I had 47 saved TikToks before I bought the flight. By the time the trip ended, I'd visited 14 of them, two of them twice, and the best meal of the week was at a place I'd never seen on any creator's feed.
This is what actually happened — what worked, what got cut, and what I'd do differently. If you're sitting on a save folder of travel videos and trying to figure out how to turn them into a real trip, this is the one.
Phase 1: The folder
The 47 videos were a mess. Some were ramen, some were skyscrapers, some were three-second pans across a vintage shop in Shimokitazawa with no caption. The first thing that became obvious is that most travel videos don't tell you where they were shot. The creators tag it "Tokyo" if they tag it at all. A video about "the best yakitori in Tokyo" might be a 12-minute walk from your hotel or a 75-minute train ride, and you can't tell from the video.
My first attempt was a Google Sheet. I tried to manually pull every place name out of every video. By video 8, I gave up. By video 12, I'd lost the link to one of the originals. The Sheet was unfinished, and I still had 35 videos to go.
The actual fix — the one that made the trip happen — was to share each video into an app that extracts the places automatically. (Yes, this is the SavedTrips pitch, and yes, that's why I'm writing this post.) After 25 minutes of one-tap shares, I had a structured list: 142 places extracted across 47 videos, each tagged with a category and a neighborhood, and almost all of them with a Google Maps link.
Phase 2: The first cut
142 places is too many. The trick is the first cut. Here's what I removed:
- Anything I couldn't actually walk to. A few places were 60+ minutes outside the city. Dropped them unless they were headline experiences.
- Anything I had three of. Five "best ramen in Tokyo" videos pointed at five different shops. I picked the two most-saved and let the rest go.
- Tips, not places. Some videos were more "wear the lightest layers in summer" than they were specific destinations. Useful, but not on the map.
- The obvious tourist traps. This is personal, but if a creator has 4 million followers and 12 spots in the same Don Quijote, that's not the trip I'm planning.
That cut got me from 142 to 58. Still too many. The second cut was harder: what's the actual trip? I knew I had seven full days, two of which I wanted to keep mostly empty. So I had five "active" days and an arrival/departure buffer. At three to four places per day, that's 15-20 places, with margin.
I ended up with 24 places saved as "definitely doing." Another 16 marked "if we have the energy." The rest archived.
Phase 3: The voting
I was traveling with three friends. Up to this point, I'd been planning solo. The next mistake would have been to send them a Google Doc and watch nothing happen.
Instead, I shared the list with them in the app and we did a round of voting. Three buttons per place: must-do, want-to, if-time-allows. Took everyone about ten minutes to get through it.
The voting unlocked something I didn't expect. People agree more than they realize about food and disagree more than they realize about everything else. All four of us voted "must-do" on the same three ramen counters. But on the museums, the hike to Mount Takao, and the day-trip to Nikko, the votes were all over the place.
That's actually the planning shortcut: the things everyone agrees on are the spine of the trip. The things only one person wants become solo afternoons.
Phase 4: The geographic clustering
This was the boring part that nobody on TikTok shows you, and the part that makes or breaks the trip. The most-saved 24 places were scattered across nine different neighborhoods. Without clustering, we'd have spent 90 minutes a day on trains.
The fix: pick a neighborhood per morning, a neighborhood per afternoon, and stitch dinner in between. Shibuya morning + Harajuku/Omotesando afternoon, with dinner in Shimokitazawa. Yanaka morning + Ueno afternoon, with dinner back near the hotel. Tsukiji breakfast + Ginza walk + Akihabara afternoon, with a no-plan evening.
The app I was using did this automatically — geo-clustered the saved places into walkable days based on the trip length. But it's not magic; you can do it yourself with a paper map and 20 minutes. The point is to do it at all.
What I'd do differently
Three things stand out, looking back:
- I undervalued the buffer days. The two unstructured days were the best two days. We wandered into a stationery shop in Yanaka that had a 90-year-old shopkeeper showing us the difference between three almost-identical pens. Nobody had filmed it. Nobody will ever film it. It was the highlight of the trip.
- I over-trusted the food videos. Some places are filmed for the food. Some places are filmed because they have a neon sign and the food is fine. The ratio is closer to 50/50 than I assumed. Reading the reviews after extracting helped a lot.
- The whole-group voting saved a probable group-chat fight. Without it, I'd have made a unilateral itinerary, sent it to the group, and had to absorb a slow-motion negotiation over text for the next two weeks. The voting compressed two weeks of group-chat resentment into ten minutes of taps.
The real lesson
The save folder is not the trip. The save folder is the raw material. The trip is what happens after you cut the list, vote on it, cluster it geographically, and accept that two of the days won't go to plan.
If you're sitting on a Tokyo save folder right now, my advice is: stop trying to be your own travel agent in a Google Sheet. Use the videos as a starting point and let an app do the structural work. Then do the part it can't — the voting, the cutting, the slow afternoons.
The real trip is in the editing.
— —
If you want to skip straight to the part where the videos turn into a structured list, SavedTrips does that. Or read our full Tokyo travel guide for our curated picks across the city.