Facebooks Like Button

Facebook's Like button ages Google overnight

Godfrey Parkin

26 April 2010

Facebook's decision to change the “Fan” button to a “Like” button, and to roll out the capability of placing these buttons on ordinary non-Facebook web pages, has instantly transformed the web. In the process, it has left Google's search looking very vulnerable.

For several years it has been assumed that the next leap forward (web 3.0) will be “the semantic web”, an internet in which web pages contain the intelligence that allows search engines to understand the qualitative aspects of those pages. A lot of extremely intellectual people (the kind that make rocket science look like finger painting) have been working on models for coding pages to facilitate sophisticated semantic searches.

To date, most have failed simply because the emergent semantic structures are way too complex for ordinary website owners to embrace and exploit.

Google, today, is essentially a robot that is incapable of making most types of value judgments. It can crunch through the quality of code, the usability of text and layouts, the essential themes of a page. But it can't tell if a recipe produces a good cake, if a joke is funny, if a news story is true, or if an image is awesome. And it can't tell whether a page is inherently pleasing.

It tries to approximate these value judgments by looking at the number and relevance of links to a page coming from other web pages - a link is considered to be a human endorsement. The more “popular” a page is, according to its link structure, the more likely it is that Google will recommend the page in its search results.

Clever as this is, it is still very much the Model-T of search systems. It rewards pages that are pure in code, clinical in usability, and correct in grammar and spelling. And it assumes that the few who go to the trouble of formatting a hyperlink to a page in their own web pages, are representative of the masses who love a page but just don't have the urge to link to it.

The result is that in response to search queries Google typically produces safe, elitist (rather than populist) page recommendations. It promotes BMWs and relegates VWs. Facebook, through its Social Graph approach, now has the power to change this, by making the complexity of semantic engineering irrelevant.

Instead of website owners passively hoping visitors will give a page the thumbs up with a hyperlink from their own site, they can now place Facebook “Like” buttons anywhere on the site. The “Like” button instantly creates a link, without any hassle, from a liker's Facebook profile, and notifies that person's friends to go and check out the page. Facebook is nearing 500 million active users. The number of links spanning the web is set to grow exponentially.

If Facebook is smart, these will be “no-follow” links which Google's spiders cannot index, effectively freezing Google out of the knowledge base. The potential now exists for Facebook to launch a search engine that reflects the true popularity of web pages. Coupled with everything that the service knows about each member's interests and demographics, search results may be filtered by multiple dimensions (e.g. show me the pages about “spring fashion trends” that are most popular with South African 16-24 year old females who also like Desperate Housewives).

Hook into that power a highly-targetable Facebook version of pay-per-click advertising that website owners can place as easily as Like buttons, and Google may well become the MySpace of search, once unbeatable and now humbled.

Those corporate marketers who still regard Facebook with disdain had better wake up and start Liking it.