How best to deal with bilingual sites

[CarlosC]

The websites I am making are in Japanese, and I've been asked to wear the "SEO expert hat" in my company. For the most part, all the visitors that come to the sites we design are Japanese, thus we want to be SEO for google.co.jp, yahoo.co.jp and so on. So if I can, I'd like to ask for advice from other bilingual or multilingual site designers...

1) Let's suppose that the site you're making doesn't use English at the very top level. For example, you are asked to design sony.co.jp. You will want to use keywords and a desription for your core visitors, but, should you also through a few English keywords in there as well? I'm a tad confused about this.

Using a non-English language for things like ALT tags is no problem, but how about LINKS and SITE directories?

For example, in English we might have a page called www.domain.com/aboutus.html. In Japanese, we have 3 writting systems (kanji, hiragana, katakana), and none of course are usable for link names. Most often, Japanese designers will use the "romaji" spelling for a link.

Like, www.domain.co.jp/shokai.htm (shokai == introduction, kind of like About Us)

Many of the young designers are asking me, should we use romaji spelling? English? Does it matter? Obviously when a Japanese searches for something in an Search Engine, they will use one of the three writting systems, and not romaji (well, maybe 95% of the time).

As for site structure, the same questions come to mind.
In English we might have a site structure like..

www.domain.com/fruits/apples/

So, should we use romaji such as...

www.domain.co.jp/kudamono/ringo/

???

Finally, who are people naming files, such as images? apple.gif or ringo.gif?

Well, I'm new to SEO (go easy on me!) so I'd like to heard as much as possible from anyone working on non-English sites.

Thanks!

CarlosC

[huebdoo]

Double byte characters and SEO have been a mystery to myself. I have been lucky that the products that I marketed in Japan had some brand recognition and we also did quite a few OEM deals with Japanese companies so marketing was up to them and not us.

As for naming... do it as you would normally, you might also want to check to see exactly what search engine the Japanese are using presently (they used to have quite a loyality to Yahoo.co.jp) but that could have changed with Google

As for your links and site directories / insure that everyone knows what is going on and design and development are all on the same page for any cluster fu*** that happen on the backend or during development will cancel any gains in SEO. (Time is Money)

Decide one style of Japanese language and follow it (for example we went with Simplified in China rather than Traditional) - Once again its all about insuring everyone is on the same page and you dont get any confusion.

Images ... well how many searches are there for a ringo? spelled in english like this versus Apple? - that will answer that

Hope that helps - Good luck and I really hope your PM is up to the task

[CarlosC]

Thanks for your input. Hopefully other people who work on bilingual sites will jump in.

Cheers,

Carlos

[Chris_D]

I'm not sure what the question is?

Is the site content in kanji, hiragana, or katakana? Or all?

Image file names as such, are less important - it doesn't really matter what you call a .gif in the total scheme of things - look at the number of sites/ CMS that use numerals for image names.....

File names and paths - again - don't have to reflect search terms. Look at the URL of this page here at SEW. I'd probably use romanji if thats an accepted compromise.

The core issue is to talk to your audience in their language.

[bill]

You will want to use keywords and a description for your core visitors, but, should you also through a few English keywords in there as well? I'm a tad confused about this.Don't bother with English keywords/description for the Japanese market. Use Japanese.

Using a non-English language for things like ALT tags is no problem, but how about LINKS and SITE directories?You can stick with the English folder and file names. It makes no difference in Japanese SEO. If you want to make the links a bit more understandable for your more savvy visitors then go with the romanji folder and file names, but this will do nothing for SE ranking.

[CarlosC]

>Is the site content in kanji, hiragana, or katakana? Or all?
Japanese is a mixture of all.

>Don't bother with English keywords/description for the Japanese market. Use >Japanese.
Copy that.

>If you want to make the links a bit more understandable for your more savvy >visitors then go with the romanji folder and file names, but this will do nothing >for SE ranking.
Thanks for the advice, I'll follow that.

Cheers,

Carlos

[Olney]

Using English file names may be safe currently but Google & Yahoo can actually list Japanese URL (not in romaji)

I used to think native Japanese URLs were bad but they are actually being indexed & displayed correctly. It also depends on the way the crawlers find your site (what is the URL of the backlink etc.)

[CarlosC]

Yep, but I thought the idea of domain names in 2-byte characters was an idea that never took off. Like, even sites like Excite don't bother with http://www.ƒGƒLƒTƒCƒg.co.jp.

