Latest / Building Brands with Fexingo: Identity, Reputation, and Long-Term Business Equity / How Spotify Built a Brand on Personalization and Data
Transcript
- Lucas: So we talk a lot on this show about brands that are built on a clear product philosophy or a founder's vision. But there's another route to brand building that's become dominant in the last decade — and that's building a brand on personalization and data. And the clearest example I can think of is Spotify. Luna: It's funny — I don't think most people think of Spotify as a 'data brand.' They think of it as a music app. But the way they use data is central to how people feel about it. Lucas: Exactly. Spotify started in 2008 as a legal streaming alternative to piracy. Back then, the value prop was simple — access to millions of songs for free with ads, or a subscription without them. That got them in the door. But the brand really took off when they started to personalize the experience. And the key moment was the launch of Discover Weekly in 2015. Luna: Right — that was the first big algorithmic playlist that felt like it actually knew you. I remember people tweeting screenshots of how eerily accurate it was. Lucas: Exactly. Discover Weekly used collaborative filtering — basically, the algorithm looks at what you listen to and compares it to what similar users listen to, then recommends songs you haven't heard. It became a huge driver of engagement and retention. By 2018, Spotify reported that Discover Weekly had over 40 million unique users per week. And it wasn't just a utility — it became a talking point. That's when the brand started to feel personal. Luna: But Discover Weekly is just one feature. The real brand explosion came with Spotify Wrapped in 2016. I mean, that thing is basically a yearly cultural event now. Lucas: It really is. Spotify Wrapped took personalization and turned it into a social moment. In 2025, over 120 million users shared their Wrapped results on social media. That's free marketing, but more importantly, it makes users feel like the brand is reflecting their identity. 'I'm in the top 0.5 percent of Taylor Swift listeners' becomes a badge of honor. And that emotional connection is incredibly sticky. Luna: There's a risk though, right? If you lean too hard on algorithmic personalization, you can lose the human curatorial touch. Some people complain that Spotify's playlists can get repetitive or too narrow. Lucas: That's a real tension. Spotify has tried to balance it with human-curated playlists like 'RapCaviar' or 'Today's Top Hits.' But the algorithm is the core. And they've been smart about expanding beyond music. Their podcast push — starting around 2019 with acquisitions like Gimlet and Anchor — used the same data playbook. They recommend podcasts based on listening history, and they now have over 5 million podcast titles on the platform. Luna: And then audiobooks too. In 2022, they added audiobooks for premium subscribers. So they're becoming a full audio ecosystem. But the brand still hinges on that feeling of 'Spotify knows me.' Lucas: And that's the core insight for any brand builder. Spotify doesn't just sell access to music. It sells a personalized identity. The data is the product, in a way. But the brand is the feeling of being understood. And they've been remarkably disciplined about not exploiting that data in creepy ways — at least not yet. Their privacy policy and data usage have faced scrutiny, but they've avoided major scandals that would erode trust. Luna: That's a good point. Trust is fragile. If users felt Spotify was selling their listening data to advertisers in a way that felt invasive, the whole brand could collapse. So far, they've walked that line well. Lucas: They have. And they've also been smart about making the brand feel human despite the algorithm. Their marketing is quirky, a bit irreverent. They do these out-of-home ads that are data-driven jokes — like 'Dear person who played 'Sorry' 42 times on Valentine's Day, we hope you're okay.' That kind of thing makes the brand feel like a friend, not a machine. Luna: That's actually a really interesting example of using data for brand warmth rather than just efficiency. A lot of companies have user data, but they don't use it to create that emotional resonance. Lucas: Let's talk about the business numbers for a second. As of mid-2026, Spotify has over 650 million monthly active users, with about 250 million paid subscribers. That's a conversion rate of around 38 percent from free to paid — which is remarkably high for a freemium model. And their revenue is annualizing around $15 billion. The brand is a huge part of that growth. Luna: And they're profitable now too, right? I think they turned the corner on profitability in 2024 after years of heavy investment. That's a big milestone. Lucas: Yeah, they posted their first full-year operating profit in 2024, and it's been growing since. But the brand faces challenges. Competition from Apple Music, Amazon Music, and YouTube Music is intense. And there's the ongoing tension with artists over royalty rates. Taylor Swift pulling her catalog in 2014 was a big deal. Spotify has tried to improve relationships, but the perception that they underpay artists is a brand liability. Luna: That's a good segue to a broader question: can a brand that's built on personalization survive if the personalization starts to feel manipulative? I'm thinking about TikTok, which also uses algorithms but has a very different brand vibe — more addictive, less warm. Lucas: That's a crucial distinction. TikTok's algorithm is optimized for engagement and time spent, often at the cost of user well-being. Spotify's algorithm is optimized for discovery and satisfaction. The brand is built on positive reinforcement, not dopamine loops. That's partly why Spotify Wrapped feels celebratory, while a TikTok recap might feel like a time-suck audit. Luna: So the lesson for other brands is: if you're going to use data to personalize, the goal should be delight, not manipulation. Lucas: Absolutely. And it's worth noting that Spotify has invested heavily in AI too. In 2024, they launched an AI DJ feature that uses generative voice to introduce songs and provide commentary. It's another layer of personalization that makes the experience feel more human, not less. Luna: I've used it — it's surprisingly good. The voice is modeled after Spotify's head of cultural partnerships, Xavier Jernigan. It feels like a radio DJ who knows your taste. Lucas: That's the key — it's not just an algorithm spitting out songs. It's a curated personality. And that's something any brand can learn from: don't just personalize, anthropomorphize. Give your brand a voice, a personality, a sense of humor. Data can fuel that, but it shouldn't replace it. Luna: Alright, so if I'm a brand manager listening to this, what's the one concrete takeaway? How do I start building a data-driven brand without becoming creepy or cold? Lucas: I'd say start with one personalized touchpoint that surprises users in a positive way. It doesn't have to be as complex as Spotify's algorithm. Maybe it's a birthday email that references a purchase they made a year ago, or a recommendation based on a recent interaction. The goal is to show you remember them as an individual. Then, make that moment shareable. That's what Wrapped did — it gave users something to post. That's when personalization becomes brand building. Luna: And be transparent about how you use data. Spotify has a 'Privacy Center' where users can see what data is collected and control their settings. That builds trust. Lucas: Right. And one more thing: don't let the algorithm run everything. Keep a human hand in the mix. Spotify still has human editors for major playlists. That balance between machine efficiency and human taste is what makes the brand feel both smart and warm. Luna: I think that's a great note to end on. Spotify is a fascinating case study of how data can build a brand — but only if it's wielded with empathy and creativity. Lucas: And speaking of building this show with care — we've been doing these episodes for over a hundred episodes now, and it's been amazing to see how many of you keep listening. We actually keep the show ad-free by design, and that's possible because a small group of listeners chip in over at buy me a coffee dot com slash fexingo. It's a simple way to support the work, and it really does make a difference in keeping these conversations going. Luna: Yeah, and it's not about big pledges — even a one-time coffee helps cover hosting and research time. We're grateful for every bit of support. Lucas: Alright, back to the topic — I want to leave listeners with one final thought. The next time you open Spotify and it serves you a perfect song, remember: that feeling of being known is the result of a deliberate brand strategy. And it's a strategy any business can adapt — by using data not just to sell, but to connect. Luna: Well said. Thanks for joining us on Building Brands with Fexingo. We'll see you next time.