Social data has proved revolutionary in transforming how companies can understand consumers, by providing a more dynamic and revealing data set. To prove this point, there are numerous data science start-ups engaged in what is known as social listening – analysing the active content being created and shared on social media. However, whilst that may sound revolutionary, there is far greater value in the concept of social clustering.
Both methods of analysing social data can have hugely positive benefits…
- The ability to extract insight from social conversations and interests
- Understanding and increasing content relevance
- A method of tracking brand perception
- Providing greater insight into audience understanding of competitors and a market
- Generating ideas for marketing campaigns
- Improving the customer experience
- Helping with strategic product decisions
- driving an effective marketing and sales strategy
….. and much more.
However, while both can develop a far greater understanding of customers, one can do so to a far more detailed degree, across a much larger proportion of people.
Let’s start with social listening. What is it? The concept is widely accepted to be defined as: The real-time tracking of conversations on social media around keywords, phrases, brands and industries, which can help leverage insight.
Obviously, this kind of analysis can provide valuable insight into understanding what people are talking about, the trajectory of trends, how consumers view competitors, the wider market in general and the influence that this all can have on both the communications and marketing strategy as well as the core proposition of a brand.
With such dynamic and revealing insight as can be gained from using social data, a brand can understand and act on real-time consumer analysis, with a greater appreciation of how a small percentage of the online world might perceive their brand.
The limitations of listening
It is important, however, to underline the phrase ‘small percentage’. Social listening relies on social media users actively participating in conversations, either by posting themselves, or by sharing other peoples’ posts. The majority of social media users are, however, generally less vocal, in the sense that they do not create or amplify content or conversations. Instead, this 86% of users are ‘Watchers’ who use social media only to view content, as opposed to actively participating.
The fundamental drawback with social listening then, is that, whilst you might be gaining real-time dynamic information, this is only representative of a small percentage of people who are using social media. This means that social listening tends to only focus on people who have a very skewed opinion: this vocal few are talking about particular subjects in a positive or negative light, or simply because it is a topic being featured in the media.
On top of this, there is inaccurate sentiment analysis, where the platform misreads content that is meant to be sarcastic, for example; there is the inability to monitor and analyse images such as memes; and the conversations are static – they have already taken place, meaning that there is little to no predictive element.
On the other hand, social clustering can do all the above and more. Essentially, rather than analysing only the active conversations, it instead clusters users into different passion segments, based on the celebrities, brands or activities that they follow. This means that even those social media users who don’t actively create or amplify are considered, simply by analysing what they follow and are therefore interested in.
The result is far greater detail in the resulting insight, widening the scope beyond only creators and amplifiers, covering a much larger proportion of consumers and providing a much more granular understanding of who a customer is and what they are passionate about.
Moreover, the difference in data set – the tracking of passions and likes as opposed to simply shared content – allows for a far more comprehensive understanding in what motivates a consumer to buy. The insight is not based off a vocal opinion but rather off the subjects that people have actively shown an interest in by following or liking it.
Add to this the fact that this method allows you to time stamp this ‘passion’ data, meaning you can see how a customer’s passions will change over time, and you have a highly accurate and dynamic method of not only understanding what motivates certain consumers, but also of predicting how those motivations and passions will affect their purchase decisions.
Now that is data science at its best.