Home » Vertical » BFSI » The Right Way to Adoption Of Social Analytics Today

The Right Way to Adoption Of Social Analytics Today



CS audit


Social Analytics is a sure way to understand perception of customers voicing in the digital space. Since year 2007 lots of tools have hit the market that promise you not only the captured data of customer comments on a particular brand or product but also some degree of Reporting including Dashboards. A popular report generated is the Sentiment Analysis.

Quickly in about 6 years almost 600 such software tools had hit the market only to realize that differentiation was becoming difficult and not every customer was happy with the restrictive subscription model. Radian6 was one of the popular tools that was subsequently acquired by SalesForce. It was a highly expensive tool even for large customers. Since then every Technology Major has acquired one of these Social Listening Tools to ensure their place in this fast growing market that will soon replace the brick and mortar Voice of Customer Management.

Fortunately or otherwise, there is no regulatory compulsion on Organizations to capture social data. And this practice is left to the ones that want to differentiate themselves from others.

4 Steps in Social Analytics

Data Collection: A quick differentiation among the social media tools was the accuracy of the sentiments predicted and another was the extent of data collected by the tool, typically using crawlers and scrapers. This has necessitated companies to use multiple tools in order to arrive at a fairly comprehensive collection of data with comparable accuracy. Now this is only the data collection part.

Data Sanitisation: Once these data are collected, the data are needed to be normalized or sanitized, sometimes categorized to help prepare reports. This can also be termed as a QC stage.

Data Analysis and Reporting: Once cleaned and categorized, the data is projected in the form of Reports and charts. Today most social tools are dishing out their own version of dashboards and reports in a one size-fits all method, with very little customization. Companies do not find these relevant always to their situation and end up using the raw data in csv format into their own environment to create the reports they need. This space is highly fragmented and left to the company’s general ability on Business Intelligence, usually handled by their BI vendor.

Predictability: Beyond generation of reports, companies who build predictive models using statistical tools to forecast outcome, now have the Big Data as well to be put to use, to make the prediction more accurate. These models are used to modify some of the conditions of business to ‘manage’ a desired outcome. The investment in this space is the least, though this might be the the most effective use of Social Analytics

Key Issues


  • Data Collection is never comprehensive and often contains gibberish information. Local country information may be absent in tools sold by the vendor from a different country.gamification
  • Data Sanitization is best done manually. Most tools today do not apply this as they are keen to provide real time data.
  • Data Reporting by most tools fail in accuracy. Data Analysis often uses text mining principles and the ability to enrich the text mining capability directly translates into the accuracy of the reports. Another issue is that Reports are too generic and may not translate to how an enterprise wants to view
  • Predictive modeling is still a poorly budgeted initiative, though the most effective finale to the effort of Reporting. Reason for this is the poor quality of data and reports in the previous steps

Why Social Analytics Is Still Only for The Big and the Rich?


  • Social Analytics assumes that a brand is spoken about in the digital media. By reasons of business size and number of transactions, only the large brands with large customer base or transactions are more likely to have customers talk more about them. Except where review sites specifically list businesses to be rated by customers that may include small businesses too.
  • The expense involved in Social Analytics is so high that only the Big can make a budget for it
  • Once into Social Analytics, there is no looking backward. One also needs to put up a Social monitoring as well as Response system in place to manage the customer voice.

In conclusion


Its better to avoid the one-size-fits all or full services tools and take  a more personal responsibility in driving each stage, if necessary with a different vendor partner. Each stage, in our opinion is a business in itself and we should see segments of vendors in each of these areas.

  1. Local Data Collection relevant to one’s geo is important. So, any use of world class tool from another geo needs to be supplemented by manual crawling through the use of local programs. Many businesses are emerging in the data collection space and we should see more comprehensive solutions soon.
  2. Data cleansing has to almost always involve manual intervention until such time that text mining automation reaches maturity. A team would be needed for the same. Here again specialist vendors are a better choice than trust the social listening tool provider.
  3. Reporting is best done by the inhouse reporting engine including any visualization tools. The reports need not come from the Social listening tools.
  4. Use of Predictive analytics using the social data needs adoption and maturity. Again this is left best to the Predictive Analysts who specialize only in this area.


Contributed by N. Ramamurthy Krishnan, Researcher, Digital Media


Please send any feedback or queries to content@csat360.com


Do follow us on


Find what you are looking for