Using text analytics and social media analytics programs for sentiment analysis can help companies identify risks to their brands early, as well as perceptions of their products and potential target markets, says analytics consulting company Olrac SPS South Africa senior technical consultant Klay Martens.
Sentiment analysis can form part of a company’s risk management strategy and can help, it identify perceptions of it among people and businesses, enabling a company to take corrective action or engage with customers on a pressing issues.
This analysis can also be used to identify gaps in the market or to identify competing products prior to development and to improve the success rates of products and development, he adds.
Some social media are easier to analyse than others – for example, feeds from social network Twitter are more easily analysed for sentiments because they are more concentrated and focused than searching through social network giant Facebook or posts on commercial network LinkedIn.
There are strategies to deal with the differences in how information is gleaned from social networks, such as using the Boardreader program that collects new public posts on social networks, enabling the text analytics program to analyse the posts.
However, the volume of data that must be analysed is significant and the processing requirements of text analytics are a key factor when determining which information to analyse.
“Sentiment analysis is a powerful tool in the marketing industry. Companies, especially service companies and large companies, can use the tool to identify risks to their reputations or problems raised by customers early to manage expectations of customers and take remedial action,” he says.
The system can also be used to assess the sentiment of different suppliers or service providers in a company’s supply chain, as well as to identify potential areas of the market that are underserved.
However, while South Africa’s social networks are not as large as those of more developed countries, they provide a broad spectrum of information that companies can use to inform their strategies, says Martens.
Linked with other programs and data, the geographic location of sentiments can be determined, enabling companies to analyse sentiment according to region and to distinguish between beneficial and risky regions.
Olrac SPS can advise companies on how to implement sentiment analysis programs, he adds.
The programs can interpret natural languages and a wide range of different languages, as well as scan and identify key concepts much quicker than a human can. Very abstract concepts present a bigger problem, but the program can interpret most factual concepts.
“Text analytics and the development of natural language processing were driven by the need to enable machines to interpret large volumes of text and extract structured information. This enabled the text to be reported on or analysed using mathematical algorithms,” concludes Martens.