Looking back in time it’s interesting to see how the Industrial Revolution, which was essentially based on the progression of technology, totally transformed the world as we know it today. And turning the clock forward, Human-in-the-Loop systems/workflows are paving the way for the future.
Since the early days of social media analytics the biggest challenge has been around understanding the true intention of posts, their overall meaning, who’s behind the scenes, and the next steps whether it be to engage and hopefully produce a sale or be alerted to something brewing on the horizon.
Extracting these valuable insights is only possible if you intimately understand what’s being said and that just cannot be done by looking at simple combinations of keywords that many times produce a flood of false positives/negatives. And that’s why large organizations not only use social media analytics tools, but also have teams dedicated to sifting through data to find the needles in haystacks.
Imagine being a Fortune 500 company paying $75k/year for an enterprise social media analytics tool, which is manned by humans, and must operate 24/7/365 to ensure nothing is missed. With the need for 2 or 3 full-time employees per day (in communications, marketing, PR, security, etc.) to manage global operations at an average salary of $100k/yr that’s a minimum investment of $375k+ per year. Then you have to cross your fingers and hope the most critical things can be manually found in time.
On the flip side, what if 90+ percent of the manual labor (repetitive tasks) like looking at social media posts that triggered false alerts could be labeled as such by humans, models learn from that feedback, then the accuracy and quality of alerts continues to improve to the point where humans aren’t doing the heavy lifting. That’s the magic of Human-in-the-Loop systems.
The times are changing and the old school world of scratching the social media surface is out the window.