Images not indexed in Google - possible reasons?

[marvin]

I rank well in Google SERPS (top 5 results) for several keywords and have done so for some time now (over 6 months).

On the pages which rank well for these keywords I have several pictures (relating to the keyword) which are just not getting picked up by Google images. I have searched this forum for possible reasons but I seem to be doing everything right, eg.

I have no robots.txt file which might restrict search engines from crawling.
I use the keywords in the image title and alt tag, for example - the html for a typical image of mine would be:-

<img class="xx" src="photographs/keyword/keyword_1_medium.jpg" alt="keyword blah blah blah" />

On a typical page I may have up to about 10 images relating to the main keyword, the HTML of each one following the same format.

Does anybody have any suggestions why my pictures aren't getting displayed in Google Image SERPS?

NB - I have the same problem with Altavista and Yahoo (MSN search does, however, display my images).

[Quadrille]

Is the page <title> tag relevant?

Is the page text relevant? In particular, is there a picture caption which is relevant?

My reading suggests you need AT LEAST three out of four of:

1. image filename (keyword.jpg or keyword-whatever.jgp NOT keyword_whatever.jpg)

2. page / site relevance

3. ALT attribute relevance

4. caption relevance.

My experience confirms page / site relevance + ALT relevance.

[marvin]

Thanks Quadrille

The Page, Site and Alt tag are all relevant to the image.

I have not come across the term "caption" before. Is this something separate from the alt tag?

I had no idea that "keyword-whatever.jpg" is better than "keyword_whatever.jpg", I'll fix these. Is this also true of urls? (for example would www.mysite.com/keyword-whatever.htm be better than www.mysite.com/keyword_whatever.htm)?

[Quadrille]

I use caption in a 'print' sense - the 'legend'; the words on the page describing the image.

I suspect that for SEs, it's a bit looser; text near or around the image, rather than necessarily the text immediately describing it.

[jbinbpt]

Have you tried this in Google Images? site:domain.com

[marvin]

Thanks for the suggestion jbinbpt

When i try that I get "Your search - site:domain.com - did not match any documents."

The same thing happens when I try site:www.domain.com

[jbinbpt]

I would run a limited test by moving some images up so they are in the same directory as your code. I read somewhere, that keyword density is a factor.

[marvin]

Quadrille and jbinbpt

Thanks for all of your help. I've just thought of something which may explain my lack of image results.

Until recently I had incorrectly been using <image src=".... instead of <img src="... on my site. This displayed well on IE so I wasn't aware of the problem. However, about a month ago I started validating my pages (and making them firefox friendly) so I discovered the problem and fixed it.

Is this likely to explain why Google image search has not been picking up any images on my site?

[jbinbpt]

From W3C....
[w3.org...]

[loudspeaker]

I've had this problem for a long time - at least in my case, Google seemed to ignore images on pages it considers too low-ranking. This frustrates me because I consider some of my photos pretty good - Google image results on relevant keywords are filled with low-res garbage and yet the spider won't even take some of my photos (it "visits" only once in 2-3 months taking no more than a few dozen images at a time).

I think it will quickly grab an image if your PR is 4-5 on the page that embeds that image. For lower-ranking pages, it may take months and even years (seriously, I have plenty of excellent high resolution images that have been there for almost 2 years and Google image spider hasn't even touched them).

[jbinbpt]

Loudspeaker,
I don't believe that quality is an issue. Quality is subjective and the algo used is not. Are all of your images hi-res? Do an image search and look at the image sizes. The general rule of thumb was 72 dpi was the best a monitor could display. Perhaps Google does not like big images?

[Quadrille]

I don't think the image quality or size comes into it.

I've seen no evidence that the SE Algo has any way - or interest - in the image file itself, beyond the fact that it is an image file.

I've plenty of images indexed that are good, bad indifferent, small, large, whatever.

I suspect age and stability are a factor - I don't think the spider works as hard as the content spidering (that's a guess!).

With my sites, the ones where images are most often indexed seem to be the ones where the pages are most focussed; I suspect because they have to rely on 'circumstantial evidence', a tightly focussed site reduces their risk of error.

The algo has no way of knowing whether a picture called paris.jpg on a site about Paris, France is actually the Eiffel Tower, Paris Hilton, or someone's Mom in Paris, Texas. So a focussed page reduces their risk. This also explains why some results are downright weird in picture search!

The only site of mine that seems never to get pictures indexed is the one where the pictures are not hosted at the same domain - all my other sites, they are, but always in a separate folder.

[HarryM]

I think there has been a change in the way Google indexes images. The traffic I used to get from images a few month ago has dropped considerably, and many of my images have disappeared from the index. This includes quality images on well-focussed pages.

In my case this appears to be related to the supplemental page issue. I suspect that Google removes images from its index where the containing page does not meet certain criteria, possible PR.

Google has always been slow to index images, which it seems to do periodically rather than continuously. So it may take some time for new images to get indexed.

[piatkow]

small illustrations on my main pages get indexed, images in my photo gallery are mostly not indexed although small subsets seem to come and go from time to time.

