Tabelog vs AskU
- FourSeasons Jul 16, 2008 04:11 AM
Does anyone know how Tabelog and Asku, the two major internet driven user websites in Japan, determine the rating and ranking of the restaurants?
I am just wondering the type of demographic that led to very different conclusion for Tokyo area by the two groups. Is it due to income group, gender, occupation, geographic or anything specific?
For example, in Sushi, Sushi Mizutani (ranked no. 2), Kanesaka (ranked no.9) in Tabelog is nowhere to be found in the top 30 in Asku. No.1 ranked Sushi Nakahisa in Asku is nowhere to be found in the top 20 list in Tabelog. There are only 2-3 common names in the top 20 of both list.
It is the same with 日本料理 category, no.1 ranked 京味 (in Tabelog) is not in the top 30 list of AskU while the no.3 ranked Kitcho (in AskU) is not in Tabelog at all. At a glance, only 3-4 names share the top 20 list in both sites.
I notice most "best" lists in major cities are usually quite similar, even if the rating companies are competitors. So I am interested to know why it is different in this case. Would appreciate any feedback on this issue.
I'm pretty sure they both take aggregated scores and generate an average. The major difference seems to be that ASKU simply sorts based on the value of the absolute average, while Tabelog takes into consideration sample size. For example, if five of us go to eat at Sushi Hirai, a new sushi shop in Ebisu, and rank it on both sites the highest possible score, it will immediately register at the top on ASKU. It will not on Tabelog. When I have posted best of Tabelog/ASKU stuff in the past here on CH, I actually filtered out ASKU lists where shops had a very small number of reviews. At this moment, a sushi place called Nogizaka Sushi Yuki is rated third for Tokyo sushi shops, with only 6 reviews. So I would have eliminated that from my comparison list as all the surrounding ones have 20+ reviews.
Anecdotally, I've noticed that the ASKU reviews are not as recent as TL. The other one just has a fresher audience. Beyond that, TL just has so many great features- i.e. being able to drill down to a single page of "best of" lists for various cuisines for a specific part of town. It's also got good mapping features and photo galleries. It's just a better resource. Regarding demographics, I have no idea. However, I do notice the same reviewers coming up all the time on TL. I can't tell if that is because those individuals are so prolific in their coverage or if they have some special membership status. Further, a problem I have with TL is that many of the reviews will cover lunch service, which as we all know, can be dramatically different than dinner- especially at Japanese restaurants. Scores will reflect the same on the aggregate of course, regardless of what time of day the review is covering. You have to watch out and read closely for the lunch thing. In the end, you have to scrutinize the reviews to find out why places are ranked the way they are.
I've found them both to be fairly random in terms of actual rankings, so the variations are based on randomness as much as anything. The ranks also tend to be self-perpetuating - i.e. if a particular place scores highly on Tabelog, more Tabelog users will go there to try it out and write even more reviews about it.
One problem is that there are so many restaurants in Tokyo, and so many good ones, that it's impossible for anyone to try them all to compare them. So people only go to places they've heard good things about.
Anyway, whenever someone mentions that a restaurant scored highly on Tabelog I like to point them to the site's Mexican page, which tells us that El Torito is the #2 Mexican restaurant in all of Tokyo.
re: Robb S
Well presumably, Japanese reviewers will be much more apt judges of say yakitori or sushi, than Mexican cuisine. But you're right, the rankings are indeed self-perpetuating as many reviewers are now referencing the rankings when they post their reviews. I do this of course as well, but I'm not scoring the restaurant. Though I always want to be clear where I've found out about places. I don't think anyone is under any impression that these rankings are absolute or all that meaningful. But the sites themselves are great resources, especially when you drill down to the neighborhood level. And aside from random pop-in to try new places on one's own, they are a way to expand to for the most part good restaurants beyond what is listed on sites such as your's. In the end, they are doing a good job of presenting the scope of what's out there.
I certainly agree that they're good resources, and I've discovered great places through both Tabelog and AskU - places I probably wouldn't have found otherwise.
And of course if someone goes to a highly ranked place and they turn out to be good, it is interesting to hear that it's highly ranked on Tabelog. But if the only thing I know about a restaurant is their Tabelog ranking, I'll wait for more data.
Great points Silverjay and Robb S. It's very similar in the U.S. with popular sites like Yelp and Citysearch. People see High Scores on Yelp or Citysearch and end up visiting those restaurants first, etc.
I think that's what makes the Japan Chowhound board so useful for me: Everyone's insight on places that may not be listed anywhere, or that one might've missed while casually perusing "Top 10" lists on various sites.
It's great insight from great people that led me to small (and great places) like Toki no Ma (thanks Robb S! :) and Rokurinsha (Silverjay)! (^_^)/
I live in Kyoto and use tabelog.com frequently to make decisions on where to eat. It's a site that ranks restaurants from reviewers' ratings. I also post comments on the site for restaurants I've visited.
It's a good guide for Kyoto. I'm familiar with the city and the restaurants recommended there. I'm not sure if it's the same for other cities in Japan, though.