Cheers,

Carlos

[Olney]

There's two different things here.
Native Japanese Domains = IDN Domains like アニメ.com

But I was talking about Japanese URLs
www.widgets.com/アニメ/ or even www.widgets.com/アニメ.html

The IDN domains will finally take off after this new Window update Vista & the new IE7. This will allow Japanese domains to finally have a bit of merit in terms of SERPs as of now & always English based domain names have absolutely no value to the Japanese SEO market.

As far as the URLs go in English browsers the name might look like www.widgets.com/6t%4%H% etc. but if you link Correctly google & yahoo are already able to read this correctly in the SERP. I still need to do a bit of testing in the Japanese versions of Google & Yahoo to give data.

I'm providing much data already for the past few months on the IDN market (mainly Japanese & Chinese) as finally they will begin to gain value in terms of marketing & especially SEO merit.

IDN domains are also already showing up in search results but since people couldn't use them because of IE they were put on the backburner.

Changing completely to an IDN domain or Native Japanese URLs is a bit premature. I actual work for an SEO company in Tokyo & am privately researching myself how far the big guys have already gone with these native domains & Japanese URLs.

If you need a local Japanese SEO company you can PM me.

[CarlosC]

Very interesting about the Kanji file paths. I didn't know that was coming.

I'm wondering the implications of uploading a file called ホーム.html to the server vs home.html

Cheers,

[bill]

Native language path names and files names have been used for several years on DMOZ. Check out their Japanese language version to see for yourself:
http://www.dmoz.com/World/Japanese/

When this first came out and when IDN names were first taking off back before Microsoft dropped RealNames support from IE (remember them?), I did some informal testing with Japanese business people of varying computer skill levels. Although some did agree that the URLs and path names were easier to understand the majority preferred the romaji versions for input. They complained about the extra keystrokes required to convert the path to Japanese.

By default the Japanese version of IE forces you into single-byte mode when typing in a URL in the Address bar. From a usability standpoint that makes current versions a real problem for any sort of IDN or native file/folder names.

There are a lot of places where developers have forced single-byte input of information (e-mail addresses, URLs), so IDNs will also mess up tons of website information forms that force the same type of behaviour.

Although IDNs are a great idea, and I've been following them since they were first introduced, I think that they have been too late in coming. A lot of development has been done on the web using the existing standards and take-up of these sorts of names will probably take quite some time.

[Olney]

I understand your point especially when IDN Domains were introduced about 5 years ago. I also was heavily involved in what was going on in bit valley here in Tokyo. SEO has actually changed worldwide since they were introduced.

5 to 6 years ago I had to coordinate with Japanese OLs for major projects because company managers didn't understand the net.

Now it's 2005 the new IE7 WILL be IDN compatible. It was a good idea but couldn't get any use initially because the most used browser in Japan didn't support it. This will changed things. A SEO company 5 years ago in Tokyo did So/So but now Japanese are coming up with the times.

As far as marketing & ease of use IDN Domains are better. Currently for a Japanese person to find a web site. You tell them the website name no matter what it is & they type it in, in Google or Yahoo Japan. Then it usually comes up. If the website names were just kanji they can actually start to remember domain names immediately.

Also currently English romaji domain names have absolutely 0% keyword value in SERPs. Having IDN domains change this. Google & Yahoo already started to display IDN domains in searches. The big boys know what they are gearing up for. It's just natural progression.

English is cool & all that but if you are working in an office with all Japanese that don't speak English & base what's easier to remember then it's IDN Domains all the way. I'm American & can type in Japanese & switch encodes in less than 1 second flat it's about as troublesome as I already have to press the spacebar with every Japanese word.

I also think that tools I'm using now to analyze the market perhaps may be different that when IDN domains were first introduced. Basically I'm saying I personally helped advanced the market recently & will continue to launch a few more web projects specially for IDN progression.

There are a few Japanese sites that the company promotes the web site name in kanji like kurumaerabi.com we have data that shows people are trying to type in the IDN domain in the browsers everyday consistantly. Or they are trying to search for the native Japanese term with the .com extension before trying to search for it in romaji

People doing the searches for kurumaerabi or the .com 179
People searching for the Japanese kanji with the dot com about 1,500 last month

Bill since you had interest before I'd love if you could contribute to our research, especially if you own an IDN domain portfolio. The general IDN use is not immediate but it will start next year finally. Too many major companies are factoring them in recently for you to take is as they are phased out.


Oh also you shouldn't use ホーム.html instead of a defauly index.html home is not something that you really need to optimize. I still need to do a bit of research on how many of these sites with Japanese folder names rank well or better, but it is factored in.

