The algorithm erodes the travel recommendation: what comes after?

The shape of the problem is not that people are discovering new places – that’s wonderful, truly – it’s that the recommendation engine has become a blunt instrument being wielded on very delicate things around the world.

The algorithm doesn’t care about any of this since it’s optimised for engagement rather than for experience. It is a machine that has never felt the specific pleasure of a hidden courtyard, or the peculiar joy of a local market stall, or the complete disaster of turning up somewhere that 40,000 people also “discovered” that week. The potential of social media, as an organic force for real-world discovery, has turned sour. Or, as Sam Blenkinsopp, co-founder of Trippin, puts it, “the promise of authenticity” born from creator-led recommendations has collapsed “into an oversaturated sea of midness that has flattened taste globally”.

So if TikTok has failed to get us closer to more human travel recommendations, where should we turn? What does the alternative actually look like – and who still deserves our trust? The rest of this piece tries to answer those questions, drawing on people who have been thinking carefully about what makes a recommendation worth following, and what travel could look like if the industry got this right.

Sam Blenkinsopp built Trippin a decade ago on exactly the premise that the best recommendations come from real people – “People actually creating culture on the ground like DJs, chefs, designers, artists.” The result of ignoring that, he says, is what urban theorists call AirSpace: a global monoculture of exposed brick and Edison bulbs, where you can traverse half the planet and have the same experience you’d have at home.

“What people want,” says Sam, “is to understand the rhythm of a place. How it feels to spend time there. Whether they can fit into it, even just for a few days. That’s a very different brief to a drone shot of a perfect family on a sunbed in Trinidad. The real Trinidad is the lady who’s been making doubles for her community for years, with a hot sauce you won’t find anywhere else.”

Meanwhile, the truly useful recommendation, Sam argues, requires specificity most travel brands “won’t touch”. “People want to know about the political reality of a place. Whether they can be openly gay. Whether it’s safe for people of colour.” Not to mention how to support local businesses, choose sustainable accommodation and actively benefit the place they’re visiting. These aren’t just niche concerns, he says – “They’re central to how a younger generation makes travel decisions.”

Both voices touch on the distinct elements required of a good recommendation. For one, format matters: “We’re seeing a convergence of inspiration and utility,” says Matthew. “Content needs to offer both a compelling narrative and be genuinely useful. For us, that might mean integrating itineraries, maps or practical detail into more editorial storytelling.” For another, quality derives from purpose. “Travel isn’t perfect,” Sam says. “But it can be less extractive. And when it’s done right – when people are thinking about how they move through spaces, how they respect local customs, where their money goes [...] tourism [can be] a genuine force for good. That’s what a great recommendation should be working towards.”

What connects these approaches is a shared thesis that the antidote to algorithmic spectacle isn’t better spectacle. In fact, it’s smallness. The beach with no Google Maps Pin or the unreplicable hot sauce. These aren’t consolation prizes for people who missed the thing everyone came for, they are literally the whole point. The recommendation that earns trust is the one that understands that.

After all, the travel recommendation crisis isn’t really about social media; it’s about who we listen to – and whether they’ve done anything to deserve it. A viral video with half a million views is not a recommendation, it’s actually crowd instruction. And if we keep following the crowd, we’ll keep arriving to find the carols cancelled and the local gem replaced by another version of somewhere we’ve already been.