qualitative part is also picked up from the website in form of “like-ability” of the object, from the “like” button.

If you see any movie listing page(e.g. Inception) on IMDB, you will see the following meta data for open graph in html source:

<meta property=”og:type” content=”movie” />
<meta property=”fb:app_id” content=”115109575169727″ />
<meta property=”og:title” content=”Inception (2010)” />
<meta property=”og:site_name” content=”IMDb” />

The age old meta data for Google and other search engines is as follows:

<meta content=”Inception (2010)”>
<meta content=”Directed by Christopher Nolan.  With Leonardo DiCaprio, Joseph Gordon-Levitt, Ellen Page. In a world where technology exists to enter the human mind through dream invasion, a highly skilled thief is given a final chance at redemption which involves executing his toughest job till date, Inception. Visit IMDb for Photos, Showtimes, Cast, Crew, Reviews, Plot Summary, Comments, Discussions, Taglines, Trailers, Posters, Fan Sites”>
<meta content=”Reviews, Showtimes, DVDs, Photos, Message Boards, User Ratings, Synopsis, Trailers, Credits”>

Now as explained earlier, normal meta data would work with bots indexing the data and algorithm pulling it up for search results. In case of Facebook, the “Inception” page has 85K “likes”, and Facebook has the frequency data for that. Therefore for this object, Facebook search results will have both the object and these pages as the front runners, when a search for inception is made.

The interesting point to understand is that Google first puts the pages in search results, and assigns the future importance to a web page on the clicks on the search result link and other performance factors (such as landing page relevance). Therefore, the search result delivery has no contribution from the users who have seen the page: it initially comes from the metadata and SEO which site owner has done, and user contribution is attributed only once users start clicking data on the link in the search results. In comparison, Facebook could use the algorithm based on popularity of the page using the “Like” button, on the basis of what is liked by more number of users.

Although not hinting anything on search, Mark Zukerberg has indicated, that this data will be picked from portals like Yelp(for local listings), IMDB(for movies), Pandora(for music) and Tripadvisor(for travel destinations). Now all the web pages on these sites have “like” button, and sites like IMDB, Tripadvisor and Yelp have open graph metadata.

This will only allow Facebook to have strength in limited fields, and there is no talk of web search as of now, just the talk of social search on Facebook. One does not really know how Facebook is crunching all this information in the back end, but with such high percolation in the Internet universe, and growing addition of open graph metadata, this information is good enough to have decent search results for many important searches, like on movies, places to visit, news items and so on.

We will wait and see if it is just speculation. Interesting times, none the less.

If you have insight, an opinion or business practice details to share with our readers, please do send across your contribution to nikhil AT medianama DOT com. Do take a look at our guidelines for guest columns

Share on Facebook0Tweet about this on TwitterShare on LinkedIn0Email this to someone
  • Ken

    Hi Sandeep,

    Firstly, well written. Currently, Fb has like button, based on the activity the search engine places the article on top of the results, agreed. Don't you think by having the dislike button as well, the activity could go up and strengthen the results further?

    • Sandeep

      Hi Ken, thanks.

      See the key for a search engine is to relatively rank web pages, the algo could work well alone, with only the likes, and realtive ranking of that. A "dislike" may also work well, but not essential. In any case, if FB goes for search, this is pre-search result data, post search results performance of webpages, has to add in to this algo, which currently Google and others are following, therefore it is this pre-intuiutive information and post click information, added together, meaning both algos working simultaneously to give importance to web pages, and best service to users…Regards, Sandeep

  • Haha

    Semen-tic web FTW!

  • DDD

    Very interesting Sandeep, new information.

    • RTI

      Sandeep, stop posting on your own articles. First rule of posing: Noone compliments with under an alias, only disses with one.

  • gupta

    This can change a lot of things, if it goes this way..

  • Geetu Ahuja

    sounds very interesting sandeep…this is an evolved way of social bookmarking.. just imagine how exciting times can get where i can just sit in front of my computer and look at what people like me like .

    another layer for marketers to manipulate information..

  • ram

    Let me put on my high-school-English-teacher hat to grade this 'essay':
    Article says nothing new, infact there isn't any insight to be gleaned from it. Author makes empty assertions which convey no meaning or are superfluous
    "Facebook’s plan appears to be to introduce the open graph" –it doesnt just 'appear' its right there for all to see
    Today, Google is the undisputed king of search — and what is you're point exactly?
    "This will only allow Facebook to have strength in limited fields" – how about the 'field' of earning money which is what matters in the end
    Citing his own articles smacks of shameless self-promotion. I wish tech writers stopped writing articles which is just a collection of briefs of what they have read somewhere (and on a serious note-without citations). Maybe they seek visibility among their peers, a tendency which I understand given the tech scene in India (write a few blog posts, and some might consider you the authority in that area).
    Sandeep if you had to improve, make judgments, make mistakes but take a stand (not bury you're thoughts in mindless drivel "I am not the first one writing on it")- , be original and stop saying nothing using 1000 words. In short stop wasting everyone's and you're time.

  • Ujjwal Greeb

    Its really informative post