Blocked thumbnails. increased image search referals

[Luxoria]

I don't know if this is common knowledge but I figure I would post my assumption/findings.

I run a ecommerce site and each product has well described image file names. I noticed in Google image SERPs the thumbnails for other products came up when searching for specific products. (because on the bottom of my website has a 'featured' list with several other products w/images) The solution was to simply block the thumbnail folder in robots.txt. A few weeks later, after the thumbs dropped from the image SERPs, the traffic referrals from Google image search doubled. =)

I am not 100% positive if this is coincidence or cause and effect. My assumption is the latter.

In Google webmasters, image impressions have doubled and CTR has increased (by how much, I am not sure).

Would anyone care to comment on this and either support or poke holes in my assumption? (the site has been live since Jan 2011. And the majority of product images shown no impressions in Google Webmasters before the block)

[deadsea]

Google says that "image quality" is a major factor in how well images rank. In my experience, "image quality" is a euphemism for "bigger images". It does not surprise me at all that blocking thumbnails with robots.txt allows google to focus on the bigger images and as a result give you more traffic. I found that increasing my images to 800px width from 400px width gave additional traffic. I've seen REALLY huge images (10000px across) that seem to rank well because of their size.

With thumbnails and robots.txt, there is one thing to watch out for now. Given that googlebot renders pages for page preview now, you just need to be careful that your pages don't look too bare without the thumbnails on them in the previews.

[lucy24]

If I read your original post right, the thumbnails are for pictures of other, related products. So a user might search for widgets and be shown a thumbnail from the widgets page-- which by definition is not a widget. "What a stupid site-- that's not at all what I'm looking for!" But if they're only allowed to see the main picture, then the picture will show the thing they searched for.

Besides, if you make pretty widgets, it cuts down on the number of people who are just trawling for images to hotlink. I've had to robot-out a few directories for that reason alone.

[Luxoria]

Thanks for the advice of not blocking to many images. I looked at a few of my pages in Google search and it shows the thumbnails in the previews. So I am thinking for the screen shot Google loads the whole page images and all.

I am thinking about blocking the 300x300 images so that the 800x800+ images might provide even more boost. (click on the 300x300 to bring up the 800x800 type deal)

lucy24, Yes, you read my post correct. When searching for say Butterfly, the butterfly image came up in the SERP, but when clicked it will take the user to say Man Cave. The butterfly image was thumbed at the bottom of the Man Cave page (rotated randomly).

[jabas]

How can you block thousands of thumbnails if they are located in your main pictures (the large ones) folder ? I guess with robots.txt, but it would make a giant file, wich I think is not the best thing to do.. Have an idea on how to solve that ?

[deadsea]

You don't want to make a large robots.txt file. Some crawlers will have problems processing it.

Googlebot supports wildcards in robots.txt so you might be able to do something like:
Disallow: /images/*thumb*.jpg
if they are all named similarly.

Otherwise 301 redirect them to a different directory and block that.

[Luxoria]

That's a good idea using wild cards.

The shopping cart system I use separates thumbnails, medium size, and largest size pictures into different folders. So the trick was just to block the images/thumbs/ folder in robots.txt. =)

[jabas]

That's the problem, most of my file name are like this: "ebd9ce9b10d6b22f489k7247aafa5.jpg", so I can't apply wildcards..

[lucy24]

most of my file name are like this: "ebd9ce9b10d6b22f489k7247aafa5.jpg"

You may live to regret that :) I realize that once you get beyond a certain size, you're going to run out of variations on

dixie_looking_cage
dixie_on_cage
dixie_in_cage
dixie_runs_away

et cetera et cetera. But it will save you much grief if at least part of your filename gives some clue what it's a picture of, or what page it's associated with, or which product class it illustrates, or, or, or. Like that.

[jabas]

@lucy24 No because it's a md5 hash + the image extension, it is supposed to be unique, I guess.

I think I'm going to do a massive 301 to the new file names (this time containing the product/item/page name). I'm just afraid of the reaction of google.. They might not like it.

[Bewenched]

Very interesting. I'd been thinking that image sizes just might be related somehow to the Google panda/quality theory.

On our product pages we have thumbnails that if clicked use lightbox to display the much larger version. I wonder if there could be some correlation to our site being hit so hard by Google's rabid panda.

Thoughts?

[tedster]

I'm not saying there can't be some correlation, but I do work with one site that uses Lightbox variant to enlarge thumbnails and they did not have any impact from Panda - in fact, I think that's a pretty common scenario.

[deadsea]

If your thumbnails open the full size in a lightbox, and you want the images to rank in image search, it is very important to let googlebot know about the full sized images from high pagerank pages. One easy way to do this is to link to the full sized image, but use javascript to open the lightbox when a user clicks on the link. Something like:

<a href="large-picture.jpg" onclick="lightbox('large-picture.jpg');return false;"><img src="thumb-picture.jpg"></a>

As far as the image search algorithm goes, linking to an image and showing an image in the page are the same. By linking to the full sized image, you get the benefits of having the full sized image in the page, without using all the screen real estate.

[jabas]

You can also use shadowbox instead of lightbox, the syntax makes it easier for search engines :

<a href="largepic.jpg" rel="shadowbox"><img src="smallpic.jpg" /></a>

That will automatically be handled by shadowbox.

[levo]

The right way to block thumbnails and keep instant previews is to use x-robots-tag header:

"Q: I want to block my images from being indexed, but I’m happy with them appearing on a preview image; how can I juggle the two?
A: In order for images to be embedded in previews, it is important that they are not disallowed by your robots.txt file. In order to block crawlable images from being indexed, you can use the "noindex" x-robots-tag HTTP header element."

[sites.google.com...]

[lucy24]

Well, google can say that until they are blue in the face, but in fact you don't have to do anything at all, because the previewbot does not think of itself as a robot. Why else would it keep snuffling around in a directory that has been explicitly roboted-out? In fact I have just this instant (really) uploaded a new .htaccess to keep it the ### out.