Relevancy and The Future of User Experience on Facebook

Gs-450h Last summer, I had a friend over for dinner. When he got to my house, he pulled into the driveway in a brand new Lexus GS 450 Hybrid. I congratulated him on his new purchase and told him that I was looking to lease the same car next year when my current lease was due. He had no idea. My response was, "yeah, I posted something on Facebook like a week ago about it."

Evidently, he didn't see my post.

I went on to ask him when he got the car and how much he paid for it. His story was fascinating. The way he pit dealers against each other to get the best deal possible was revelatory. He didn't step foot in one dealerships; He did it all over the phone, and ended up getting the car plus all the fees involved for below market value.

Dumbfounded, I confessed I had no idea he was such an aggressive negotiator and that I wanted him to negotiate my next car lease. His response was, "I'd love to." But then he said, "I'm surprised you didn't know about this since you're on Facebook all the time. It's all over my Facebook page."

I was surprised as well since none of these posts made it to my wall. I only saw links about real estate from him (he's a realtor,) but none about cars.

The point of my story here is this: Facebook does not recognize or elevate relevant content between friends.

It would have been awesome if Facebook somehow realized that my status update in which I mentioned I was looking to lease the GS 450h was related to my friend's posts about his purchase experience and somehow got us to connect.

But Facebook doesn't work that way (yet.)

Right now, relevant content is discovered on Facebook by chance. My ability to find relevant content to me is depending on 1) the frequency in which my friends share content and 2) the time at which the content is shared and 3) the time at which I check my wall. All three conditions have to align for me to see that one piece of content that's relevant to me.

It sucks, doesn't it? If I'm willing to sift through countless irrelevant posts from my friends, the least of my expectations is that the relevant posts are brought to the top of my wall where I can easily see them.

But how do you define "relevant?"

A Real Opportunity

Since I do work for @Edmunds, I automatically switched to "find a solution" mode. When I got into the office the next day, I talked to my colleague @HowardOgawa about my experience. After bouncing ideas off of each other, we decided to take this on as a challenge.

Our objective: Elevate Relevant Automotive Content and Conversations to Friends That Care About Them within the Facebook Ecosystem.

We decided to build a Facebook app to do this. Our app at a high level would passively listen to the stream of activities (i.e. status updates, links, checkins, photos, ..etc) coming from the app subscribers and try to mine the data for automotive relevancy. As relevant data is found, other subscribers are notified. Subscribers would also get to indicate some of their friends as "auto experts," which in turn will render automotive content coming from those individuals even more relevant to that user (granted they subscribe to our app, of course.)

Soon after we started looking at the data in a user's Open Graph, we realized that we couldn't mine that data efficiently. Something was missing from the structured data. As we dug deeper, we were convinced that for the data to be meaningful for us, it had to be segmented or categorized.

Content Segmentation and Relevancy

It's hard to determine whether a link a user shares on his/her wall is a link to an article, a YouTube video, a Flickr image or an audio file. The type property of the link object in the Open Graph always returns "link." Sure we have access to the optional message the user attaches to the link and the description that is captured with the link, but that isn't enough to determine the type of that link, and most importantly, the category into which the content of the link falls.

Howard and I went back to the drawing board. It was pretty clear to us at this point that in order to truly recognize relevant content on Facebook, the Facebook structured data had to include segmentation or category.

A shared YouTube link about the President giving a speech in Egypt should be categorically distinguished from a shared YouTube link about Arcade Fire rocking out at The Hollywood Bowl. The former falls under "politics" and the latter under "music."

When that segmentation is embedded into the Open Graph, relevancy becomes much easier to discern and users can specify what content they care about from what friends. I'm sure many of my friends on Facebook would rather see less of my political posts and more of my entertainment ones. With segmentation, they will have that choice.

Facebook Committed to Relevancy

About three weeks ago, a Facebook spokesperson was quoted in a New York Times article saying, "We’re always looking for better ways to help people discover the most relevant content on Facebook…"

This was great news to me! As a Facebook user, this would help me a whole lot. But according to the article, the approach that Facebook is taking won't help me in my particular use case. The same questions remain unanswered: How will I be able to see relevant posts from my friends? How can I specify what specific topics I trust which specific friends with? How can I ensure that my wall is 80% relevant to my real life needs?

Potential Solution: schema.org

Schema Interestingly enough, around the time the Facebook story broke, TechCrunch reported that Google, Yahoo and Bing were collaborating on a structured data initiative, or schema.org. The goal of this initiative is to help websites optimize their HTML and crawlable data structures to make their content more accurately searchable.

The question here is, why isn't Facebook working with these companies on this initiative?! Facebook already has the social sentiment component that all three of these companies lack. All it needs now is to ensure that the content people share to their wall is meaningful and structured, which in turn will help with the relevancy goal and will help me find the content I really care about.

