Acquired Taste
Three companies, three bets on personalization, everyone wants a piece of your taste.
Three stories this week sitting at the intersection of personalization, and the fight for who owns the customer relationship. From credit card wars to teen repeat songs, everyone is building a platform, the question is whether anyone is building trust.
***
I once listened to a white noise song, for six weeks after that my Spotify Discover Weekly was nothing but white noise. I tried triggering the algorithm back by clicking on 100 hip-hop songs, but nothing worked. So I stopped using Discover Weekly altogether.
I must not be the only one complaining, because Spotify just released a new feature at SXSW called Taste Profile. It lets you see how the algorithm understands your listening habits and, for the first time, correct it. You can tell Spotify what you want more or less of, and add context like “I’m training for a marathon” to shape what shows up in your feed.
Taste and profile are two buzzwords overused lately, even Paul Graham wrote a blog post about it. My prediction is that we are going to see more of this across AI-driven platforms. LLMs are predictive and give median answers, so the best way to steer the output toward something personal is context plus memory. That’s what Taste Profile is, your context layered on top of Spotify’s memory of your behavior.
It’s smart for Spotify to get in early. It increases personalization, which its users love most, and turns passive listeners into active data contributors, increasing stickiness and loyalty.
Knowing the success of Spotify Wrapped, I would not be surprised if Spotify manages to make Taste Profile a shareable moment too.
***
Capital One brought on board 150 employees from Hopper and launched a standalone travel app, booking, rewards, lounge access, flight alerts, real-time trip management, all in one place. It’s the next frontier in the high-end credit card battle.
Learnings from Accor, Selina, Sonders, and Sweetgreen, who all wanted to ditch vendors and internalize the tech; it never works. They ended up losing more money and time than anything else, often going back to a tech partner if they hadn’t gone bankrupt in the process.
That said, the strategic shift is real. We are at an inflection point in the premium credit card wars, moving from perks (DoorDash credits, points multipliers) to platforms (integrated travel ecosystems). Capital One’s bet is that owning the digital travel experience creates stickier customer relationships than giving you discounts on other apps you like to use.
I mentioned the super-travel app last week, banks have been investing heavily in travel in the past few years, as wealthy individuals tend to spend more on experiences than physical goods. But their remit seems limited to travel. They are not yet embedded in the day-to-day lives of their customers, so my bet on Google stays.
***
BILT Rewards acquired Sion, a commission management platform used by more than 8,000 travel advisors, for $30 million. It’s a bet on human taste.
Every company is racing to replace humans with AI, BILT is doing the opposite. It’s investing in more than 8,000 people who know and love travel, advisors who have been to the places, who understand the difference between a hotel that photographs well and one that delivers.
While Capital One builds a digital platform to own the travel journey, and Spotify builds a platform for you to do the work, Bilt placed a bet on white-glove services and on proprietary data. It makes sense, AI can legally only scrape public data, so the knowledge of the travel advisors using BILTs platform is something competitors can’t replicate. That gives Bilt a competitive advantage, making its membership more valuable.
But data alone isn’t enough, you also need brand. When everybody zigs, BILT zags, and for that I salute them. The move is smart. Yet I have noticed online that BILT’s users are complaining about the rollout of its new credit card. Winning on the advisor side is good, but it won’t be enough if Bilt fumbles the same trust on the core product.
Everyone wants to “own taste.” What they actually want is to be the last voice you hear before you decide. The Spotify example shows how fragile that is. One wrong signal and it keeps feeding you the same thing. That is not personalization, that is a loop. Letting users fix it is not a feature, it is recovery.
BILT’s bet is different. Taste is judgment, not just data. It comes from being there and knowing what holds up beyond the photos. The risk is simple. The moment it feels off, people stop trusting it. And when trust drops, so does influence.