James Purchase is VP of product management for Attensity, a pioneer in natural language processing and sentiment analysis.
For years, brands have been mining social networks to study consumer metrics, including check-ins, likes and retweets. Combined with census-like data – gender, age, location, etc. – brands have created countless dossiers on consumers and audiences in the name of improving service and sales. Dubbed “listening,” the approach has been passive and objective, a mathematical equation of huge amounts of publicly-shared content about a brand.
Living in the past will only get brands so far
With time, brands found that social metrics weren’t as useful as they seemed. Everything that was tracked had already happened: a restaurant meal posted after the fact; vacation pics shared after the trip; the movie critique (no spoilers, please!) of a film already watched… and so on.
This after-the-fact view left out a huge piece of the marketing and customer-service puzzle: a contextual understanding of how people feel about products. Experts in the field know this as “detailed sentiment,” which is largely considered the golden key to customer service and marketing.
Detailed sentiment analysis outlines that familiarity we feel among our family and friends — the understanding that lets others anticipate what will grab our attention. Combined with demographic information, sentiment analysis tunes in to our needs and wants.
The way forward for brands
Detailed sentiment is the secret ingredient brands need to understand what makes their customers happy. Brands that ignore it will ultimately lag behind those that do.
Consider, for example, the phrase: “Dang, that phone is the bomb!” or “Beyonce’s Grammy Awards outfit is sick!” Both of these statements have colloquial meanings that – while most au courant individuals would recognize as strongly positive – most data-crunching, social-analytics algorithms perceive as strongly negative. The danger here is marketers using incorrect data.
So how can brands accurately measure detailed sentiment without employing an army of high-priced data analysts? Most technology can only go so far in interpreting data — usually by assigning ‘positive’ or ‘negative’ earmarks. But next generation Natural Language Processing (NLP) has surfaced recently and is able to interpret human conversation.
How Natural Language Processing works
Next gen NLP merges advanced text analytics with big data to help brands become more emotionally intelligent. Designed specifically to decipher the nuances of language, it is an important tool available for understanding the voice of customers — a brand’s best source of trend leads.
Groups of businesses are already forming around this technology, since it allows them to harness customer conversations from a variety of channels, including social media, surveys, contact center notes, chat, voice, text and email. The sheer amount of passion and emotion online makes NLP indispensible to understanding the contextual information available on social channels.
Add in a deeper layer of structured data about your customers — such as a recent graduation, marriage, job change, online searches — and you begin to see a contextual and more humanistic understanding that produces detailed sentiment awareness.
Why does all this matter?
With the ability to search for patterns and focus on specific areas of interest, algorithms employing NLP give businesses foresight on when noteworthy events are taking place, like a sudden surge in negative or sarcastic language coming from a specific location. This allows brands to build strategy around story arcs based on a more complete view of the customer.
Intellectually, we know that it is in a brand’s best interest to interpret digital commentary in a human way, and NLP is delivering on that need today. As the volume of data increases online, the future of better marketing and customer service will depend on listening programs with the goal of applying emotional intelligence to their business strategy.