Latest / Social Media Marketing with Fexingo: Instagram, TikTok, LinkedIn, and Platform Strategy / How Brands Use Social Media for Customer Service at Scale
Transcript
- Lucas: If you've tweeted at an airline in the last five years, you've probably gotten a reply within minutes. That is not an accident. That is a deliberate, costed, headcounted strategy. Luna: And it's not just airlines. Brands across retail, telecom, even government agencies are moving customer service onto social platforms. But is it actually better, or just faster? Lucas: Great question. Let's start with a concrete number. Delta Air Lines runs a dedicated support handle on X — @Delta — and according to industry reports, they handle over 50,000 interactions a month through that channel alone. That's not counting their Facebook page or their WhatsApp bot. Luna: Fifty thousand a month is essentially a small call center's worth of volume. And they're doing it in public, where every mistake is visible. Lucas: Exactly. That's the risk and the reward. When you reply to a complaint in public, millions of people can see how you handle it. Delta's agents are trained to acknowledge, apologize, and move to DM quickly. They don't try to resolve complex issues in the thread because that just drags out the negative visibility. Luna: So the public reply is basically a signal — 'we see you, we care' — and then the real work shifts private. Lucas: Right. And the economics are compelling. A phone call to a traditional support line can cost a company between five and ten dollars per contact, depending on average handle time and overhead. A social media interaction — even a DM conversation — often costs around one dollar. That's a massive savings when you're dealing with millions of contacts a year. Luna: But that math assumes you can handle volume without scaling headcount linearly. How do brands actually do that? Lucas: Two ways. First, chatbots on platforms like WhatsApp and Messenger handle the first tier. Think password resets, order status checks, flight change eligibility. Those are pattern-matchable requests. Second, the platforms themselves have gotten better at routing — X's business tools let you assign conversations by topic, agent availability, even language. Luna: Zappos famously took the opposite approach for years. No scripts, no time limits, no chat bots on the phone line. But they've also integrated social support. How does that fit? Lucas: Zappos is the outlier that proves the rule. Their entire brand is built on service as marketing. A legendary ten-hour phone call with a customer is a story that gets shared millions of times. But on social, they've actually been more pragmatic. They use bots for basic inquiries on Facebook, and reserve their famously empowered reps for complex or high-emotion cases on X and Instagram DMs. Luna: So they're segmenting by channel. Phone gets the full white-glove treatment. Social gets efficiency. Makes sense. Lucas: And then you have a company like Spotify, which has turned customer service replies into a form of entertainment. They have a dedicated @SpotifyCares handle on X, but they also let their main account occasionally respond to complaints with humor. Remember when a user complained that Spotify wrapped was late, and they replied with a custom playlist called 'We're Working On It'? That got millions of impressions. Luna: That's the dream scenario — a complaint becomes a viral moment that reinforces brand personality. But it's also risky. One wrong joke and you're in a PR crisis. Lucas: Absolutely. That's why most brands keep humor on a very short leash. The majority of social care is blandly professional. Delta, for example, has templates for almost every scenario. 'We apologize for the delay. Please DM your confirmation number so we can look into this.' No jokes, no personality. Just speed. Luna: And yet, speed is a personality in itself. If I get a reply in two minutes, I already feel better, regardless of what the reply says. Lucas: That's backed by data. A study by Conversocial a few years back found that 42 percent of customers who complain on social expect a response within 60 minutes. And the brands that meet that threshold see significantly higher customer satisfaction scores than those that take longer. Even if the issue isn't resolved, the acknowledgment matters. Luna: So what happens when a brand just ignores social complaints? We've all seen those threads where a customer tweets at a company ten times and gets radio silence. Lucas: That's a brand that hasn't done the math. Because a single ignored complaint can be screenshotted, shared, and turned into a story. Remember the United Airlines guitar incident? That was a song, not a thread, but same principle. A service failure that goes viral can cost millions in brand equity. Luna: And it's not just big brands. Small businesses on Instagram face the same dynamic. If you ignore a DM complaint, that person might post a negative review or tell their friends. Lucas: Exactly. The bar for social care is actually lower for small businesses — customers don't expect 24/7 coverage from a local coffee shop. But they do expect a reply within a few hours during business hours. And that's achievable with a simple automation: an away message for DMs that says 'We'll get back to you within 2 hours,' and then actually doing it. Luna: I want to talk about AI bots on WhatsApp. I've used them for a few brands — Domino's, KLM. Sometimes they're great, sometimes frustrating loops. Lucas: WhatsApp Business has become a major channel, especially outside the US. In markets like India, Brazil, and Mexico, WhatsApp is the primary messaging app, and brands have invested heavily in chatbots that can handle ordering, tracking, returns, even insurance claims. The key is that the bot needs to hand off to a human seamlessly. If the bot says 'I don't understand' and you're stuck, that's worse than no bot at all. Luna: KLM's bot, for example, can rebook flights and answer baggage questions, but if you type 'I'm angry,' it routes you to a human. That's smart design. Lucas: Emotion detection is becoming standard. Many of the enterprise social care platforms like Sprout Social and Hootsuite now offer sentiment analysis that flags angry or urgent messages for priority handling. So the bot handles the routine, the human handles the rage. Luna: And then there's the question of metrics. How do you measure success in social care? It's not just resolution time. Lucas: Brands look at first response time, resolution time, and customer satisfaction score — CSAT — which they often survey via a quick DM after the interaction. But there's also a newer metric called 'social care efficiency' which is basically cost per resolved interaction. And some brands track 'escalation rate' — what percentage of DMs require a human after the bot. Ideally you want that under 30 percent. Luna: The ultimate goal, though, isn't just to solve problems. It's to turn a support interaction into a loyalty moment. A great service experience on social can be more memorable than a great marketing campaign. Lucas: That's the thesis of the entire 'social care as marketing' movement. And it's why so many brands now have dedicated social care teams that sit alongside marketing, not in a separate call center silo. They share the same brand voice guidelines, the same campaign calendar, the same performance reviews. Luna: If these marketing conversations have sparked something you've actually used — maybe a tactic for your own brand, or a way of thinking about a problem — I think we can both say this show has been worth making. And honestly, if today's episode was worth a coffee to you, that's the link — buy me a coffee dot com slash fexingo. It keeps the show ad-free, and it means we can keep digging into these details without anyone telling us what to say. Lucas: Yeah, we really appreciate that. It's a small thing that makes a big difference for us. Now, back to the future of social care — where do you see this going in the next year or two? Luna: I think we're going to see more generative AI that can handle complex conversations, not just scripted flows. Imagine a bot that can write a personalized refund email in your brand voice, approved by a human with one click. Lucas: That's already happening with tools like Intercom's Fin and Zendesk's AI agents. They can generate full replies from knowledge base articles. The human just reviews and sends. That could drive that cost per contact down even further, maybe to twenty cents per interaction. Luna: And also faster resolution. If the bot can answer a nuanced question without escalating, the customer is happy and the brand saves money. Lucas: But we'll still need humans for the edge cases — the scenarios that require judgment, empathy, or a creative workaround. The best social care teams will be the ones that blend AI efficiency with human warmth, and do it publicly without breaking character. Luna: And that brings us back to the airlines. If my flight is canceled, I don't care if it's a bot or a person — I just want someone to fix it fast. Lucas: Exactly. The channel is just the medium. The expectation is the same as it ever was: solve my problem, and treat me like a person while you do it. Social media just made that expectation visible to everyone. Luna: And for brands that get it right, that visibility becomes their best ad.