Imagine if Amazon uses the right semantic tags to describe items on their pages. When users share an Amazon link on Facebook, it's no longer just a "link." It's now a "link to a book called ABC by author XYZ and it's currently listed for $xx." This granularity adds meaning to the "link." Meaning that is translated to metadata that algorithms can computer, manipulate and correlate, all of which can easily produce true relevancy.

The Real Business Potential

Creating a relevant experience on Facebook is great and I'm pretty sure Facebook will get there one day. But there's a potential here to create an experience that far surpasses that. An experience that's not only great, but awesome.

Facebook knows how to do social very well and its objective is to keep users on its platform for as long as possible. But to do that, showing relevant content isn't enough. They need to think about creating a user experience that is detailed, localized and actionable. But in order to do that, they would need to partner with subject matter experts in each content segment (e.g. travel, retail, automotive, finance, …etc) to provide the missing data points that will enhance the relevant experience Facebook is building and make it detailed, localized and actionable.

What do I mean by detailed, localized and actionable? Here's an example:

Mary just read a review of "Under the Tuscan Sun" by Frances Mayes on oprah.com. She decides to share that review with her girlfriends on Facebook and she does.

What Mary doesn't know is that oprah.com's content is semantically structured which allows Facebook to understand what the content of this link is all about. Also, Mary doesn't know that Facebook uses Amazon.com's APIs to enhance the experience for Mary's friends by offering them more detail about the book (i.e. price) and locality (i.e. availability at Borders down the street) and a call to action (i.e. Amazon buy button.) All of which is customized to each friend as they see Mary's post on their wall.

The next day, Kirstin, one of Mary's Book Club friends and Facebook friend, logs on to her Facebook. Kirstin has previously indicated in her Facebook preferences that Mary was a good source for literary/readying content. As a result, Mary's book link is now at the top of Kirstin's wall since it's a piece of content that is likely relevant to her. Kirstin is so compelled by the review she goes ahead a buys the book, by clicking on the Amazon link attached to the post and without ever leaving Facebook!

This could be applied to any segment. Facebook can partner with @Kayak to allow users to find travel deals to Heathrow when reading a link about London. Facebook can partner with @Edmunds to allow users to see the price of a vehicle and contact dealers nearby when watching a YouTube video about Toyota Prius.

The possibilities are endless.

What I'm talking about here could be huge. Google, Yahoo and Bing can get the structured data, but they don't have the social sentiment. Facebook has that, but what they need to do now is ensure the content shared on the platform is structured. Once that's accomplished, partner up with subject matter experts in every segment and use their APIs to enhance the content.

The resulting experience is not only social, personal and timely; it's relevant and actionable.

When a simple Facebook search returns all the relevant content that friends (and all of Facebook users when privacy allows) are sharing in realtime with specific calls to action that meaningfully transition the online experience to an offline transaction, why bother go somewhere else?

Does this make sense? Am I missing something? I'd love to hear what you think. You can leave a comment here or find me on Twitter at @ielshareef

2 thoughts on “Relevancy and The Future of User Experience on Facebook

  1. Excellent writeup. I’d like to hear more about the Facebook app implementation you wrote up to test your hypothesis. The structured data is certainly one part of it, but determining what is relevant to a particular user is a bit more difficult. I’ve considered doing g something similar where tweets are fed into the Apache Mahout machine learning engine to solve the problem you spoke of about o ly stumbling upon links by chance. Having some kind of engine to process the aggregate feed and recommend top articles in relation to what I post, what my friends post, and the type of link would be more effective than hoping I didn’t miss something like you spoke of.
    Pardon typos, I responded from a mobile device.

  2. Thanks, Matt.
    The app we created simply asked for permission to access the user’s stream and status (read_stream, user_status respectively.) https://developers.facebook.com/docs/authentication/permissions/
    The problem with that was that we had to basically interpret each content item in the stream to figure out if it was related to automotive. That wasn’t easy since the link object that’s exposed through the Open Graph doesn’t give you much.
    That’s why Howard and I started thinking about the best way to do this and we reached the conclusion that only a restructuring of the data will resolve this issue. Also, it seems like a the next logical thing to happen considering the push for structured data by Google.
    Facebook has a golden opportunity to truly be the ultimate destination on the web. If I can transact on Facebook, why would I leave it?!
    This won’t hurt the other companies that partner with Facebook. On the contrary, it will boost their conversions and help the business grow.
    The idea of bringing traffic to your site is dying fast. The future is to be part of an ecosystem that works, and today that ecosystem is Facebook.
    Power up your APIs and join forces with them. At the end of the day, it’s the user experience that’ll matter.
    Are you into machine learning these days? Have you spoken to Paddy about it? He’s been looking into it too.
    Good to hear from you, man 🙂
    Cheers,
    Ismail

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s