Hi Elliott,

Yes! The human element is so vital. In my view, it isn’t only because humans are better are deciphering sentiment, but also because social media is an opportunity to truly hear from customers (and, even better, to let them know you are listening and engage in conversation). Listening platforms are great at finding the conversations, facilitating internal collaboration, gathering metrics for identifying the most active conversations, finding influencers, calculating ROI, etc., but they can’t add the “personality” or human engagement that clients are really expecting to get from brands.

True “listening” is a human skill. Just like any other relationship, we listen to let people know we care about what they have to say. In a customer sense, the same applies. In addition to gathering insights, listening sends a message to customers that they are important to us.

Indeed, there is no substitute to hearing things first hand from customers.

Cheers,
Marcel
CEO, Radian6

Marcel,

Thanks for your comment. How do you see Radian6 taking steps to fill this gap?

Elliott,

You hit on an increasingly popular topic with many brands today and it’s good to see that people are starting to look under the hood a bit more about sentiment analysis of social media.

While there are no perfect systems today to deliver 100% accuracy, we have found that a hybrid approach serves our clients well until technologies improve in this area.

By that I mean everything from establishing concrete scoring criteria for every possible topic and issue upfront, to combining human and technology analysis and supporting that with frequent data quality checks every step of the way.

It’s not easy, but it does deliver the best chance for accurate analysis that brands can rely on to make smart decisions.

Mike Spataro
Visible Technologies
SVP, Client Strategy

, technology support and data quality assurance.

Mike,

There’s a wonderful phrase in psychology–”the power of thin slicing”–which says that as human beings we are capable of making sense of situations based on the thinnest slice of experience

Do you think there is a possibility that thin slicing techniques could be used or modified to fit in the online world to help monitor the human element?

elemenager,

Now that’s an interesting concept — how to get technology to improve the human dynamic.

Mike

Hi Elliott,

Thanks for the question. I have a couple thoughts on that depending on what you mean by the “gap” and also your specific business goal(s).

1. If you mean the gap between what a computer can currently do well vs a human (i.e. such as determining sentiment of a comment), and your business goal is to use the social web to build relationships with customers, to listen to them (and to be perceived as a brand that listens) in order to ultimately build community & grow advocacy around your brand… then I don’t think such a gap should be filled. You can’t build community and relationship without the personalities behind the brand and that can’t be mechanized. However, software tools are very useful to help with that in terms of facilitating collaboration amongst the people doing the engaging, tracking all the conversations/maintaining history, etc. To that end we have built workflow & collaboration features and enable you to measure the overall engagement of your team, etc.

2. On the other hand, if your business goal is gathering insights such as determining which topics or issues are most relevant to to your customers, then there are things which can help with the gap. When we talk about sentiment, we are talking about words (were the word positive or negative) and that is only a small part of the picture in gathering insight. Many people use the saying, “actions speak louder than words” and we all know that some people’s words can cause people to follow more than others (either because of who they are, or what they said). If customer A says something extremely negative, but nobody relates to that comment, engages with that person, or acts on that comment, what would you conclude compared to customer B who says something slightly negative (or positive, whatever), but everyone jumps into that conversation, links to it, comments on it, etc.?

To me, one of the most powerful characteristics of the social web is that it doesn’t just allow you to track words, but also actions. The medium itself records the actions. And, back to your question on what steps Radian6 is taking to fill the gap, we are tracking all those actions relative to every conversation. We gather all the “digital breadcrumbs” that, in fact, become a better indicator of the relative importance of an issue to your customers, than the words themselves. If customers are talking about issue A and you can see that this issue is gathering all the commenting activity, linking activity, and even traffic to your website, then you know this issue is resonating with customers and important for you to understand regardless of its sentiment. Now combine the use of these social metrics with things like sentiment and you have a powerful tool to assist the humans in gathering insight.

Imagine a human trying to go through six thousand posts in one day to figure out what the hot topics are? Then, imagine if you had a 100% accurate sentiment scoring of those – so now you have two thousand negative posts and three thousand positive posts… does that help? But imagine if you could ask, “show me the top 10 posts, from all six thousand, which have the highest customer engagement and/or commenting activity right now”? Then your human analyst can easily pull these needles out of the haystack.

I often think of the analogy of choosing a restaurant. I ask and someone says, “restaurant A is good”. But then when I walk down the street, I can see that there is a huge line at restaurant B and almost no one at restaurant A. Which tells me more, the words or the actions? That’s why we invest so heavily on gathering all the social metrics about the conversations.

I hope this helps a bit. Again, thanks for your question.
Cheers,
Marcel
CEO, Radian6