Using NLP to Produce Retail Intelligence from Social Media
GUEST POST BY JOHN DAVIS, CEO OF NOTICE AND COMMENT (N&C)
Many retailers are aware of the value in mining “structured” data from business spreadsheets, to obtain useful intelligence about their operations and customers. Fewer are familiar with the wealth of insights to be elicited from “unstructured” data such as email and social media, using Natural Language Processing (NLP) technology.
Originally developed by the national intelligence community to identify and analyze potential threats from vast amounts of “unstructured” data, tech developers in the last several years have adopted NLP methods to generate insights about a product or service provider’s customer preferences and opinions.
Social media content typically offers more real-time insights into market trends and tastes than any other data source. Consequently, technology developers are hard at work creating tools to obtain rapid business intelligence from an ocean of online market buzz and chatter – positive, negative, or somewhere in between.
Comment sentiment analysis of social media can indicate the trending popularity of a retailer’s brand, product selection, or customer support in comparison with the competition. What’s more, correlating social media sentiment to recent changes in a retailer’s product offerings, pricing, or other key market differentiators, can provide almost immediate feedback to inform smart course corrections in strategy.
Hundreds of vendors have emerged in the last few years, offering NLP-based analytic solutions to support marketing and business intelligence-gathering. Their methods vary, from analysis of individual words, entities, phrases and sentences, to relationships and events.
I won’t endeavor here to name or review any specific vendors in the NLP social media analytics space, but I will say the most effective applications typically combine several methods. Filtered keyword search, social graph analysis, and sentiment analysis are generally considered among the most useful. Some that account for the entirety of conversations provide greater contextual analysis.
The best NLP tools on the market can provide a competitive advantage, but it’s not always a simple matter. Good analysis requires subject matter expertise – in this case a retailer’s accumulated experience and knowledge. Some of these expert insights can be automated, which really amps up the power for a retailer in analyzing social media for valuable business intelligence.
No matter how powerful NLP applied to social media proves to be for retailers – and its promise is great — it’s not likely to completely replace human perception and ingenuity any time soon.
For now, NLP will likely serve as a valuable new tool for mining insights about where to conduct further market research. Retailers will still need to rely on their experience, knowledge, and instincts to succeed in a highly competitive market environment for the foreseeable future.