The algorithm killed the guidebook — and the guidebook killed itself

The algorithm killed the guidebook — and the guidebook killed itself

In May 2026, I bought a brand-new copy of a famous travel guidebook for Lisbon, flipped to the restaurants section, and found a place I had personally watched close in 2019. Seven years. Seven years of editions, seven years of paid editors, and the listing was still there, complete with the chef's name, the price range, and a little symbol indicating "good for vegetarians." The chef had moved to Madrid. The space was a phone repair shop. I am told this is a coincidence. It is not a coincidence. It is the genre.

The travel guidebook didn't lose to TikTok. It lost to its own refusal to admit that a recommendation older than its readership is no recommendation at all. We can keep telling ourselves the comfortable story — that an unfeeling algorithm crushed a thoughtful old craft — but the craft stopped being thoughtful around the time editors started getting their dinners from PR firms. The future of travel guides was decided long before the first food creator filmed a bowl of soup.

Lonely Planet sold for around $50 million in 2020 after Red Ventures bought it from BBC Worldwide, which had paid roughly $210 million for it in 2007. That is not a typo. Fodor's, Frommer's, and Rough Guides have all shed editorial staff in the years since, and the remaining guidebooks now lean heavily on aggregated content from sites like TripAdvisor, which is itself a layer over user reviews of variable quality. So when people ask "is the travel guidebook dead," they're asking the wrong question. The honest question is: when did it stop being alive?

What the guidebook used to be, and what it became

I owned a 1998 Lonely Planet Indonesia. The author rode buses for months. He listed the warung that the bus drivers ate at, the woman who rented snorkels on the beach in Pulau Weh, the bemo route that nobody else knew about. The book was wrong about prices. It was sometimes wrong about hours. But it was right about the world, because the person writing it had actually been there in the world, that year, with sandals on.

By the late 2010s, the same publisher was issuing updated editions where roughly half the listings were either the same as five years earlier or pulled from press materials. I know this because I cross-referenced two editions of a city guide I won't name and found that the "new" edition had moved 3 listings and rewritten 12 captions. Everything else was a refresh of the publication date.

When a Lonely Planet recommendation persists for seven years without anyone visiting it, that's not a recommendation. It's archaeology. And the reader can feel it, even if they can't name it. There's a stiffness in the prose that says: I have not been here. I have been told about here. The footnote-bound 19th-century travel writer Karl Baedeker, whose name became the original brand of the genre, would have laughed.

Meanwhile, the cost of a person actually getting on a plane, eating their way through 30 restaurants, and writing 4,000 words of honest copy did not fall. It rose. The advertising rates supporting that work, however, collapsed. So the math of the guidebook — pay one writer to spend two months in Hanoi — stopped working before TikTok was even a verb.

The takeaway: the guidebook didn't die of new technology. It died of an old margin problem the publisher hid by recycling.

Why the algorithm wins, and where the algorithm cheats

Now to the other side, because I don't want to romanticize what replaced it. TikTok's food and travel side has a real signal that the guidebook lost. When 27 different creators show up at the same bakery in Lisbon's Graça neighborhood within six months and film the same custard tart, something is true about that bakery that no editor in London is in a position to verify. The signal is: 27 different humans, with 27 different cameras, in 27 different moods, decided this place was worth filming. Try faking that. You'd need a budget bigger than the bakery's annual revenue.

The platform has its problems, and I don't think you should swap the guidebook for the For You page either. The For You feed will gladly serve you a place that paid a creator $400 for a 15-second clip, and the visible signal — the post — looks identical to a place that earned 27 unpaid visits. The algorithm doesn't tell you which is which. It just shows you both and counts likes.

What TikTok actually does better than the guidebook is this: it shows you what the place looks like, right now, in May 2026, in 4K, with the actual owner saying the actual words. The genre of the static recommendation — "we recommend this place" — is replaced with the genre of the present-tense moment. A guidebook describes. A 30-second video shows. That difference is not aesthetic. It's epistemic.

I will give you a concrete example. The creator @keith_eats has, across his run on the Eater "We Tried" video format, demonstrated something the guidebook fundamentally can't: he tells you whether the food is actually good while he's eating it, and you watch his face. Now, you can argue about whether that's journalism, and I'd argue back that it's closer to journalism than a paragraph written from a press kit. The pull-quote from his work that has stuck with me is the throwaway moment where he says, of a dish:

"I don't love it. I don't love it at all."

A guidebook will never write that. A guidebook is contractually obligated to love every recommendation. That is, in fact, the central problem of the genre.

But here is where I think the conversation needs to grow up. TikTok travel recommendations are not, by themselves, a replacement for what the guidebook used to be at its best. They are a stream of moments without a structure. You can spend two hours scrolling and end up with 14 saved bakeries in 9 cities and no plan. The platform makes the choice; it doesn't help you decide. The algorithm is generous with options and cheap with judgment.

What's actually missing is the editor — and I mean the function, not the LinkedIn title. An editor's job was never to write the prose. It was to decide what belongs in the canon for a given city, defend that choice with reasons, and update the canon when reality changed. The guidebooks stopped doing this somewhere between 2008 and 2014. TikTok never started doing this. So the reader is left to do the editor's job themselves, which is why travel planning in 2026 feels exhausting even when you have access to more information than any traveler in human history.

The takeaway: the algorithm wins on freshness and humility. It loses on structure. Whichever product gets both wins the next decade.

What replaces the guidebook, and what the reader should actually do

Here is my position, plainly. The guidebook is not coming back. The lonely planet decline is permanent, and no amount of nostalgia for that yellow spine is going to reverse it. The team at Red Ventures isn't going to invest in 800 freelancers riding buses in Bolivia. The economics don't work and they're not going to start working.

What replaces the guidebook is some combination of three things: a verified stream of real visits (TikTok and its successors), a structural layer that organizes those visits into something a planner can use, and a privacy layer that doesn't dump the traveler's coordinates into a marketing pipeline. That third piece is the one I see almost no one talking about. The guidebook had a quiet feature that everyone took for granted: when you walked into a restaurant from a Lonely Planet recommendation, the restaurant didn't know it. Your data didn't follow you in. Most modern recommendation systems can't say the same.

I built GeoTok because the structural layer was missing. In May 2026, when I open the app, I see a curated, current map of what people are actually filming this week in the neighborhood I'm in, without anyone tracking which place I tap. The point isn't to replace TikTok. The point is to put a structure around it that the platform itself has no incentive to build.

For the reader, the practical change is this. Stop trusting any single source. The seven-year-old listing in a famous guidebook is not your fault, but believing it is. Watch three or four current videos from creators who aren't using the same opening sentence. Look at the date stamp. Look at whether the owner appears on camera. Look at whether the comments include locals correcting the creator's pronunciation — that's a tell that the place is real and beloved, not a tourist set piece. And keep a small list of places you want to remember, in a tool that doesn't sell your list back to a hotel chain.

The genre of the travel guidebook isn't being killed by an algorithm. It's being replaced because it stopped doing the job it was paid to do, and a generation of writers and editors let the rot happen quietly. The next version of travel writing — whatever we call it in 2027, 2028, 2030 — will have to earn the reader's trust the way Baedeker earned it: by being right about the world, this year, this month. As of May 2026, the candidate that's closest to doing that is a 30-second video by someone you've never heard of, filmed yesterday, with the price written in chalk behind them. That's not a romantic answer. It's just the honest one.

— Aleks, GeoTok, May 2026