It’s All Semantics
Digital semantics is the next horizon in reputation analytics.
The question of how to use the notoriously fickle indicator of sentiment has plagued the digital analytics field since inception. Some heavily caveat use of sentiment, create custom frameworks or throw it out all together. It comes up in project pitches, vetting interviews and fine print. Many start up companies have been founded and perished on promises of solving this puzzle.
Here’s the answer: let’s evolve.
I’ve worked with Red Bull, where “sick,” and “wicked” are high praise, and with Zappos where the accusation of selling $/%# products was turned into a rallying cry. Numerous crisis clients have seen ordinary words used as smears against them in social memes. Sarcasm is a train wreck for sentiment analytics and humor is not much better.
Past models have taken four main forms:
- using existing sentiment analysis products with heavy warnings;
- creating custom lexicons that are applied via manual processes;
- analyzing a sample set with human resources; or
- machine-toned sentiment analysis that is trained by a human.
None of the options are good. However, don’t despair. We can now leapfrog this whole sentiment morass by using emerging advanced analytics capabilities to understand how words are being used in context.
Semantic analysis involves using machine learning techniques to score how words are used in the wild and find synonyms that are being used in the same way. Machine scoring quantifies the sentiment of the word and how it was being used by the given commenter. This opens up a whole new, and more accurate, world of understanding what consumers or the broader public is thinking on an issue.
The context of consumer language reveals how they think, why they said what they said, and with whom they intended to communicate. The ability to apply numeric pattern recognition enables brands to compare their own messaging with the way their consumers, brand fans and category influencers talk about key market themes. Brands can now unlock insights into how to align with audiences and understand when, where and how to connect in order to maximize impact.
How do you currently use sentiment analytics?
Let Us Know in the Comments Section Below!