Rohini Srihari has written a great post on how NLP technology could be used for marketing. It is indeed clear, that sentiment analysis by itself cannot be a source of signal and be the base for a decision making. And this is not only because understanding sentiment is algorithmically hard, it is also because there can be irony involved. Or the sentiment is about author’s life circumstances, except that expressed in social media while sitting in coffee shop #name.
What NLP technology can offer (and our products for English, Russian and Chinese do) is:
- Sift sentences / mentions of the target product / person / brand / business branch.
- Detect sentiment orientated towards these targets.
- Report back in aggregate and / or detailed form.
Having other key information bits at hand, such are age, location, popularity (number of followers, engagement level with author’s posts) companies could leverage this to understand where is the most negative folks and why, and what audience is the happiest to the moment. This helps marketing do direct interaction, lead positive momentum, solve particular tricky cases and satisfy their customers.
While NLP technology cannot offer human level interpretation of written free form texts, it can really quickly pass through the vast avalanche of data pieces and give you a sense of general sentiment about your brand plus give you all the access to particular data portions of interest. Especially, when you think of getting in direct contact and use social media as *the* channel for the customer relationship management.
It is time, when using NLP as a service opens up so many opportunities to integrate into your existing software systems and get much socially closer to the market you sell for.