Has anyone noticed that image search traffic has vanished

[Archbob]

I think it happened sometime between Tuesday and Wednesday when traffic from images.google.fr, images.google.de, and other international image searches just all of a sudden disappeared. I still get the same amount of search engine traffic from regular google and from places like docs.google.com but the referrals from the non-english image search vanished. I find this weird as when I search images.google.fr and others from my images, they are still in the same position. Was there some change to the way google analytics counted these referrals as since now they are gone?

[agent_x]

Now you mention it yes, they do seem to have disappeared apart from one hit from images.google.co.id yesterday. I normally get more than that.

[Lake]

Someone mentioned that they rolled out the new image search design to those countries? (As in, they used to show the actual page vs. making users click a button now) Is this true?

[Rasputin]

I am also seeing this.
If i access our site through an image then look at analytics real time, the page is shown as a referral from google organic, not images.google, so I think the visitors are still reaching the site from images but analytics is reporting them differently
But i cant be sure because I dont know how they used to show in real time and since google images is such a small % (about 1%) of our total I wouldn't otherwise notice if they disappeared altogether - certainly they prefer not to send traffic from images nowadays!

[RedBar]

Google's making a total mess of images, I reported here in January:

[webmasterworld.com...]

I had tens of thousands of images indexed yet many have been removed over the past couple of months or so even though they were my own and unique.

[dhaliwal]

they had done this few years back for all countries except Germany and France (because they were fearing some local laws there).
maybe, now they have done it for these two countries as well.

they have stolen lot of traffic from our websites in this way.

Google and Bing, both are letting people steal images from websites.

[Archbob]

My images aren't being removed, they are still there, its just the referrer traffic that is gone. I mean its not a huge deal even though it is like 8% of my traffic because that traffic had like a 90% bounce rate.

[aristotle]

doesn't google have its own copies of your images and shows them on its own site. So people can see them there, without needing to come to your site at all.

[Archbob]

No, google doesn't when you download image its from your servers, not google's.

[aristotle]

No, google doesn't when you download image its from your servers, not google's.

But can't they get the image without coming to look at your page that has the image. That's what I meant by not needing to come to your site.

Also, I believe that they can get a lower-resolution copy of your image directly from google's servers.

[Dimitri]

@aristotle, just try the image search yourself, and all your questions will be answered.

Since a couple of years, Google is displaying the full resolution image, served from YOUR server, but displayed at Google's page. So, yes, people can find images, view them in full resolution, download them if they want, and all this without ever visiting your site, but still consuming your bandwidth.

In the past, your site was showing in the background of the "preview" image, giving a chance for visitor to see your page and eventually be interested in visiting it.

Also, don't try to use anti hot linking script on your images, otherwise Google will kick you out of the image search.

Now, the only "positive" point, is that at least when you click on the full resolution image, it takes you to your site, whereas at Yahoo/Bing/Yandex (which are doing the same now), it shows the full size image. So basically, you can still get a chance toget some trafic from Google Image search, while you won't get any from Yahoo/Bing/Yandex.

[lucy24]

But can't they get the image without coming to look at your page that has the image. That's what I meant by not needing to come to your site.

Also, I believe that they can get a lower-resolution copy of your image directly from google's servers.

I remember experimenting on this point when they first rolled out the new-style image search a few years ago. (Disclaimer: Details may have changed in the interim.) At the time, they would request a fresh copy of the image as soon as the user clicked on its thumbnail--but they wouldn't actually display the new version. You can use this to your advantage by rewriting to a smaller file so they don't waste as much bandwidth. It was especially easy to tell if the original was .jpg, because their archived version is always in .png format.

[aristotle]

People copy your images to pinterest, and google's image results show those pinterest copies instead of your originals

[Dimitri]

People also "Like" (Facebook) or "Tweet" (Twitter), "your" image from Google Image search, meaning the link goes to Google image search, and not your site...

[chrisv1963]

It's very simple: Google and Pinterest are stealing our photos, bandwidth, visitors and income! They have been doing this for years and it's about time that somebody is doing something about it. I'm sick and tired of seeing my work being stolen by these companies that make millions/billions!

