New Signals in Search: The Bing Social Layer

The technology search engines employ to retrieve that needle in the information haystack is staggering and humbling; it’s one of the things that get us fired up to come to work at Bing every day.


Today we announced an exciting expansion to our long-term partnership with Facebook.  Specifically, we are rolling out some new features that allow you to take your friends with you into your Bing experience, both at as well as within the search experience at We will enable a great search experience for people queries, by bringing in information from your Facebook friends and people who share networks with you, and we will show you what your friends have liked (using Facebook’s public like platform) as you navigate through search results in Bing.


Why is this interesting?  At the customer experience level, search is getting more social and more personalized.  We think that’s a great thing to help you make better, more informed decisions.


But we think there is something more profound going on under the covers, an inflection point in the search industry that will enable more interesting social scenarios in the future. 


Traditionally, search engines rely on a large number of clues to help us determine what you are looking for.  We call these clues “signals.” Search was built on a concept of these signals that told engines what was probabilistically the most likely piece of information you wanted based on the words you entered.

This has worked pretty well over the years and helped search improve a lot – early signals like meta tags to give the engines hints on page content and reverse IP to provide more locally relevant results, which has evolved to the mobile phone with the addition of geo-location data.  Eventually the industry developed more complex signals like anchor text and popularity models to try and bring a human element into the mix.  In Bing, we look at more than 1000 signals to try and get you the best result.


Our focus at Bing on helping customers make better decisions allowed us to rethink this model a bit, as we believe in some ways the current set of signals is not perfect for the way people actually make decisions and accomplish complex tasks.


The fact is the real world isn’t defined purely by how information is connected; it’s also defined by the connections between people. 


But the signals that engines have come to rely on to help you find what you’re looking for are not really representative of those human connections and the role they play in making decisions in real life. 

So we asked ourselves: what if we could make your friend’s opinions visible to you in a simple way when you’re searching?  What if we could put that information at your fingertips and bring your trusted sources to the forefront for many of your searches? What if there was a “social layer” in search that could make the whole process of connecting to information and making decisions more social, more personal and more useful?


For many decisions we make every day, both small and large, it’s the advice of others that help us answer a question.  Examples include:

·      Should I see the movie Inception? Would I like it?

·      Should I buy an extended warranty for my new car?

·      What is the most interesting tourist site in NYC?


These types of questions are best answered if you ask someone who knows something about you.  A friend might know you like Leonardo DiCaprio but you don’t like movies with guns.  They might know you should get the warranty because the car you’re buying is prone to problems after the fourth year – because they own a car from the same manufacturer. 


All these questions are best answered by a richer set of signals that take into account not only relationships between data but relationships between people..


We think it’s time for a real, robust, persistent social signal. Facebook has led a transformation of the Internet already.  It has reached and passed 500 million members, and the amount of content created inside Facebook each day is staggering.


Customer data really support these conclusions.  Indeed we see it all the time: 50% of people say that when making a decision, they take into consideration thoughts shared by others in their circle of friends.

What’s most exciting to us at Bing is the evolution of this new signal, based on something more than just pattern recognition or keyword matching. This new signal will allow us to do a better and more comprehensive job predicting what resources and content are most relevant to you because, in addition to all the other signals we use, other people you trust have found them interesting. It means better, more personal search experiences and better tools and input to help you make decisions.


Today’s launch is only a start of this journey but it reflects an important milestone in evolving search into something more than a collection of machines and algorithms.  It shows the promise of what can happen when we take the greatest decision engine – the people you trust – and combine that with the world’s greatest resource for information – the world wide web.


This is just a first step in ensuring that people are a first class entity in Bing.  We will build more exciting experiences on top of this social layer, as the web continues its journey from a collection of documents to a more full-blown digital society, where people matter as much as pages.


For more information on the new features, check out our Learn More Page.


Enjoy, and as always we want to hear what you think about Bing, so keep the feedback coming.

Satya Nadella – Senior Vice President, Online Services Division


View Video
Format: ???