What the Barcelona TikTok food creator scene gets right that LA and London get wrong

What the Barcelona TikTok food creator scene gets right that LA and London get wrong

I have spent the better part of May 2026 watching how three cities surface a restaurant on TikTok, and I have come to a conclusion that will annoy a lot of people in Los Angeles and London: Barcelona is winning. Not because the food is better. Not because the creators are smarter. Because the city is the right size. Barcelona has roughly 1.6 million residents inside the municipal boundary and roughly 5.6 million in the metro, which puts it in a network band where a food creator can plausibly know, or be one introduction from, every other serious food creator in town. LA's metro is around 13 million. London's is around 9.7 million. The math does not care how talented anyone is. The math says coordination collapses past a certain size, and recommendations collapse with it.

This is the part where I am supposed to soften the claim. I am not going to. When I pull a list of Barcelona places that keep showing up across multiple creator accounts in our GeoTok dataset, I get places like Rocambolesc, La Madurada, Xopo, Prodigi Restaurant, Nectari Restaurant, La Balabusta, and El Tribut. Seven names, all real businesses, all repeatedly cited by food accounts that also cite each other. When I run the same exercise on LA or London, the lists go in seven different directions and never reconverge. That is not a content problem. That is a network problem.

The midsize-city sweet spot is a real economic phenomenon

There is a body of work on network density that economists and urban theorists have been writing for years, and the short version is: small networks coordinate easily because everyone knows everyone, large networks coordinate badly because nobody can know everyone, and there is a middle band where the network is large enough to be interesting and small enough that signals still cross. Barcelona sits in that band. So does Lisbon. So does Copenhagen. So does, on a good day, Mexico City.

Los Angeles does not sit in that band. The LA food creator scene is technically the largest in the world by raw count, and that is the problem. A creator working out of Highland Park has no functional overlap with a creator working out of Venice. They cover different sets of restaurants, and even when they cover the same restaurant they describe it differently because the audiences are different. I have watched eight LA creators in May 2026 cover the same Sichuan place in the San Gabriel Valley and produce eight different framings, two of which were factually wrong about the menu. Nobody corrected anyone. Nobody had a reason to.

London is the same disease with a different accent. Soho creators cover Soho. Hackney creators cover Hackney. The lines almost never cross. A Camden creator and a Peckham creator are operating, for practical purposes, in two different cities. The TfL map insists they are connected. The creator network insists otherwise.

Barcelona, by contrast, behaves like one creator graph. A food account covering Eixample and a food account covering El Born are not just aware of each other, they are actively responding to each other. Rocambolesc, the dessert spot from chef Jordi Roca, currently sits around a 4-star average across roughly 350 reviews, and it appears in clip after clip from accounts that otherwise do not overlap. That kind of cross-account consensus is the signal you want when you are trying to find a real recommendation. The takeaway: the city's size is not a constraint on the creator scene, it is the feature that makes the creator scene work.

Density beats variety when you are trying to eat dinner tonight

The standard rebuttal here is that LA and London simply have more variety. That is true and beside the point. Variety is a supply-side feature. Recommendation quality is a demand-side problem. When I am visiting a city for four days, I do not need access to ten thousand restaurants. I need ten that are actually good, ranked in a way I can trust. Barcelona's creator scene gives me that. LA's drowns me. London's siloes me.

Look at the shape of the Barcelona list I am working from. Prodigi Restaurant runs around a 4.6 average across roughly 93 reviews and lists itself as European, Catalan, Mediterranean, and Spanish all at once, which in another city would be a red flag. In Barcelona it is just how the cuisine is described. Nectari Restaurant carries around a 4-star average across roughly 409 reviews and lands in the same Mediterranean-Spanish bracket. La Madurada, which is the American-leaning steakhouse on the list, runs about 4.3 across roughly 102 reviews. La Balabusta brings the Mediterranean-Israeli-Middle Eastern angle at around 3.5 across roughly 103 reviews. These are not the same kind of place, but they live inside one creator graph, and the graph can compare them honestly because the same people have eaten at all of them.

I cannot run that exercise in LA. Or rather, I can, but the output is nonsense, because the creator who is plausible about Koreatown is not the same creator who is plausible about Beverly Grove, and the two are not in conversation. London has the additional friction that the food press, the food podcasts, and the food creators are three different ecosystems that barely cite each other. A 4.6-rated Catalan place in Barcelona gets surfaced by the same network that would surface a 3.5-rated Israeli place across town, and that is good for the eater. It means the network is grading on a single scale.

This is what I mean when I argue for the midsize-city advantage in creator ecosystem dynamics. It is not that Barcelona has the best food, it is that Barcelona has the best functioning recommendation graph. And a functioning graph at a 4.0 average will beat a fragmented graph at a 4.8 average every single time, because in the fragmented graph you cannot tell which of the 4.8s are real.

When the same five food accounts keep ending up at the same dinner table, the city stops being a list of restaurants and starts being a conversation.

That is paraphrased from a London-based food writer I spoke to in early May 2026, and it captures the dynamic in a single sentence. Barcelona has the dinner table. LA and London have the list.

What this means for anyone planning a trip in 2026

If you are flying to Barcelona this year, the city food creators comparison gives you a workable shortcut. Pull five food accounts that cover the city in Catalan or Spanish, not the English-language travel accounts, and watch what they cite in common. The intersection of those five lists is, in my experience, around twelve to fifteen restaurants. Eat at any six of those twelve and you will have a better trip than if you tried to follow a single LA-style mega-influencer with two million followers and zero accountability.

If you are flying to LA, you have to do the opposite. You have to pick a neighborhood first and then a creator who actually lives in it, and you have to ignore the citywide accounts almost entirely. The citywide LA food account is a marketing format, not a recommendation format. The same is true in London. The Time Out style list is fine if you want a survey of what exists. It is not a substitute for a network that has eaten at the places it is recommending.

I will say one more thing about the dataset we are working from. Barcelona's tourism economy is dominated by a small number of anchor sites, and the Basílica de la Sagrada Família, with roughly 165,331 reviews and a 4.7 average, is the gravitational center. Food creators in Barcelona orbit those anchors more tightly than LA or London creators orbit anything, because the city simply does not sprawl in the same way. That spatial constraint is part of the network density story. When the city is walkable, the creators eat at each other's recommendations. When it is not, they do not.

The practical conclusion: if you are using TikTok to plan a food trip in 2026, you should bias toward midsize cities, and you should be skeptical of any creator scene in a metro larger than roughly eight million people. The recommendations get noisier as the city gets bigger, and there is no algorithmic fix for that.

This is the part where the GeoTok pitch makes itself. We built GeoTok because the manual version of this work, which is exactly what I have been doing for the last five days, does not scale. If you want to see the creator graph for a city you are visiting rendered visually, with the restaurants that multiple accounts converge on actually flagged, that is what we do. We do not invent the network. We just make it legible.

The honest summary in May 2026 is that the creator economy keeps getting larger, and that is a problem for recommendation quality, not a feature. Barcelona's scene is the proof that smaller can be sharper, and that the midsize city advantage is doing more work than anyone in LA or London wants to admit. I do not think that gap closes. I think it widens, and I think the cities that punch above their weight in the next two years are going to look a lot more like Barcelona than like the megacities everyone else is still trying to cover. The food press will catch up to this in 2027 or 2028, write a few think pieces about Lisbon and Copenhagen, and pretend they figured it out first. They did not. The creators in those cities figured it out, by accident, because the network forced them to. That is the whole argument, and that is what LA and London have to solve before they can produce a recommendation graph anyone can rely on. Eat well in May 2026.