Google is indexing my affiliate thumbnail images!

[Nostalgic Dave]

I've got a strange problem here... Google has been indexing my affiliate product thumbnail images and attributing them to my domain in Images Search. I suspect this may be causing image ranking problems for the REAL images on my site. The affiliate images are on the affiliates domain, not mine so I don't know what I can do to stop this.

Can anyone please offer advice on how to fix this?

Thanks!

[phranque]

X-Robots-Tag: noindex HTTP Response header returned with those image requests.

[Nostalgic Dave]

Thank you phranque, I had not heard of the x-robots-tag before... very interesting, but since this is something in the header, I'd end up blocking all the images on the page from being indexed, not just the affiliate images. Correct? Unless there is way to block the images from just outside my own domain?

[rainborick]

You can send the X-Robots-Tag header for just selected URLs like your affiliate images via .htaccess, so it wouldn't affect the other images on a page. I'm sure if you ask over in the Apache forum you could get the specifics.

Of course, you could also block these URLs via your robots.txt file and they would eventually fall out of the index.

[Nostalgic Dave]

Thanks rainborick, I posted my question in the Apache forum after Googling for a solution with no luck. Obviously blocking an external URL is not common practice. Seems to me that indexing these is a bug with Googlebot, I can't possibly imagine why Google would want to do this.

[aakk9999]

The affiliate images are on the affiliates domain, not mine so I don't know what I can do to stop this.

X-Robots-Tag: noindex HTTP Response header returned with those image requests.

I think pharanque and rainborick have missed that the thumbs are hotlinked from the affiliate domain. You cannot send X-Robots-Tag response headers from your domain as the request goes to affiliate domain, not yours. You cannot block a particular image via robots.txt either for the same reason - they are not on your domain.

What you could try to perhaps do is:
- construct the img src via external javascript and block this javascript in robots
- use a small iframe to hold the image and noindex the iframe

However, you must think what would any of these (or some third action) do to your page preview in Google.

[Nostalgic Dave]

Interesting thoughts aakk9999... I was thinking about caching the images on my server, then blocking the folder that I store them in. An iframe may be the easiest, and if those images didn't show up in G's page preview, it wouldn't matter, they are not the focus of my website as far as the user is concerned. Should I be worried that an iframe containing affiliate ads could somehow be seen as a spam technique by Google?

[rainborick]

I'll admit I'm confused. If the image doesn't reside on your site, there's no way to use an X-ROBOTS-TAG header, and no way to block it via robots.txt. But it also shouldn't be indexed as belonging to your site. It just doesn't make sense. Google might associate the image with a page on your site in the standard search results if your page has the best quality and relevance. But if the image doesn't reside on your site, it should have no impact on your other images' rankings in image search.

It seems to me that the acid test would be to go to Image Search and use the site: operator to see if Google is indexing this image within your domain. If not, there's nothing to be done and no need to worry that I can see.

[Nostalgic Dave]

"It just doesn't make sense"... you got that right! This smells like a Google bug to me.. however, in thinking about it.. what about situations where a website serves it images from another domain/server for quicker response, server load, whatever. Then this would be a NICE thing to see. In my case, not so much! I've done the site: operator on my domain, then clicked on "images" in the sidebar, and there are thousands of images from my affiliate listed. Hovering over the image even shows my domain name!

[phranque]

I think pharanque and rainborick have missed that the thumbs are hotlinked from the affiliate domain.

yes indeed i either missed or misread that and eventually picked up on that detail in the apache thread.

Need to block indexing of images from affiliates domain:
http://www.webmasterworld.com/apache/4487711.htm [webmasterworld.com]

[phranque]

"It just doesn't make sense"... you got that right! This smells like a Google bug to me..

apparently it happens.
although it goes off topic, the OP in this thread tells essentially the same story from the opposite point of view...

Image Search ranks other websites hot-linking to MY images:
http://www.webmasterworld.com/google/4424682.htm [webmasterworld.com]

[indyank]

But it also shouldn't be indexed as belonging to your site. It just doesn't make sense.

I have been giving the same kind of example in the other thread where a scraper is being credited as the source for the images. I have also explained the reason why google is crediting the scraper in those cases. In this case the OP is equivalent to the scraper in my example :)

But Google might not be causing image ranking problems for the REAL images on OP's site due to this.

Solution to OP:

Why don't you serve all those images from your own sever (if you are allowed to do) ? if that is possible, you can serve all the affiliate thumbnail images from a folder on your server and use robots.txt block or x-robots noindex tag or by adding a "noimageindex" meta tag on pages serving those affiliate thumbnails. But while using "noimageindex" meta tag, you will also have to ensure that you don't have any other indexable images on those pages. So the x-robots noindex tag for all images served from a specific directory is a preferable option.

[indyank]

You can use the iframe technique only if the affiliate has separate html pages that link to each of their own thumbnails. Otherwise you will have to create those html pages on your domain and that would again result in google crediting those html pages on your domain for those thumbnails!

So you will also have to ensure that you use "noindex" and preferably "noimageindex" tag as well on all those html pages or serve those html pages from a directory that you block via robots.txt.

[edited by: indyank at 5:38 am (utc) on Aug 24, 2012]

[Nostalgic Dave]

Indy, thanks... I'll probably cache the images and x-robots noindex the folder.

[Nostalgic Dave]

I could do the iframe trick, because I'm actually creating the ads myself via their api, so I have tons of flexibility here. Would the iframe treatment possibly send any bad signals to Google?

[indyank]

As i said, if what you serve via the src attribute of an iframe is a url (html doc) on your domain, you will still be credited with the affiliate image unless you block those urls.

[Nostalgic Dave]

Yes, I understand Indyank, but my concern is with how Google might see iframes full of ads. Could they see it as a spammy technique possibly? Seems like we have to walk on eggshells with Google these days.

[phranque]

iframes full of ads

isn't iframes the most common implementation of externally-served ads?