Social Media and Search have had a intertwined relationship in their emerging development. Now with real-time search of social media streams, the two are more related and supportive than ever. Today’s panel of experts takes a look at what social search is and what it isn’t, who’s working on it and getting it right, and perspectives on making search and discovery more relevant to users. Panelists include: Max Ventilla of Aardvark, Marc Vermut of Fine Point Solutions, Brynn Evans, Scott Prindle of Crispin Porter + Bogusky and Ash Rust of OneRiot. I have coffee and they are playing some Inertia rock this morning while everyone set’s up. Let’s get started.
What is Social Search and how does it benefit the social graph?
Brynn Evans is first up. She is a digital anthropologist looking at what people are doing onilne and how we design for social interactions to be integrated with Search. Search is no longer just a question inside a box. But, how can our friends help us. Search is a process over time. She defines three flavors of social search. Collective, Friend -filtered, and collaborative. People exhibit two main social strategies. Some want to ask the network for help and guidance before doing a search on their own. Others distinctly want to embark alone first and see what they can gather. It’s not until they can’t find what they are looking for when they ask friends or networks.
Next is Max Ventilla as a Google zookeeper after his startup Aardvaark was acquired by Google. Subjective questions generate majority of revenue. But and unreliable and there is a social cost of asking a favor. How do you find the best person to answer your question? Each individual’ network is growing exponentially. In queries of subjective nature, average query is 19 words as opposed to Google 3-4 words. Social context is different than the social graph, the context is frequently sufficient. Went to same school, etc. Intimacy (more than authority) facilitates trust.
Ash Rust is Director of Search Relevance at OneRiot. Advertising is matched with realtime web and focused on indexing what people are talking about in social media. Delivers very realtime both in browser and app modes for a swarming type of distributed tool. Ask the masses for best answer.
Prindle is thinking about social search in the digital marketplace. “Give customers something good to talk about in social media, and they will talk. That conversation becomes content for social search, helping to drive additional traffic and conversation.” This sharing used to happen in email, but is not happening in public. Advertising space for many years has been start in TV ad, then drive other marketing efforts. New model example: Best Buy’s @Twelpforce as example of social search platform. Using TV to extend to wider audience.
Where is Social Search going?
Will these new forms of search overtake the Google Box? 20-40% of queries can be handled by social search currently according to Rust. It depends if they will overtake. Different responses are useful in different search. Brynn comments that we often go to Google for an answer, but still need the next level of social relevance. I have an answer, what do I do with it? Then network can follow through. Rust chimes back in with how to rank in social algorithms based on history of streams and their diversity of content. Inference of smaller groups or networks and may highlight authority.
Brynn biggest problems are who is an authority, how to index them, and how to index friends. OneRiot and Aardvaark answer these in different ways. Aardfavrk is a response to the tremendous social activity happening online. OneRiot recognizes the need for different search. For example, a simple search for “killer whale” might bring up lots of information about whales, Orca, etc. But not necessarily what the ongoing conversation is about social responsibility of killer whales in SeaWorld shows after news item of killer whale attack.
Monetization of Social Search and Advertising
Reverse delivering search information starts with answers, then builds brand. Nash talks about the monetization of Social Search. Traditionally, you draw people coming to you with a clear commercial intent. Then they ask you want they want and you can answer with incentive. New model focuses on not just PPC. First deliver organic answers, then deliver sponsored ads not only at cost of PPC, but discounts on bid based on delivering good responses. Bad responses will not be rewarded. Asking people first if they want a sponsored ad typically is over 50% for opt-in. Social search will allow advertisers to target user intent and not just users.
New model is not about buying keywords, but creating content the social search indexes. Headlines are being tracked more than keywords. Links are still more important than keywords. Anchorlinks and number of people linking to your content is more important than the actual keywords and will continue to drive ranking and indexing.
How do you Measure Social Search?
Finally, in measuring social search, @ventilla on social analytics: “Who you are connected to is more important than how you describe yourself.” True analytics of social search is still TBD. Will be a growing area of research.