Integrating social media into business
Chris Nott 100000MPDE Visits (5233)
Social media generates lots of interest for many who are keen to explore the benefits of big data technology, but a recent study published by IBM’s Institute for Business Value shows that CMOs are struggling to make progress in coping with social media. Organisations are unaware of how best to extract the data they want from the noise and they are unsure whether that data can be used to provide any business value. To compound the challenge, organisations are demanding answers from analytics increasingly quickly.
I was fortunate to present on IBM’s implementation of social media analytics for The Championships, Wimbledon, one of the four Grand Slam tennis tournaments, at the Institution of Engineering and Technology’s Data
In order to understand the collective interests and views of people as soon as they are expressed, keywords on certain subject matter, such as names of players, were put into the Twitter fire hose. Natural language processing was used to understand what people are saying about a specific topic to then enable further exploration. For example, by filtering the results of the analytics on a particular topic such as an individual player, clearly defined conversations can be identified with the associated volume in real time such as when they are playing a match.
Furthermore, a re-tweet by someone or an organisation with a large following can have huge reach. Perez Hilton, the US television personality, re-tweeted news of Serena Williams losing her fourth round match on 1st July, contributing to peak reach of over 32 million users. By understanding its followers, an organisation can plan and drive interaction with specific organisations or users in order to drive up “Reach” to get a message heard by a wider audience or a specific community. A re-tweet from a trusted source can be hugely valuable to an organisation to get its message out and acquire new followers in the process.
Trending themes can be identified as they emerge using a glossary of possible terms important to the organisation. Trending sponsors were highlighted during The Championships and this is useful not only for Wimbledon to measure the value of rights granted, but also for sponsors to assess the impact of activation. Coupling this type of trending with filtering, such as by player, allows relationships between those themes and other topics of interest to be identified.
Wimbledon was able to use analytical insight from social media in a virtuous circle of interaction. Social analytics was used to inform and tune the content of the Wimbledon website in real time, thereby making it more relevant, which in turns drives more traffic and further informs site content.
We have seen that volume in social media analytics is important, but the spread of sentiment may provide deeper insight into what people are really thinking, and this may differ from what is otherwise perceived. Sentiment in traffic content can be indicated using analytics. Tweets referencing a specific topic were scored -5 to +5 for sentiment during The Championships, and were able to take account of sarcasm and other language subtleties. The depth could indicate if people really like or dislike something, for example. This can help CMOs understand the buzz about a topic of interest, perhaps an advertising campaign, directly from views expressed by customers, and it could then be used to help predict what is likely or unlikely to work.
In turns out that very few users of Twitter enable geo-tagging of tweets on their smart phones. However, a large proportion of users have location information recorded in their profiles. For some analytics this may be good enough. It was useful during Wimbledon to illustrate the geographic spread of interest in a player at a point in time. However, reliance on such data without regard for its limitations is likely to lead to spurious results. This is an example of the challenge of veracity in big data because you cannot rely on the location from which the tweet was sent being the same as that recorded in the user’s profile.
These analytics can be combined together into a single picture showing location, volume, traffic content and sentiment over time.
In addition, Twitter conversations can be linked to enterprise data to illustrate the relationship between conversations and locations, organisations or other events based on a timetable, for example. Locations could be picked out and highlighted based on the amount of chatter driven by particular circumstances, alerting something of unexpected interest. During The Championships, insight on the volume of social media activity was combined with match schedules and court locations.
Understanding the interests and views of individuals expressed directly on social media is more representative and likely to be more useful than relying on views otherwise formed by analysts alone. The experience from all of these scenarios at Wimbledon shows the potential to integrate social media into the business.