Social media, while a relatively new phenomenon, has become ubiquitous globally and is changing how people interact. People share their thoughts and opinions and engage with their friends, family, acquaintances and even strangers through these platforms. This means social data has become very valuable for businesses to understand, communicate and acquire customers. Businesses are using these platforms as broadcasting and listening tools.
Powerful solutions like these also present problems; how should businesses extract the right insights from user-published social data? While there are different listening tools available in the market, the real challenge is in deriving tangible value, by separating noise from the collected data. Social data is very noisy and hence transforming raw data into high quality workable data is a complex heavy duty task. This, along with profiling the actual users behind the data, is hugely dependent on the application of big data technologies and data science.
As the volume of data grows the technologies in this area are also maturing in tandem. The Amazon, Apache and Google ecosystems offer capabilities that would be considered revolutionary just four or five years ago. It used to be that only a limited amount of structured data – a few gigabytes a day – could be processed, and at an unacceptably high cost. Today tools like Google Big Query and Dataflow mean data volume is not even the primary consideration because they enable the processing of many terabytes of unstructured data every hour.
Newer data science methods also allow us to use social data in order to predict future outcomes and analyze ‘what if’ scenarios. If we look at marketing analytics we can see that businesses have been investing heavily on tools and practices to monetize social data. For example, telecom companies listen to social media and create innovative products to increase customer loyalty, reduce churn, cross sell, up sell, provide customer care and increase their market share. Another example I would like to quote is that of e-commerce companies, where they thrive in a market that is very close to perfect competition and where almost all the players sell identical products, have relatively small market share and the buyers have complete information about the market including feedback from peers about the quality of products. Flipkart, an Indian ecommerce company, allows users to interact with their friends through a chat application and get their opinion before buying the product. For the customer, their decision is vetted by fellow users on their purchasing decision; for Flipkart, this information becomes social data for further analysis. For e-commerce companies, speed is a key factor of success. They gain a competitive edge and this is now becoming the exclusive domain of social data.
We are also witnessing software giants like Google, Hootsuite, Salesforce, Qlikview and Tableau investing heavily in research and development in order to develop social media analytics. Tier 1 telecom operators, multinational FMCG companies, and pharmaceutical companies have adopted their products and have started to listen, engage, measure and analyze content in social media. Additionally, asset management tools are also on demand for managing social assets as these companies have social presence in multiple platforms, countries and for multiple segments.
Latent Semantic Analysis technique in Natural Language Processing (machine learning) is also providing results closer to human analysis. Hence the unstructured information from social media can be clearly differentiated from noise and can be used for critical decision making process. Most analytics are based on volume metrics and content. Volume metrics based analysis helps us to understand the level of engagement, reach, impression, number of reviews, likes, favorites and trends. Content-based analysis helps us to understand market sentiments and is also helpful for segmenting user profiles based on different attributes.
The speed of progress in social media analytics has been impressive. However, one area of deficiency is the integration of social media platforms with other mainstream systems such as ERP, CRM & KYC. The slow rate of integration is partly due to user psychology and partly because of the business purpose for social media platforms. Twitter users like to differentiate their social media identity from their real identity, and hence do not want to consider their social media profile as a substitute for an email ID or phone numbers. Businesses hence are unable to update their CRM systems with social media profiles and monetize the information as they are yet to distinguish whether a statement is from a real or fictional identity. Even if the person’s opinion is real, it is difficult to reach the person as part of the sales process.
The original purpose of social media was not to help businesses use the data created for business analytics. Hence greater efforts are required to explore the commercial potential of this space. However, a paradigm shift is happening with user psychology as well as for social media platform vendors. This will facilitate the integration of social media data with other kinds of data. In turn this will generate superior RoI. A case in point is social care: the delivery of customer care via social media, which is being widely adopted by companies to offer a faster and better response to their customers.
Despite the challenges, social media provides us with a wealth of new insights. The businesses that prosper will be the ones that gained those insights first.
