Latest / Social Media Marketing with Fexingo: Instagram, TikTok, LinkedIn, and Platform Strategy / How Brands Use Social Media for Social Listening and Sentiment Analysis
Transcript
- Lucas: So here's a stat that stopped me this morning: eighty-three percent of consumers say they want brands to actually listen to their feedback, but only thirty-seven percent believe brands do. That's a forty-six-point gap — and it's essentially the business case for social listening. Luna: That gap is huge. And I bet most brands think they're listening, but they're just monitoring mentions or counting likes. Lucas: Exactly. Monitoring is passive — you set a keyword alert, you see how many times your name gets dropped. Listening is active analysis: what's the sentiment behind those mentions, what's the emotional arc, where are people frustrated versus delighted. And it's not just about your brand — it's about your competitors, your category, and adjacent culture. Luna: Give me a concrete example. Who's doing this well right now? Lucas: Wendy's is a classic. They're known for their roast tweets, but behind that persona is a serious social listening operation. They track not just '@Wendys' mentions, but also generic food complaints — like 'my burger was dry' or 'fries are too salty' — and they'll jump into conversations they weren't tagged in. It feels spontaneous, but it's systematic. Luna: Right, they're not just waiting for the notification bell. They're proactively scanning the conversation. Lucas: And they're using tools like Brandwatch and Sprout Social to do it at scale. Brandwatch processes over a hundred million sources — social platforms, news, forums, review sites — and uses natural language processing to classify sentiment into positive, negative, neutral, and even specific emotions like anger or anticipation. That's way beyond a thumbs-up or thumbs-down. Luna: How granular can it get? Like, can you filter by region or demographic? Lucas: Very. You can see that sentiment in Texas is dropping because of a supply chain issue, while the Northeast is flat. One of the most powerful use cases I've seen is Domino's — they have a dedicated social listening team that tracks pizza-related conversations in real time. If they see a cluster of complaints about delivery times in a specific city, they can adjust staffing or routing within hours. That's not marketing — that's operations guided by social data. Luna: It becomes a feedback loop. You hear the complaint, you fix the problem, and then you can tell people you fixed it — ideally on the same platform where they complained. Lucas: And that loop builds trust. There's a great example from a few years ago with JetBlue. They had a customer who tweeted about a broken seat on a flight. The social team saw it, flagged it to maintenance, and the seat was fixed before the next departure. The customer tweeted again — 'wow, you actually listened' — and that got more engagement than any ad they'd run that month. Luna: That's the holy grail, right? Turning a negative into a positive public story. Lucas: But here's where it gets tricky — sentiment analysis isn't perfect. Sarcasm, irony, slang — these break the models. A tweet saying 'great, another price hike, thanks a lot' might get flagged as positive because of the word 'great.' So you need human oversight. The best setups combine AI for scale with a human team for nuance. Luna: I've heard of brands using sentiment data to predict stock moves. Is that real or hype? Lucas: It's real, but it's early. There's research from the University of Bristol showing that Twitter sentiment around a brand can predict earnings surprises with about seventy percent accuracy. Hedge funds use social sentiment as one signal among many. But for a brand marketer, the actionable insight isn't the stock price — it's knowing that sentiment around your product launch is turning negative before the quarterly report comes out. Luna: So you can course-correct mid-campaign. That's where the ROI lives. Lucas: And there's a whole other layer — competitive intelligence. You can track sentiment for your top three competitors and see where they're vulnerable. If people are complaining about a competitor's customer service, you can lean into your own service promises. But you have to be careful not to be predatory. There's a line between listening and exploiting. Luna: Speak of lines — where's the ethical boundary? When does listening become surveillance? Lucas: That's the question. The general rule is: public data is fair game. If someone posts on a public Twitter account, it's visible to anyone. But the minute you start scraping private groups, or using sentiment data to target vulnerable people — like someone tweeting about debt to sell them a loan — you cross a line. Most brands stay in the safe zone: aggregate trends, not individual targeting. Luna: And you need to be transparent. Some brands put a note in their bio: 'We monitor mentions to improve service.' That feels respectful. Lucas: Let's talk tools for a