The marketing world has been buzzing this week about a new study from social superstar Coca-Cola. Despite the brand’s massive social following and celebrated dedication to highly-sharable content, recent analysis by the Coke team found that online buzz had no measurable impact on short-term soft drink sales.

Many clung to the study’s findings that TV and display ads showed more sales correlation than social chatter, questioning whether marketing’s increasing focus on creating conversation has value at all. In response, Wendy Clark, Coke’s Senior Vice President, Integrated Marketing Communications and Capabilities, published her own op-ed, encouraging marketers to drop the narrow-mindedness of viewing media channels in isolation. She writes,

"A recent study from my own company suggested that social buzz or chatter does not generate sales lift. And, taken in isolation, this is true. But, today’s progressive marketers know better.

None of our plans are simply social, or TV, or mobile, or experiential. On the contrary, it’s the combination of owned, earned, shared, and paid media connections – with social playing a crucial role at the heart of our activations – that creates marketplace impact, consumer engagement, brand love and, brand value… No single medium is as strong as the combination of media."

We couldn’t agree more.

And there’s another, equally important point that must be made: “Buzz” is not sentiment. And the latter’s value vastly exceeds the former’s.

Structured sentiment data vs. unstructured buzz

Buzz refers to the volume of chatter around a brand. “How many people mentioned Coke today?” Generating this conversation around your brand is a worthy aim. But pulling meaningful learnings from general brand mentions is difficult. Coke’s own studies confirm it: The brand had both a team of human raters and software from a vendor analyze 1,000 social media mentions for sentiment. They found that, when a human would rate a mention as positive, the software labeled it negative about 21% of the time.

Structured sentiment data, on the other hand, is quite easy to reliably boil down into concrete, actionable insights, even for software. By sentiment data, I mean consumer conversations with a quantifiable sentiment indicator attached – a rating, a like, a helpfulness vote – something to indicate how the consumer really feels, that is unimpeded by things like slang and sarcasm.

So while it’s always good to get people talking, chatter can’t be your endgame. Deeper value comes from analyzing conversations for trends – and acting on those trends to improve your business. Sentiment data is proven to help brands:

Increase sales. Yep. Shoppers exposed to reviews and Q&A on product pages show 161% higher conversion rate and 195% higher revenue per visit.

Identify advocates. Finding your best customers isn’t just about who talks about your brand most – it’s about who has the most positive things to say about you and your products. Analyzing sentiment data reveals which people are highest on your brand, and which are voted the most helpful, so you can target these advocates with personal experiences or special promotions – showing them you appreciate their advocacy.

Improve segmentation. Which types of customers love your products most, and which are less enthused? Which products score best among Millennials, but dissatisfy Baby Boomers? Sentiment data reveals which segments you should aim toward certain products, and which you should point elsewhere.

Build better products. What’s the one thing people who rate your product four stars wish it did differently? Is there one feature that would take a product from four stars to five? And that two-star product: Can it be saved, or should it be scrapped? Insights from structured sentiment data inspire and guide R&D teams, letting them know exactly what customers wish they would build – and then build it.

Don’t stop trying to get people talking. Don’t stop creating highly-sharable and talkworthy content that doesn’t necessarily tie directly to a sale. But be sure you’re capturing data that can give you true insights – data that goes beyond how often consumers mention your products, and reveals how they honestly feel about them.