Banner Image

Business and Technology Insights

Understanding, Appreciating, and Evaluating Unstructured Data

 
December 13, 2016

Companies often struggle to make smart business decisions using the data that theyve collected. They see the value in investing in analytics, and are spending heavily to gain a competitive advantage. Several challenges prevent them realizing the best return on that analytics investment.

Many fail to understand the key metrics they should be measuring and they struggle to identify and evaluate the right subsets of the data they target.

Most companies fail to tap into social media or mobile data for new insights. Theres a huge amount of untapped unstructured data such as text, images, voice calls that cannot be ignored. A thorough analysis of this data may lead to improving customer satisfaction and brand image and also provide cues for new products.Soumya_TechTarget

But analyzing unstructured data adds complexity to the analytics project. This information comes from multiple channels and not all unstructured data is relevant and reliable. The thumb rule is to review social media posts for signals pointing to particular issues, and then collect more data around each issue through targeted questions or survey.

In the end, a lot of analytical data gets trapped in business silos. To enable better decision-making you need to be get the data that matters to the managers who can make best use of it. Sales and marketing data may not be available to the Quality control and management group. Thus, while the later does receive the social signals, these signals are often not taken into account in real time while designing the marketing strategies. Retail companies who do it have reported huge benefits from aggregating data from multiple channels and also overlaying social-media content with it. Yet, large manufacturing companies like those manufacturing automobiles or computers have not been able to incorporate this into their decision-making pipeline so successfully.

To learn more about how to derive value from analytics please read my essay in the latest edition of Perspectives.

Dr. Lipika Dey, Prinicipal Scientist, TCS Innovation Labs, New Delhi Dr. Dey graduated with an Integrated M.Sc. in Mathematics from the Indian Institute of Technology (IIT), Kharagpur, and holds a master’s degree in Computer Science and Data Processing. A PhD in Computer Science and Engineering, her interest areas include text data analytics, information fusion, machine learning, and data mining.