[bill]

It's good to see someone passionate about IDNs. After all these years the promises have jaded me a bit. However I've supported the idea all along. I hope they do finally get something working on a worldwide basis. It will only enhance local search, which is a good thing.

As far as the SEO value of IDNs I wonder how much of a factor it will really be. English keyword domains are great for type-in traffic, but SE algos don't really give that much weight to KWs in the URI path. The SEs know it's too easy to game the system this way so the weighting is decreasing.

Also there is an increasing trend being promoted to use local ccTLDs. Where do you think that will leave these VeriSign controlled .com, .net and .org IDNs? Countries like China have outright shunned that system and made their own. (just playing devil's advocate here)

We're still at the stage of a lot of what-ifs and conjecture to a large degree when it comes to IDNs. I do hope that the release of IE7 helps spur some interest in this type of domain. I work for a fairly old and large Japanese corporation and I see what the uptake on technology is. There are still plenty of Windows 95, 98, & NT machines in service. There will not be a speedy uptake of Windows Vista in the Japanese corporate world and only a limited uptake of IE7.

[CarlosC]

>just kanji they can actually start to remember domain names immediately.
Yes, but as you know there are many renderings for a kanji. If a company said on the radio, "Visit us at: ame dot co.jp", listerners might wonder 雨.co.jp　飴.co.jp　天.co.jp, so I'm not so sure that it will make the Internet magically easier for Japanese :D I think with the 6 years on bland English grammar/memorization they are forced into, most can utter the alphabet fine. ;)

I work for a fairly old and large Japanese corporation and I see what the uptake on technology is. There are still plenty of Windows 95, 98, & NT machines in service.
Indeed! I "worked" at one myself and many of the browsers were pretty darn old (machines running on 2000). I can't see them installing the next Windows on their thousands of Pentium 2/3 laptops. :D

Please continue this conversation... it isn't everyday I get the chance to speak to people involved with SEO & Japan. :D

Cheers,

[Olney]

Hey Carlos I think we went a bit off subject but I'll continue to explain.

The IDN market for Japan is not going to be made or successfully implemented by international Japanese. You know the ones that live in the big cities & can remember some English or the ones that vividly try to have foreign friends & live American, British, or Austrailian style. My personal research is not based on them at all. It's the ones who can't function in English a bit which is nearly all of Japan.

Some people argue that English is worldwide & everywhere in Japan but if you are bilingual (reading & writing) a copyline in English is fine but details is certainly not.

If I told a Japanese person my site is Japanese Celebrity site is geinojin.com they will not be able to write it in romaji. Romaji was made for us they never use it. To simplify writings they would just use hiragana. There also are various ways to write words that are said in Japanese in English. Japanese have never had a reason to practice writing Japanese in English. For example the term geinokai could be geinoukai or geinokai just like Tokyo is really Toukyou. But if they see it they can remember 芸能界.com or 東京.com

About the misspelling of kanji. This is as common as a person explaining a person's name. If your company is selling umbrellas & it's called Ame dot co .jp More than 95% of Japan will understand that it's the kanji for rain or the DJ would ask. Radio is hard for any language & a URL. I'm mainly speaking about visually being able to remember a domain name especially riding on the Yamanote in Tokyo.

kakaku dot com used to only have the Japanese kanji & dot com on their site. Do you know how many Japanese easily remembered the site but tried to type in the kanji & dot com? literally everyone. Most can't remember how to spell it in English though so the process fopr them to find the site if it isn't bookmarked is to just type in 価格.com into Google Japan or Yahoo Japan.

You can use the Overture Japan tool at my site & see how many searches are done for that 価格.com a month. Currently it's 1,014,998 for last month. If they own the IDN this number will go down trememdously in a year because of the new IE7. Not immediately but in a year or 2 if the new IE7 is initially released in February.

I'm American but I'm 100% basing all this on Japanese internet users now in Japan. The average person. Even being bilingual & going to school in Tokyo I'd still rather function in English. For major brands even Japanese people don't type in the brands in English letters in search engines. They type it in Katakana. Even stuff we know as being Japanese Nintendo, Playstation, Mitsubishi, Sony etc.

Most of this stuff isn't just available in English because it's just second nature to most Japanese.

On another note I was searching around & found some sites with Japanese in the URL may rank well in Google & Google Japan. Not IDN domains but the URL (directory name etc.)

