Banned from image search - Panda connection?

[freejung]

Here's another interesting datapoint. I have an image site that was pandalized. You might think I would have noticed this before, but I actually pay almost no attention to image search. The traffic from it completely sucks - they almost always bounce, they almost never click ads, they just take the image and go. So I really don't care how I rank in image search and I've never attempted to optimize for it specifically.

However, some of my images used to do fairly well, and all of them would show up somewhere in the results for their exact name.

Now all but one are gone. The one that's left is not a particularly successful or interesting one either. Nonetheless, it's the only image from my site that appears in a site:domain.com search in Google Images. This image also appears prominently in the results for its name.

For one image with a particularly unusual title, my copy of the image on flickr (I only have a few images on flickr, this happens to be one of them) shows up, along with several copies on other sites, most of which link to me as the source of the image. The original is nowhere to be found.

I find this quite odd, and wonder if it could be related to my Panda problems.

I don't use a framebuster script or any other tricks. My images are downloadable from within the popup box that Google puts over the original page. I have all of my images tagged with schema.org metadata.

So... I don't get it. Why have my images been removed? Why does one remain? Why that particular one? It doesn't make any sense to me at all. If there is some technical problem preventing Google from indexing my images, why does it still index one of them? If that one image is OK, why aren't the others indexed? The template for the pages that display images is identical for all images on the site. This one is no different from the others in any way.

I'm thinking this might be a clue as to the problems the whole site is having. If I were Google, I would be extremely suspicious of an images site that only has one image in image search. It's an anomaly.

I know some of you have done quite a bit of work with image search -- does anyone have any idea why this would happen or what might be going on?

Edit to add:

Looking at traffic from google.com/imgres, it's clear that the problem occurred around the first of April, about a week before I was pandalized in Panda 2.0. Coincidence? Seems unlikely...

I don't have hotlinking disabled, I'm not blocking Googlebot from crawling the images that I know of, and I don't think I changed anything around that time. I think I'll check with my webhost and see if they made any changes, and spend some time investigating my server logs. There has to be a reason for this.

[aristotle]

What about the pages that contain the images? Are they still in Google's index, and if so, do they still rank for their main keywords?

[HuskyPup]

I don't have hotlinking disabled,

Do you mean you have hotlinking enabled in cpanel?

If so, that's your problem, it took me months of testing to resolve that issue.

[freejung]

What about the pages that contain the images? Are they still in Google's index, and if so, do they still rank for their main keywords?

Yes, and yes at least as well as can be expected given that the site is pandalized. These pages do not rank as well as they used to, but they still rank for their names as long as the name is not a common keyword.

Do you mean you have hotlinking enabled in cpanel?

This question is ambiguous. Cpanel has a feature called "hotlink protection" which adds code to your .htaccess file preventing your server from serving image files in response to a request with a referrer string other than your domain (and other domains that you whitelist). I don't use this feature, as I prefer to write my own .htaccess code.

About a year ago I used .htaccess to prevent hotlinking by redirecting invalid image requests to my licensing terms. This was a response to bandwidth spikes caused by hotlinking. I discovered, as did you (I read your posts at the time), that this interfered with the indexing of my images by Google, so I removed the code and purchased more bandwidth. Subsequently, my images returned to Google image search, and have been in it up until the beginning of this April.

Currently I have no restrictions whatsoever on access to my image files, either in .htaccess or in robots.txt. I am not using a framebreaker script or doing anything else that I know of that would interfere with Google image search in any way.

The one image that is still in image search shows up just fine and can be downloaded from the popup that appears in the image search results.

I can think of three possibilities:

-My webhost may have done something to prevent bots from downloading my images as a response to scraping. However I went to Google webmaster tools and told it to download an image as googlebot, and that worked fine.

-I am using some unusual CSS on the image pages. This was never a problem before, but maybe it is now. I can change that.

-I have incurred some sort of penalty, probably not related to Panda as nobody else seems to be reporting this problem. I'm actually kind of hoping this is the case, as it might explain the drop in traffic that I experienced in early April, which I thought was caused by Panda. Maybe if I can get this penalty lifted my traffic will recover, and it never had to do with Panda at all. Seems an odd coincidence though.

[lucy24]

I discovered, as did you (I read your posts at the time), that this interfered with the indexing of my images by Google, so I removed the code and purchased more bandwidth.

All you have to do is whitelist blank referrers. That will let robots in.

[freejung]

All you have to do is whitelist blank referrers.

Good point, I expect that would work. However, I don't particularly want to prevent hotlinking. Now that I have plenty of bandwidth, I actually want people to hotlink my images (as long as they give me a regular link too of course).

I had never submitted an image sitemap before. I just submitted one with all of the individual images associated with their appropriate URL. We'll see if that helps. If not, I'll probably submit a reinclusion request.

Also, interesting note -- I was considering removing all of my individual image pages and putting all of the images on the category pages as a reaction to Panda, but I just re-read Google's guidelines for publishing images, and they actually suggest that each important image on your site should have its own html page where you can display information such as the caption and perhaps offer versions at higher resolution, which I do. In fact, I am following the guidelines to the letter in all respects, as far as I can tell - they even suggest using a Creative Commons license, which I do. I'm glad I didn't remove the image pages!

[lucy24]

I actually want people to hotlink my images (as long as they give me a regular link too of course)

How will you know?

[freejung]

Ironically I usually find out through Google image search.

[freejung]

*facepalm*

I was being filtered by safe search. It didn't even occur to me to check, as none of my photos could even remotely be considered to be "adult" content, they don't even have people in them!

Contacted Google, I expect it'll be resolved.