As a child, I could sometimes be found crawling around the unfinished attic of our house. It didn’t even have floorboards. What was I doing up there? I was adjusting the TV antenna. My father would be stationed in the basement, in the family room, a walkie-talkie in hand, his eyes glued to the flickering screen of our Zenith TV set, instructing me through the crackling speaker which way to turn the antenna in order to achieve the clearest picture. Slithering along 2x4s from one end of the house to the other, I was more concerned about losing my balance and plummeting through the ceiling than about achieving optimal picture quality for “The Carol Burnett Show,” “The Bob Newhart Show” and “All in the Family.”
Our mission in adjusting the antenna in those bygone days of over-the-air broadcast television was to root out static by improving the signal-to-noise ratio. It’s the same mission every company is working to achieve today when it comes to the art and science of social media monitoring and analysis. In this context, improving the signal-to-noise ratio means honing in on the small fraction of consumer-generated posts and comments that are relevant to the brand and that may yield actionable insights.
Simple monitoring systems use basic search tools to identify and segregate posts based on keywords. More sophisticated systems use computational linguistic algorithms or natural language processing systems that break the posts down structurally (paragraph, sentence) and grammatically (subject, verb, object), providing a deeper understanding of what is being said – and by whom – and also enabling unexpected topic discovery. What bubbles up from consumer conversations may be completely counterintuitive and unpredictable.
Eliminating noise is not easily done, even with the ability to tune the analytics algorithms. Sometimes the challenges are unique to the brand. Just last week, for example, a leading solution provider was telling me about his team’s experience monitoring social media for Bank of America, commonly known as BOA. Of course, there is no shortage of online consumer conversation about BOA. The challenge in this case was that “boa” also means “good” in Portuguese. Every morning and evening, millions of people in Brazil and Portugal would tweet “good morning” and “good night,” causing the conversation volume on BOA’s dashboard to light up like a Roman candle.
It’s a noisy world. Adjusting the signal-to-noise ratio is critical when it comes to social listening – just as it was critical when it came to watching 1970s sitcoms in our family room.