Junk chars in Adsense ads

natural language chars encoding will not display correctly

[dhatz]

I've emailed this Q to Adsense techs, but since it's holiday and they may take a few days to respond, I'll ask the WW gurus as well:

I have enabled AdSense on some english language pages, in which the

meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1"

metatag is used.

Since AdSense serves geolocation-dependant ads, some ads are in national language (in testing I saw ads in greek and japanese - which didn't display right because they were written in a different charset encoding)

These ads (e.g. in greek) appear as junk characters.

What is the recommended method to avoid this phenomenon?

[tombola]

I use UTF-8 encoding on all my pages and never got one complaint.

<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />

[robho]

I use UTF-8 encoding on all my pages

Doesn't make any difference - the ad is on the iframe page, which has no encoding, and it seems to be garbled before it even gets there.

I've reported to Adsense today a page of mine which gets an ad in Japanese (on a page in English, viewed in the US).

The ad headline and text appears as all question marks even if the main page is UTF-8, even if the browser is unicode, and even if the browser (tried several) is set to a Japanese font (I can view other Japanese and mixed-font pages fine).

So it appears Adsense is messing up the fonts (at least sometimes) before they are displayed.

[zhenghua]

I've reported to Adsense too, I told them it may because of the GeoTargeting, some advertisers use non-ascii characters in their adverts but still target widen areas (most of their pages are iso-8859-1 encoding, ascii only). I've suggested Google to use UTF-8 encoding in their javascripted iframe. No reply so far...

Zhenghua

[Sanenet]

Errr... isn't this a client side problem caused by not having the regional language charset installed? You can have as many meta charset descriptors as you want, but if your system doesn't support Japanese or Cyrillic you're stuffed.

[dhatz]

Errr... isn't this a client side problem caused by not having the regional language charset installed? You can have as many meta charset descriptors as you want, but if your system doesn't support Japanese or Cyrillic you're stuffed.

No, I definately HAVE the charset and encoding settings required for the ads.

Let's take a concrete example of adsense ads in greek language: the ads display fine for me, when they appear as sponsored listings next to the G SERPs, and G uses iso-8859-7 charset for serving its SERPs to me.

Those same ads appear GARBLED when served via the AdSense js script on content sites, using 8859-1 charset.

According to my tests, it looks as if they got garbled BEFORE they are served to the user, ie by the Adsense server.

Adsense tech suggested this:

Please verify that the encoding for your Internet Explorer is enabled by
taking the following steps:
1. While viewing the page URI
in your browser, right-click on the page and select 'Encoding.'
2. Select 'Greek (ISO)' or 'Greek (Windows)'.
3. You may also choose 'Auto-Select' so that IE can automatically detect
the encoding required for the page. This may require that you have the
language pack for the page language installed.

Ofcourse I had tried every possible combination before emailing them ...

[Sanenet]

G uses iso-8859-7 charset

Where do they specify this? Can't find it anywhere. Surely, if you're specifing charset-1 page wide, then it will default to that for the Ifrane. Have you tried changing the charset pagewide?

[dhatz]

I meant that in G SERPs, in the HTML code, G specifies 8859-7 as encoding for the content it's serving to me.

In that case, greek ads / sponsored results (served next to regular results) appear fine.

[Sanenet]

Well, 8859-7 is ISO for Latin/Greek, so yes it would appear correctly for that code. But I get "charset=UTF-8" in Google (which supports anything, doesn't it?).

Try setting your pages to charset to UTF-8 and see if that works

[robho]

As mentioned above, the browser character set (unicode, japanese, whatever) doesn't make any difference, the?'s appear to be what is being served by Adsense. (I also tried different browsers, all of which display other Japanese/Chinese/Arabic etc pages fine).

I got a reply back from Adsense about this. They suggested blocking the ad. No mention of how I'm supposed to block ads which are not shown locally, which would be the majority of the affected ads, and no mention of any plans to fix the bug.

[dhatz]

I got a reply back from Adsense about this. They suggested blocking the ad. No mention of how I'm supposed to block ads which are not shown locally, which would be the majority of the affected ads, and no mention of any plans to fix the bug.

Just another friendly reminder to Google folks who read this forum, that non-latin character set Adsense ads for some languages (e.g. GREEK) DO NOT DISPLAY RIGHT on CONTENT SITES (but display fine alongside G's SERPs)

I see this specific problem all the time for greek language ads (charset iso-8859-7). I've NEVER seen a greek ad display correctly on Adsense content site.

When reporting this by email to adsense-tech, I got a very polite, canned but not helpful response like the one quoted above.

Btw, I've triple-checked this (although it's clear where the problem is) on several systems using IE 6.0, 5.5 and 5.0, using either auto-detect, and/or explicitly iso-8859-7 or UTF-8 fonts.

8859-7 Ads in G's own SERPs display fine. Ads in CONTENT sites never do. And I read greek text in my regular browsing all day long (ie no problems in my setup).