It is always worth checking with safe search both on and off, an embedded character string within the title can be enough to flag the images as "adult".

[HuskyPup]

NOT keyword_whatever.jpg

If you want to waste your time altering all your files names please feel free to do so however I have several thousand images all with the underscore and all ranking at the top. I can assure you that this is not necessary regardless of what may have been written in certain blogs etc.

This also applies to html file extensions.

[HarryM]

NOT keyword_whatever.jpg

I think this dates from a time when SEs only recognized words if they were separated by a space or hyphen. An underscore was taken as part of the word. However Google can now pick out words even in a string such as 'keywordwhatever', so an underscore should no longer be a problem.

Having said that I wouldn't use an underscore in a new file or image name.

[Quadrille]

Having said that I wouldn't use an underscore in a new file or image name.

Me too.

But with images, filenames seem to be a fairly small factor - and stability matters - so I'd not change existing (listed) files.

[iridiax]

I think there has been a change in the way Google indexes images. The traffic I used to get from images a few month ago has dropped considerably, and many of my images have disappeared from the index. This includes quality images on well-focussed pages.

In my case this appears to be related to the supplemental page issue. I suspect that Google removes images from its index where the containing page does not meet certain criteria, possible PR.

Yes, I can confirm this. Some of my pages recently ended up in the supplemental index (verified by site:example.com), and Google removed all of the images on those pages from Google Image Search. When some of those pages mysteriously popped out of the supplemental index, the associated images were suddenly back in the image index.

[HarryM]

When some of those pages mysteriously popped out of the supplemental index, the associated images were suddenly back in the image index

Thanks, iridiax. That gives me some hope!

[zeus]

ok here is :

forget _ or - thats not of any value, if we are talking under 2%

images will be picked up if its a PR1 or PR8 dossent matter

quality of images, also dont matter.

what is a problem is there damn moderate filter - try a search for you images with this filter off and you see some pictures, they just ca not control that filter, it filteres everything.

[zeus]

< The next 3 posts were spliced
into this thread from another location >

oh that really is getting a pain in the a.. everytime they try to filter bad from good, mostly they take 95% of all images from a site, which have no nude images at all on the site, but somehow they filter everything out.

4-5 month ago they where on a good parth, they FINALY made some changes to there filter and I saw many sites had there images included again, in the google images and I also saw no dirty, so a good job done, but now they are at it again, I see a lot of competition, also our site, getting filtered hmm 95-98% of the image search and we have NO nude stuff also no dirty text, we have a filter for that.

I really hope they are still working on there filter and we are really dependend on the image search, which many in our category.

[edited by: tedster at 10:20 pm (utc) on July 15, 2007]

[tedster]

Are there also updated results with the filter off? Or is it just changes to the filter itself, and not the overall scoring algo.

[zeus]

as I see they are only messing with there "Moderate SafeSearch is on" i see a lot of images go from the index and we are NOT talking naughty pictures, we are talking, space, fish, nature... everything, sites are loosing all there pictures with filter on.

If you turn the filter off the pictures are there.

I know some time ago - 8month they had also filtered everything there is for a year, but then also 8 mont ago until now I saw some images getting back into the search results.

I dont understand this, why do they mess with the filter so muc, why not just write a text, "we are not able to see the pictures, so there will be some unfit images in the results" be cause they are realy filtering A LOT of images.

[dailypress]

< The next 6 posts were spliced into
this thread from another location >

For some reason, I dont get Google Images or MSN Images referrals from a site about 7 months old. I added text under the image and hyperlinked both to individual pages.

The image dimension is small (about 100*100px). I used ALT tags as well. What do you think the cause would be? I am thinking the image dimension might be the cause, or would it be the fact that the site is 7 months old?

[edited by: tedster at 10:23 pm (utc) on July 15, 2007]

[tedster]

I've moved this question into another thread that may give you some ideas:

Also, yes, 7 months is relatively young for a website, so that may be part of the picture. Also, are the related search terms the kind of thing that would be a common image search?

[edited by: tedster at 10:21 pm (utc) on July 15, 2007]

[dailypress]

Thanks Tedster. I have searched the exact phrase I used and it doesnt show up. I learned that google mostly cares about the file name and not the ALT text.

[tedster]

I only have one client who gets good traffic from image search. I just checked it out, and I do see alt text being valued. I also see title tags from "enlargement" pages showing up prominently. In fact, these particular images do not have the keyword that I used for searching in their filename at all.

So I'm pretty sure that there's a more complex algorithm involved. Not to say that the filename doesn't matter - I know it does. But on a widget site, I'm not going to include "widget" in the name of every image file.

[DamonHD]

Google Image Search is *much* *much* slower than regular search to pick up new material. I don't count on less than six months ever for a new image to get into Image Search. I can generally get a new page into normal search in two days or less.

Rgds

Damon

[zeus]

anyone els here have problems with the filter on and off, I see today it has not changed, they are still filtering wholes sites for nothing and with filter off i then see great nature pics and more.

[piatkow]

nature = naturist = naughty