Social media monitoring as a technique has evolved to become almost totally marketing-centric, but this misses a trick.
While dashboards provide great real-time information, they can struggle to help you really learn about your consumers.
Going back to basics for a moment, we can boil social media down into four parts:
- Creation – your consumers creating, sharing, commenting and publishing their own content.
- Listening – how you access this stream of conversation.
- Learning – your ability to create and share insights about your consumers and markets.
- Interacting – acting on this knowledge, influencing the conversation and engaging directly with consumers.
As an industry we’ve come a long way in terms listening and interacting, but there’s still room for improving how and what we can learn from what consumers say. At the moment we seem stuck at measuring what can be measured (keywords, reach, sentiment etc.) at the expense of trying to really learn what consumers think.
Here’s a quick test you can do to see if you’re really learning or simply counting:
Q: Has your approach to social media so far:
- Helped you understand how and why (not just if) consumers feel the way they do about you?
- Let you find answers to questions you didn’t know you had?
- Given you the evidence and insight to innovate, solve problems or be creative?
- Given you valuable insights to share with other departments?
Those with listening systems with outputs limited to just dashboards tend to answer ‘no’ to most of these questions.
However, it’s not a lack of desire that’s holding firms back, but rather a matter of having the right tools. Computers are great at sorting and counting, but still have a way to go to unpick the chaotic beauty of human communication.
The challenge seems to be how to make the most of these systems by combining human analysis with machine processing.
We’ve seen some evidence of firms using this hybrid model to great effect, but so far this approach tends to be light touch and focuses on observing trends in the outputs of dashboards, rather than going back to the raw data.
The problem here is one of scale, how can you practically read thousands of comments and learn from your observations?
There are some recent developments that are starting to provide answers to this. These use some new techniques (such as out-tasking or crowdsourcing) that enable firms to use humans to read large amounts of information and help identify what’s important and interesting.
This then can be harnessed more easily to create genuine insights about how people think.
Our work for an engergy company shows how this can be put into action. As part of a recent innovation project they collected thousands of detailed comments about energy and how it fits into people’s lives and wanted to see what they could learn from it.
Sentiment and keyword analysis couldn’t provide them with the depth of insight needed, so we read and analysed the comments to understand what people were saying about energy, when they talked to each other.
From this they were able to learn more about what motivates people’s interest, what untapped needs they had and identify ‘hooks’ to help communicate more effectively with them.