The site has many avid reviewers. It's a good way to choose your favored reviewers by seeing if you agree with their ratings and comments. I count more on my preferred reviewers' than the overall ranking since, as someone else pointed out, that the number of reveiws don't seem to get factored in as much as the actual rating of one review.
Thanks for all the reply. But I think some miss my point. I am not debating whether the sites are useful (which I think most agree they are) or the ranking is correct (after all, every individual has different taste bud). But it just baffled me the difference the rankings are in those 2 sites, especially considered the stereotype image that Japanese tend to be more herd instinct and to make decisions based on consensus.
This is in stark contrast to my recent "research" on Chowhound, Zagat, Michelin, NYT prior to my recent trip to NYC. The views there converge with each other in supposedly a much more individualistic society: Per Se on fine dining, Sushi Yasuda on sushi, Babbos on Italian will rank no.1 and high up in the rankings of every survey. And in the top 10 of all the surveys, you will find pretty much the same 5-8 names, though may differ in exact ranking.
Even after eliminating the number of reviewers like what Silverjay suggested, there is still a gap between the preference of the users of the 2 sites. For example, when I look at the top 20 of Sushi categories, only 5 names are in the both top 20 list. (For Michelin fans, only 2 out of the 5 are from Michelin, namely Sushi Saitou and Kyubei). I think even eliminating the low number of reviewers do not produce much difference. Even more amazing, only one of the top 10 in Tabelog appears in top 10 list of AskU, and likewise in reverse role. It just seem to me to be two different groups out there trying to make a statement that "we are different and unique".
Or could Robb be right that there are so many good restaurants out there in Tokyo that the difference between the "very best" and "good" is such a small difference that it is not detectable, unlike the case in Manhattan and many other cities where there is "concentration of power" in the hands of few? But I do doubt if it is due to "self-perpetuating" like what Robb suggested, I am sure all these "best" list would attract many attentions, I am sure Tabelog fans would also view the list at Asku and flock to those favorites as well. In the end, they usually converge, but this is not the case here.
For Michelin fans out there, you will be surprised not many make into the top 20 list. Out of the 15 Sushi places with star ranking in Michelin, only 6 made it to top 20 in Tabelog, and 4 in AskU. And incredibly, the legendary Michelin 3 stars Jiro did not make it into both list. In fact, AskU users basically ignore the 5 Michelin three and two stars places. I wonder if that means AskU users are more anti-establishment? (since 3 out out of that 5 make it into Tabelog)
Hi FourSeasons, thanks for the observation. I find this topic very interesting.
I notice this too when I look at the rankings in these Japanese user-driven sites. Will it be there are too many restaurants in Tokyo ? I don't have the official numbers, but it is like there are 160000 restaurants in Tokyo compared to 20000 in NYC, and as other mentioned, most of them are with good quality and not easy to really tell the differences. And especially in Sushi, some chef will treat you better if you have established a "good relationship" with them, so it is more like personal preference in some way when one writes a review on those sites ? I am so surprise to find even the main Jiro on Ginza is nowhere to be seen in these sites. But personally, I found these sites very useful, reliable and informative, great as a reference.
The tiny Hong Kong also has a big number of restaurants, with tons of hole in the wall, hidden gems.... There are similiar situation in Hong Kong's Openrice user-driven site, maybe not to that extent, but I find the world famous abalone specialist "Ah Yat" is getting a pretty low score in there too.
Great topic. :) It's definitely interesting, but I think it just might be the sheer # of restaurants and some possible self-perpetuation on each. Another baffling example of disparity:
* Ramen Rankings in Tokyo: I've seen quite a few "Best Ramen Rankings" and the Top 10 differ drastically between them (this is beyond AskU and Tabelog, incorporating some TV Specials and "Fan Polls"). In one Top 10 Ramen Ranking (I don't have the Video Link anymore), NONE of the Top 10 from Tabelog or AskU made it into that list. Shrug.
Ultimately, reports from Chowhounders here and Blogs that we feel in tune with really help gain some insight as well, IMHO (^_~).
My impression is that Michelin is not widely accepted in Japan (or most Japanese diners couldn't care less about it). None of my foodie friends in Japan mentions it and some don't even know its existence.
When the Tokyo version came out, it was controversial and some said, "How the heck such and such a restaurant made it?" Jiro is one of them and you see many negative comments on (and off) Tabelog and AskU.
Some were offended by the guide.
IMO, Michelin is essentially for tourists. It's expected to serve to bring in more overseas tourists and it looks like it's doing its job (juding from this board).
I guess that's why the Japanese version of the guide sold out multiple printings. As a point of reference, at Ristorante Honda, which is basically next door to my apartment, it was fairly easy to get reservations on a regular basis. Since the guide came out, it has become quite difficult to get them less than two weeks in advance. And yes, there are now more gaijin eating there than before (before the guide, I only saw 1 other gaijin party eating there in 3 years of patronizing the place), but the vast majority of the diners remain Japanese. I guess they all read it in Hanako or Kateigaho.