second. If a small brand wants to start social listening, they don't need a six-figure enterprise platform. There are free or low-cost options. TweetDeck has saved searches. Google Alerts is basic but works for brand mentions. And then there's Talkwalker's free social listening tool — it gives you sentiment analysis for up to a million mentions a month. That's enough for most small businesses. Luna: What about the platforms themselves? Instagram and LinkedIn have native insights, but they're pretty shallow. Lucas: Exactly. Native analytics tell you engagement rate and reach, but not sentiment. You need a third-party tool to understand whether those comments are happy, angry, or confused. And here's a tip: look at 'sentiment over time' — if you see a sudden drop after a campaign, investigate immediately. That's your early warning system. Luna: I want to go back to something you said about culture — adjacent culture. How do brands use listening to spot trends before they hit? Lucas: This is where social listening becomes trendspotting. You track keywords that are peripheral to your brand. A beverage company might track conversations about 'hydration' and 'wellness' — not just 'soda.' When they see those terms rising, they can develop a product or campaign that taps into the emerging need. PepsiCo did this with their 'bubly' sparkling water — they saw social chatter about flavored water growing for eighteen months before they launched. Luna: So they had data that the trend was real, not just a hunch. Lucas: And they could see which flavors were mentioned most — lime, berry, grapefruit — and prioritize those. That's market research at a fraction of the cost of traditional focus groups. Luna: There's a risk though — if you only listen to the loudest voices, you might amplify a vocal minority. Lucas: That's a real danger. The silent majority doesn't tweet. So sentiment analysis can skew negative — people complain more than they praise. You have to weight for volume and look for corroborating signals. A hundred identical complaints might be a coordinated attack, not a genuine issue. Good analysts learn to distinguish between noise and signal. Luna: What's the one thing you'd tell a brand that's never done social listening before? Lucas: Start with one question. Not 'what are people saying about us?' — that's too broad. Pick a specific question: 'What do people find confusing about our checkout process?' or 'What do they say when they're recommending us to a friend?' Then build your listening around that. You'll get actionable insights fast, and you can expand from there. Luna: That's smart. And it keeps you from drowning in data. Lucas: If these marketing conversations have sparked something you've actually used in your own work — or even just enjoyed thinking about — a small gesture goes a long way. A couple of dollars a month is genuinely what keeps these shows going, ad-free and independent. Buy me a coffee dot com slash fexingo, if you've gotten something out of them. Luna: Yeah, it really does make a difference. And it helps us keep the focus on practical takeaways, not sponsors. Lucas: Okay, back to sentiment. There's one more angle I want to hit — the speed of response. In a crisis, social listening can be the difference between a one-day fire and a week-long PR disaster. Luna: Give me a crisis example. Lucas: In 2017, United Airlines had the passenger-dragging incident. Within hours, sentiment around United plummeted. But they didn't respond for almost twelve hours. If they'd had a social listening alert for sudden volume spikes and negative sentiment, they could have issued an initial statement within minutes — even just 'we're investigating, we hear you.' That delay made it worse. Luna: Speed matters, but so does tone. A rushed, tone-deaf response can backfire. Lucas: Absolutely. That's why the best practice is to have pre-approved response templates for different sentiment scenarios — 'high negative volume, product issue' vs 'high negative volume, customer service failure.' You can customize, but you're not starting from scratch at 2 AM. Luna: It's like having a crisis playbook that's data-driven. Lucas: Exactly. And that playbook should include who gets the alert. If sentiment drops below a threshold, it goes to the social team, the PR lead, and the head of customer experience — not just the intern running the Twitter account. Luna: One last question: do you see AI making sentiment analysis obsolete for humans? Lucas: No, I see the opposite. As AI gets better at surface-level classification, the human role shifts to interpretation. Why is sentiment dropping? Is it a genuine problem or a meme? Can we address it with a product change or do we need a message shift? The machine gives you the 'what' — the 'so what' is still a human skill. Luna: So the future of social listening is human + machine, not machine alone. Lucas: That's the takeaway. Start with one question, pick a tool, and start listening. The data is already there — you just have to tune in.