Forum Moderators: martinibuster
I run a bilingual site with English and Swedish content and have found that a lot of my Swedish pages just won't get anything but PSA ads, allthough they are packed with keywords that should render targeted ads.
I've spent some time analyzing this, and run some tests and found that Google can't distinguish between languages when filtering out adult content.
Some expamples
Slut means End in Swedish. It's just impossible to use on a page if you want anything but PSAs.
Sex is the Swedish word for the number six (6). And it's a definite no-no. However, this is a homonyme even in Swedish, where it also bears the same meaning as in English.
Swedish homonymes
But words with different meanings within a language will also give the same results:
Slinka has several meanings in Swedish. It could mean pop or slip as in slip into something more comfortable or slip into something more comfortable. It also bears the same meaning as the most common English word used by rappers to describe women. This word is clearly entered in Google's list.
What experiences have others from other languages? I guess visitthailand would have problems withe the very frequent Thai word that begins with a p and ends with orn?
Google has also reviewed my site at at least one occation without pointing this out for me. Still. As it is, I will have to go back to those other advertising methods for the 40% of my site that isn't in English.
A pity.
I don't run AdSense on a site with a language other than English (yet). But I have run into significant problems with trigger words on a site with a topic of crime. Words such as sex, porn and rape are frequently mentioned on these pages. Most (but not all) of the pages show unpaid PSA ads because of this.
However I've been unable to determine if a set number of trigger words or trigger word to total word ratio triggers the blocking. Have you?
For now I've removed AdSense from most of the pages of the site. I've considered writing a script to automatically change or obfuscate potential trigger words, but I am concerned that this will annoy users and hurt search engine rankings for pages that already rank highly. Obfuscating would be easy enough - "sex" could become "s*x", but coming up with drop-in word replacements would be tough and read poorly in many cases. I don't know a good synonym for "rape" that wouldn't be a trigger word.
However I've been unable to determine if a set number of trigger words or trigger word to total word ratio triggers the blocking. Have you?
No. In my case one word was enough. Re-publishing the page with a different URL and the suspective word (one was enough!) showed targeted ads.
I don't think you should scramble you text for regular visitors, but maybe you should do some evil cloaking on you pages. Show the media-bot scrambled words like s*x and r@pe.
Aaarghh! This moralism makes me crazy! Especially with a site like yours! What do they think? That problems will go away if we don't talk about them?
Some might suggest that the site isn't well-suited for AdSense then. Maybe, maybe not. It's a content site with an interactive component and there are plenty of Adwords ads (though with extremely low CPCs) that are well-targeted to the site's users. It's just a matter of avoiding the triggers so they're displayed.
I'm optimistic that Google will eventually figure out how to handle such pages so they aren't flagged as adult or negative content so long as there's enough reason for them to do so. I've seen the same problem on numerous news sites, but perhaps they're not aware of it or it affects such a small percentage of page views that they don't care.
In any case, it would seem that it would probably be effective for Google to allow certain sites to be white-listed to avoid certain triggers. After all, they shouldn't be concerned that a site about sex crimes and sex offenders is going to sneak in an adult content page after 5+ years on the Internet.
Apologies for getting this thread off-topic.
No. In my case one word was enough.
Wow, interesting. I need to do some more testing, but it appears that in at least some cases a single word doesn't cause blocking. Maybe each trigger word has a score and the total page score has to exceed a certain level to be blocked and/or word density comes into play somehow. I'll try to construct some tests and see what I find.
I say this because the single page I'm currently running AdSense on has the phrase "sex offender" on it numerous times, "rapists" once and "child molester" once. It's possible that none of those are exact matches for trigger words though.
I don't think you should scramble you text for regular visitors, but maybe you should do some evil cloaking on you pages. Show the media-bot scrambled words like s*x and r@pe.
I've considered that, but I'm worried that Google might detect that and drop the site. Has anyone tried this? Thoughts?
Also, I wonder if a page being flagged by Google as "explicit sexual content" as many of the site's pages apparently are using Google's SafeSearch filtering is a contributing factor. See [google.com ]. Scrambling text for the regular Google bot to avoid that would be a bad move.
Aaarghh! This moralism makes me crazy! Especially with a site like yours! What do they think? That problems will go away if we don't talk about them?
It's frustrating. And it's too bad because AdSense could be a good revenue source for the site to help fund hosting, mailings and other expenses since I prefer not to have untargeted and prominent advertising and I do little to encourage donations so funding is almost entirely out of my pocket.
You can do a normal page and do an image capture from the censored words.
Then, you can do a replace with any html editor (replace "child porn" for "<IMG SRC=rplcchildporn.gif width=43 height=12 border=0>"
You can do a normal page and do an image capture from the censored words
Mauricio, I hadn't thought about that - thanks for the suggestion. If they keywords I'd have to replace weren't found in most of the top search engine searches that generate traffic I'd probably implement that on some pages right away. But I need to think about it more and consider the risks.
A Google fix would be much better than jumping through hoops though.
Anyway, the problem now for instance. I have a news story about a rapper named 'Fatal Hussein', now my pages display PSA's. Also have a conspiracy theory page talking about a false autopsy report that will only show PSA's.
Of course anything with profanity I've removed or censored. I'm still trying to figure out what other trigger words cause the PSAs.
Maybe it would help if we put together a list?
That sounds like too much work.
How about using ASCII codes in your HTML? Like a displays the letter 'a' and e displays the letter 'e'. Not sure if this is kosher with Google though.
