What is the intent of your “Like” ?
With Graph Search Facebook has entered the Search game, and it intends to make search relevant with discovery and recommendation.
That’s neat.
However, one of the prime indexes by which they can sift through data and match queries to responses will be the “Like” button.
To give an example, If I use Graph Search with the query “Which DSLR camera do my friends recommend” - Facebook will search through my network ( maybe later extended network), gather all the instances where my network “Liked” an activity or instance related to a DSLR camera and aggregate them to give me a list of recommendations.
That’s cool.
Now here’s the caveat. Given that many brands have spent tons of money in audience acquisition campaigns- (remember ” Like this page” and you can win a free bar of soap? ) - it is very likely that many of these like’s are not genuine recommendations, but an act that was undertaken simply with the wish of winning a freebie.
So let’s assume that camera manufacturer A ran a Facebook audience generation campaign using an app, where the sign in was to Like the fan page, then I will have a list of those from my network who “recommended” product A.
Then let’s assume some months later camera manufacturer B ran a similar Facebook campaign for audience generation and several members of my network signed in ( by liking the page). I will also get their names recommending product B.
Now let’s say I find that a lot of similar names on these two lists.
Which one would be the real recommendation? Which one would I really go by ? Will it even have relevance to my query ?
Problem is marketers, agencies all colluded together to ensure that the number of ” Likes” was a metric to chase for.This was not Facebook’s doing. But it sure affects them. As the number of “Likes” grew so did the giddy excitement (amidst marketers/ agencies) of having a number to show as proof of efficacy for a campaign. More “Likes” became equivalent to more success.
Sadly it was the wrong metric of efficacy to start with. Still is. Especially if you are paying to buy those “likes”. Through ads, apps or contests. Yes you do get people to engage with the brand, but it should not be taken as metric to define buying intent or relevance. One has to take into consideration the intent of a person who presses the “Like” button. Why did s/he hit like? For public good, or for selfish sweepstake? Or to make someone happy ? Or to buy?
When a system has to sift through this kind of data which is ambiguous, the search results can be pretty skewed in terms of relevance.
It fits in other places as well. Just because I click on an ad word on Google and land on a page does not give a clear indication of my intent.
People will buy when they want/need to buy. At that moment they will search, they will read, they will read reviews, they will ask friends and then based on their need and budget, they will make an informed decision. The best we can do, is to ensure we are there in their mind and that our product has the relevance to match to their needs.
Also, when people really need to search something, they will search. Even if the right answer is in the 20th page of the search results- they will find and use that information which matters to them. Not some paid promotion which supposedly recommends them on their next action.
Graph Search will be important for power users and with time it will metamorphose, add functionalities and then only we can truly understand what a game changer it is.
The consumer is not dumb.
Marketers, brands, organizations, digital ninja’s, social guru’s and agencies need to start by acknowledging that first.