I personally missed a few internet opts because of my young age but I do know Japan like the back of my hand. When I was younger I worked at a copy shop & exclusively used a Mac & started to do graphics. This kid once said in front of me because he just got a job at IBM & used to work at the same place. I'll never use a Mac, I just got to use a PC because a Mac insults my intellegence. a few years later we have windows. So does that mean he's stupid know (like he was back then)
My point is technology changes. IDN domains will benefit Asia greatly & you have to clearly think like the average Japanese to start seeing the benefits. If you personally think that kanji is hard or even that things just in hiragana etc. would just be better then you might miss the boat. (Not a rant guys just trying to give you knowledge)

[Olney]

Oh also the idea of these new Japanese domains will not be to change over company brands.
A lot of new companies started using .jp for new sites & kept co dot jp for company main site's etc.

I also work with a large media company that creates campaigns for large companies like the largest beverage company every knows.

They could have an online campaign at

ooloncha.com or ooroncha.com or uroncha.com
or simply
ウーロン茶.com

These are the main reasons for use:
1. Japanese cannot spell in English well at all
2. Simple easy to remember domains in English have been registered already anyway
3. Japanese cannot write Japanese words in English well. (I noticed this after asking 10 Japanese people that couldn't speak English to even spell their own name with the English alphabet) They have to think about it.
4. Visually they can remember the domains.

[CarlosC]

Romaji was made for us they never use it.
Actually, that isn't really true. While Japanese keyboards have hiragana, I know a great deal of people who enter Kanji/Japanese via the romaji method.

>kakaku dot com
Yeah, I think that is the only one that comes to mind.

The other info is interesting.

BTW, are all .jp registering companies over $80 a year? (OT again ;) )

[bill]

While Japanese keyboards have hiragana, I know a great deal of people who enter Kanji/Japanese via the romaji method.
I had to do a survey of several hundred PC users in Japan about their preferred input method and the romaji input method was used by the vast majority. That's only looking at use in one company, but I have a hard time finding people who use the hiragana input method. I also know that the PC classes at local schools from elementary on through college level teach input using romaji, so I'd say you're pretty safe using romanized phonetic names. Just spell them out the same way you'd type them. I do this the same way for Chinese names. Just use the the same pin-yin input you would use for the names on a Chinese system.

Aside from the IDN name idea which is still quite a few years off, another domain strategy that I've seen used a lot in Japan are the all number names. The Japanese use a phonetic shortening of numbers that you can creatively combine to make a memorable little ditty in Japanese. There are still plenty of these all numeric domains left and they offer an easily remembered name that can be used immediately.

[Olney]

Great discussion Carlos,
It's going to wake me up this morning.

There's a difference between keyboard entry & writing the structured Romaji we use to represent Japanese words.
Japanese mainly use keyboard entry via English letters. This is one thing.

But for a Japanese to sit down & try to read or write Japanese words in the Romaji system is a bit different.

For the keyboard system there are a lot of characters that we don't actually use in English like the small ぃWritten "xi" or the silentう in Tokyo
Romaji system is a bit simplified for us this was my meaning.
For example Central: 中央 pronunciation & Romaji Chuo
Written by Japanese keyboard: chuuou

This is one example & my point is you can't if there is chances of mispellings of a domain wouldn't be easier for a whole nation to use a domain name that wouldn't have the chance of being mispelled? Live the term Gravia Idol Could be grabia or glavia or guravia or 3 more variables & you will find all of them over the net in Romaji because they are all correct but in Japanese it only グラビア.

I think I'll put it simply like this. I own the term Gravia Idol in Japanese. Since the domain would be meant for Japanese there are 4 posssible high spellings for the common Japanese either: GlaviaIdol.com GrabiaIdol.com GlabiaIdol.com or GrabiaIdol.com
This term gets search for 25,000 times a day according to OVT jp (Overture Japan Tool not the US tool) or 773,991 searches for October.

Once IDN domains are abled to be used after the next Windows update for the Japanese market you can have one of four high possible ways to write the term (even though there really is about 16 different ways to write it in Romaji) or you can have the only one correct way to write in Japanese that any of those 25,000 searches a day would probably relate to.

The marketing aspect is like this you can have a high searched term in Native Japanese. I'm trying to create an amazon affiliate site there to test the traffic.


PM Me & I'll give you a good place to register Japanese dot jps I won't make this public.

I will say that the registries in Japan are giving huge extremely high discounts on registering native Japanese domain names since the new IE has been officially announced. Is this a coincidence? No I believe it's an organized plan. This stuff is not in English yet. Google & Yahoo have already been indexing IDN domains & they are appearing in search results.

It's your choice to invest in these or not but I plan on developing a few IDN high ranking keywords to see how high they can get for the keywords.