If you've cross-referenced Tabelog while planning a trip to Japan, you've probably had this moment: a restaurant sits at "only" 3.5, so you scroll past it looking for something closer to a Yelp-style 4.5, assuming 3.5 means it's fine but nothing special. That instinct makes sense coming from an American review site. On Tabelog, it's almost exactly backwards.

At a glance
What a 3.5 actually meansRoughly the top 3-4% of restaurants nationwide
What 4.0+ actually meansAbout 413 restaurants out of 905,000 nationwide (0.05%)
How the score is calculatedA weighted average based on reviewer reliability, not a simple average — the exact formula isn't public
English-language accessTabelog's own app, in English, since November 17, 2025
Figures confirmedJuly 19, 2026, from Tabelog's official help page and independent research

The Tabelog Scale, By the Numbers

Tabelog is Japan's dominant restaurant review platform, and its scores look deceptively familiar to an American eye: a five-point scale, one decimal place, the same shape as Yelp or Google Reviews. The numbers behave completely differently, though.

According to Tabelog's own official help page, a score of 4.00 or higher places a restaurant in roughly the top 500 nationwide, and a score between 3.50 and 3.99 puts it in roughly the top 3% of every restaurant on the platform. An independent analysis published on note.com went further, counting all 905,000 restaurants listed on Tabelog nationwide over a two-week window in April 2020. The result: only 413 restaurants in the entire country scored 4.0 or above — about 0.05% of everything listed — while restaurants scoring 3.50 or higher made up roughly 3.3% to 4.0% of the total, closely matching Tabelog's own "top 3%" guidance.

Put plainly: on Tabelog, 3.5 is not a "pretty good" score. It's a "this is one of the better restaurants in the entire country" score. And 4.0 isn't rare in the way a great Yelp restaurant is rare — it's rare in the way a restaurant with a line out the door every night of the year is rare.

Tabelog's own published guidance, checked against an independent 2020 study of all 905,000 restaurants then listed nationwide.

Tabelog's official guidanceIndependent study (2020all restaurants)
Score of 4.00 or higherRoughly the top 500 nationwide413 restaurants nationwide (about 0.05%)
Score of 3.50-3.99Roughly the top 3%About 3.3%-4.0% nationwide

For a US traveler used to Yelp or Google, where 4.0 is a completely ordinary score for a decent neighborhood spot, this takes some recalibrating. A rough (and imperfect) way to think about it: a Tabelog 3.5 is closer to what a 4.5 or higher would mean on Yelp, and a Tabelog 4.0 has no real Yelp equivalent at all — nothing on Yelp is capped at 413 listings nationwide.

Why Tabelog Scores Are So Much Stingier Than Yelp or Google

The gap isn't just cultural pickiness — it comes down to how the number is calculated. Tabelog's help page is explicit that the score is not a simple average of every star rating submitted. Instead, each reviewer is assigned an "influence" weighting based on their review history and reliability. Higher-influence reviewers count for more than brand-new or less-established ones, and the whole score is recalculated on this weighted basis twice a month, on the first and third Tuesday.

Tabelog doesn't publish the exact formula behind that weighting. According to its own help documentation, that's a deliberate choice, intended to make the score harder to manipulate through fake reviews, review-swapping, or coordinated rating campaigns — a real problem on review platforms everywhere. The tradeoff is that outsiders can't fully audit how any single score was produced, and scores tend to move more slowly and settle lower than they would under a simple star average.

An opaque, undisclosed formula asks you to trust the platform's judgment rather than verify it yourself — one of the system's honest weak points. Tabelog's answer is that the alternative — a simple average anyone can game with enough fake five-star reviews — is worse. Whether that tradeoff sits right with you, it also means a restaurant's score can shift for reasons that have nothing to do with the food itself — including changes to the weighting algorithm (more on that below).

How Japanese Diners Actually Read a Score

Because the scale runs so much tighter than Yelp's, Japanese diners don't read a Tabelog page the way an American reads a Yelp page. A few habits show up again and again.

The 3.5 Line

Given how few restaurants clear 3.5, it functions as an informal quality line for a lot of Japanese diners: below it, a restaurant is probably fine but unremarkable; at or above it, a restaurant is worth a special trip. It's not an official cutoff Tabelog publishes, but the math above explains why it works as a rough shortcut — 3.5 already means "better than roughly 96 to 97 out of every 100 restaurants nationwide."

Review Count Matters as Much as the Number

A 3.6 built on 400 reviews reads very differently from a 3.6 built on 12. A score backed by a large volume of reviews has effectively been stress-tested by far more people, and by the platform's own influence-weighting logic — a high score on a thin review count is treated with more caution.

Photos Are a Signal, Too

Beyond the number itself, Japanese diners routinely scroll through a restaurant's photo section before booking, since photos come from reviewers rather than the restaurant's own marketing team. A restaurant with a large, recent, unstaged-looking photo library backs up its score in a way the number alone can't.

Tabelog Finally Speaks English

For years, the biggest practical obstacle for a US traveler wasn't understanding the scoring system — it was that Tabelog itself was Japanese-only. That changed on November 17, 2025, when Tabelog launched an official multilingual app with English, Traditional Chinese, and Korean support, built for inbound travelers, with booking included directly in the app.

The uptake has been fast. According to a Kakaku.com press release dated May 7, 2026, the app passed 2 million cumulative downloads within its first half-year on the market, and ranked #1 among inbound restaurant-search apps in Taiwan, Hong Kong, South Korea, and the United States. As of that same release, more than 70,000 restaurants nationwide support instant online booking through the app.

Booking through the app isn't entirely fee-free: it requires registering a credit card, largely to cut down on no-shows, plus a ¥440 (about $2.90) system-usage fee charged the moment a booking is confirmed. Check a restaurant's own cancellation terms too, since some apply their own fees on top of Tabelog's policy.

What this means practically: you no longer need a Japanese-reading friend, a hotel concierge, or a translation app screenshot to use the same restaurant-discovery tool that Japanese diners have relied on for years. You do still need to read the score on Tabelog's terms, not Yelp's.

Where the Score Falls Short

None of this means the score is beyond question. Tabelog's scoring has faced real challenges worth knowing before you treat any single number as gospel.

Starting in 2019, Tabelog applied a scoring algorithm that specifically lowered ratings for chain restaurants relative to independent ones. The operator of the Kollabo yakiniku chain sued over this, arguing it was unfair.

In June 2022, the Tokyo District Court sided with the plaintiff, ordering Tabelog's parent company, Kakaku.com, to pay 38.4 million yen in damages, and ruling the practice an abuse of superior bargaining position. Kakaku.com appealed, and in April 2024 the Tokyo High Court reversed that decision, ruling that adjusting the algorithm to address a gap between raw scores and consumer perception, and to counter increasingly sophisticated fake reviews, was reasonable.

Whichever ruling you find more persuasive, the case is a useful reminder: a Tabelog score reflects a company's editorial and algorithmic choices, not a purely neutral tally of public opinion.

The score also isn't the whole picture at the very top end of Japanese dining. Some of the country's most sought-after restaurants are so small, and so booked out by returning regulars, that they barely register on any public review platform at all — Tabelog score included. A high score is a strong positive signal; a middling or missing one on a genuinely tiny, reservation-only restaurant isn't necessarily a red flag — it can just mean the restaurant doesn't need the exposure.

The practical takeaway for a US traveler: use the Tabelog score as your first filter, the way you'd use a Yelp score at home, but recalibrate what the numbers mean before you do. A 3.5 is worth booking. A 4.0 is worth planning your trip around.