Social Listening, Social Media Management, Social Customer Relationship Management, and social media in general are all relatively new disciplines. Much of what is being created is new. Because of this sometimes there are no words, phrases, or expressions to describe things that exist in this new medium. It is even harder to describe new methodologies of exploiting these new technologies. We are building all the time and despite our singular focus on the wine industry, we too are innovating and inventing new ways of doing business that (may or may not) exist in other companies. We’ve been on a release blitz for months now and my posts have veered away from comments about wine and tech to features (apologies but most of my theory stuff has been dedicated to the big white paper I am writing). This post is about both and meant to explain two key features we have added to our software and explain the terminology that we’ve had to invent to accomodate this new challenge of describing new ways of dealing with social media.
The first phase we have adopted at VinTank, and was incredibly well explained in a recent Forbes blog, is called “texture.” When we tweet, post, comment, blog there is more than just the words we say (which is what traditional social listening monitors). There is everything around and within that content we express to the world. Like an onion we start with the center which is the words we are “listening” for (in our case every wine word we can find). Then the texture is extracted. The first layer of the texture are the relevant terms and phrases around the signal words. Things like sentiment and key descriptors (everything from variety, region, to tasting note attributes like cassis, blackberry, tannic, etc). These help you derive better context. Then we move to the other layers of the texture which include location, pictures, and embedded links (where we link-dive; more on that in a minute). Finally we try to extrapolate even more context from the user’s bio, past wine content, alternative social identities, and the posts from those identities. By pulling all of this together we can start to get a more complete context of the person and what they truly are saying. For us, customer and content context is king to derive buying style and probability. Context is the future of commerce.
That brings us to “link-diving.” I am sure there is a technical term somewhere that someone from MIT has invented as well but we were unable to find it so we made our own. When you post content that includes a link there is key information underneath that link to help us better understand the intent of your post. This is an element of texture but provides lots of interesting by-products for us and the wine industry. We call it link-diving because when this new feature finds a link it “dives” through the link, following the complete redirect path through to the targeted content. We then analyze the page and extract vital content such as meta information, Open Graph tags, oEmbed references, and even distill a summary of the page (stripping out ads, navigational elements, and so on). Without “diving” into the link, all of this rich contextual information is hidden from us and our matching engine.
First and foremost (we are still testing), almost EVERY wine article that is presented digitally will now be indexed and presented to our customers. The only three types of content that will be missed will be digital articles whose URL’s are not shared on social media, content behind “pay walls,” and non-digital content (rarer and rarer these days). This means almost every VinTank customer can cancel their press clipping services to save money and get faster results.
Secondly, we extract and normalize ALL URL’s that we “dive” into (no matter the source: bit.ly, t.co, or more). This means that we can determine which articles are getting the most social sharing signal regardless of the link shortening service being used. This works not only for external content about brands but also to determine the “virality” of content (blogs, pictures, videos, discounts, marketing campaigns, press, etc). Soon we will be adding the ability to see who is spreading your message and how effective individual “sneezers” are at spreading content.
Third, as a result of this feature our matching has increased by 60%. We now can see into these links which would be empty content for us and bring even more results back to wineries. In fact we’ve “dived” almost 3M unique links since link-diving went live one week ago. Yowza!
Finally the content we present to you now is even richer because we pull images and previews of the content behind the links that we have “dived.”
What’s coming next for both of these features? For texture we will be able to start profiling social wine customers with an algorithm and score that tells that person’s wine understanding and buying style by the statements they make about wine. For link-diving we will be able to start identifying who spreads wine messages best and which articles across the entire social ecosphere are getting the most attention. This will allow us to answer questions like “Is a blog post from Joe Roberts better appreciated on social media than an article from James Suckling?” and much more. We will now be able to determine two new factors of influence: who’s content is more shareable and who shares content the best.
Every day I am more excited about what the future will bring with social media and more than happy to help invent the vocabulary that helps us describe this new frontier.