The Daily Trap, When Will Information Find Me? The State of the Personalized Web
In time, we will learn whether or not Facebook’s Graph Search will revolutionize the way that we, as individuals, interact with social media and information. But Graph Search’s recent unveiling does provide an opportunity to reflect on the current state of personalization and discovery and an opportunity to consider our aspirations for this technology.
As interesting and as ambitious as Graph Search is, Facebook is still thinking too small. When it comes to true personalization, Graph Search is, in a sense, answering the wrong question by still forcing us to ask questions in the first place.
With the abilities of today’s computers (particularly for Facebook given their scale of operation) and with all of the information about ourselves that we’ve voluntarily put out there for everyone to see, what we should be working towards is enabling relevant information to find me rather than my having to search for it.
I want the computer to take what it knows about me from what I’ve already told it and to apply that knowledge to proactively discover content and information that I find interesting, engaging, and thought-provoking — even going as far as to introduce me to ideas that challenge my established views and that push me to consider new points of view from other’s experiences.
The launch of Graph Search suggests that even the tech giants like Facebook simply are not there yet.
For the Web to move forward effectively, in an age where information is exponentially expanding, we need to shift from a paradigm of search to discovery. Search, of course, will always have a valuable role, but its capabilities are necessarily limited amid so much information and content. We need the tools and functions that can dive deeper into relevancy rather than broader in scope.
A search only starts after we’ve typed in some keywords and told it to “go.” And for all of its utility, search lacks a sense of serendipity or “unexpected relevance,” as my co-founder Hank (and Jeff Jarvis) calls it.
To put it simply, search forces you to click through to discover. True personalization offers the opportunity for discovery to be delivered.
Also, Graph Search is built on social recommendation. Aside from the problems that naturally arise from this — do people actually like the things they “like?” — recommendations are a poor, first approximation of personalization.
Interests grow and evolve over time, and someone’s online persona is rarely a perfect match for their actual person. And as anyone with a Netflix account and children knows, we’re often a proxy for others with different interests. I, for instance, have a diverse group of friends each with their own equally diverse interests, and there’s a very real possibility that I do not actually like any given thing “liked” by one of my friends.
Today, by using recommendation systems, Target may be able to figure out that a teenage girl is pregnant, but Barnes & Noble and Amazon have yet to pitch me a product I actually care about enough to purchase.
So, if Graph Search isn’t the path to this idea of information finding me, what is?
We need to put the machines to work for us. We need technology and automation to bring content to my attention. Artificial intelligence should scan content that I like (and don’t like) and figure out the content’s characteristics so it understands why I like (or don’t like) that content. This is about understanding rather than simply graphing.
And finally, I want to automatically feed that knowledge back into discovery so it can find more interesting content to bring to me.
To get there, we first need to get over the “search and find” mentality. We have web search, social search, local search and enterprise search. Every few months Google, Bing, and now Facebook release an update that “revolutionizes” search, but it does very little to move us towards the bigger promise of the web… personalized discovery.
Innovation in search does not mean we’re making progress towards personalization. These are dramatically different concepts, and one doesn’t necessarily feed into the other.
We also need to move away from the concept of utilities that help us interact with the Web. Instead, the Web needs to mold to each of us as individuals. More importantly, we need to adopt a mentality and belief that the Web can, and should, be personalized down to the individual level.
We’re starting to see this happen in bits and pieces. Certain apps embed some level of customization or personalization, but it needs to be widespread and even ubiquitous to be truly effective.
In order to reach ubiquity, we can’t continue to pretend that social search or simple recommendations come anywhere close to the vision we hold for personalization. Most technologists agree that personalization is a big piece of the future of the Web, so let’s hold ourselves to high standards and get it right — and the first step is to finally move beyond the mentality of search.
-Gary Griffiths