[OnlyOne]

My images started dropping out of the search results around the 24th January. All my images are large high quality, high resolution original. Many of my galleries used to do very well, with sometimes as many as 10 images in the top 5 rows. Now I'm lucky if I see more than one on the entire page. It appears to be (at least partly) a censorship issue! I also think traffic throttling is at play. I have made the following observations, these last few days:

1. Galleries with images that were very popular have been the hardest hit with no images remaining in search results for some galleries. Some (about 10%) of these images contained sexy poses but many other benign images have also been removed. Where someone has copied my image, Google is now showing that instead of mine (irrespective of how family-friendly the image is or isn't!). I initially thought I was being penalized for having some mild adult content but other galleries without any of this type of content have also been hit.
2. Mid last year, I removed a fair-use copyright disclaimer from the bottom of the page when I updated some of the more popular galleries. Except for one, the galleries that still have this one sentence of text are unaffected, Last week, I put the text back, but have yet to see any improvement.
3. On the Landing Pages page in Google Analytics, over a 24hr period, I have a long row of ones in the Sessions column. This, to me, looks like as soon as an image gets clicked, it is removed from the search results for that term and possibly other related search terms.
4. For the last seven days, the number of sessions has been around the same figure.

In 2015, I was issued an image mismatch penalty which I recovered from. Last October, a scraper site hotlinked a large number of my images from one gallery and Google decided to show the Url to their site instead of mine. I decided to put a hotlink redirect on for that gallery only. Maybe this is a penalty?

[OnlyOne]

People copy your images to pinterest, and google's image results show those pinterest copies instead of your originals

Serve them up a small thumbnail when they click on the Pinit button. Google doesn't have much interest in very small images.

[aristotle]

Mid last year, I removed a fair-use copyright disclaimer from the bottom of the page when I updated some of the more popular galleries. Except for one, the galleries that still have this one sentence of text are unaffected

Can you show the text of that disclaimer. Does it mean you give permission for fair use copying? I don't understand how it works, especially for images. Please explain

[rustybrick]

Might be related, see [support.google.com...]

Search Analytics Report:

February 7 – Ongoing (Germany and France only)
A logging issue is causing loss of image search data for image searches originating from Germany and France. This issue effects only the Search Analytics data, not image presence in Google Image Search.

[Dimitri]

I learned to live with the loss of image search traffic, before the switch, I was getting 100.000 visitors per day, from Image Search, now it's like 5.000. But, I am not blocking Google image bot, or trying anything, because, 5.000 visitors from Google image search, is still not to be neglected, and it still contributes, in a way, to promote my site, and give the idea to some to visit it to look at galleries. From time to time, it happens that big news sites, are finding images at Google image search and adding a link to their article, and they are fair enough to link to my page, and not the image directly. And I can tell you that when the BBC, adds a link to you in one of their article that thousands of visitors (for a few days), even if your link is lost in the middle of the article. So still worth it, (and I guess that a link from this kind of site is given good value by Google).

[RedBar]

I was getting 100.000 visitors per day, from Image Search, now it's like 5.000.

Same for me, many, many people have never believed me!

[Dimitri]

>> Same for me, many, many people have never believed me! <<

#BrotherOfMisfortune

[Mentat]

Big drop in image referer since 1-2 February 2017

[RedBar]

Coincidence or not?

In my widget sector if you're not a .com you're in double trouble.

I have a specific .cn with 55 pages of my widget images, not one single image is cached and every one of the widget pages, although supposedly indexed, does not have the cached green v on the rhs.

Of the other pages with a cache some go back to the beginning of December.

[rustybrick]

Google confirmed for me they rolled out the 2013 image search interface changes they had in the US to other countries last week.

[Dimitri]

Oh, there was no need of confirmation, you just had to go to Google germany, france, etc, ... to notice they are now using the US image search layout.

[JS_Harris]

Having experienced the above when the changes rolled out on .com years ago this is deja-vu for me so I'll add my two cents in hope it helps someone else.

No, google doesn't when you download image its from your servers, not google's.

This is partially incorrect. Google does keep a lower resolution version of all images which it actually displays instantly when you expand an image on their site. The design simultaneously is downloading the image from your site and if/when successful your image is swapped into place on the Google page seamlessly, without reload. If you block Google from downloading the image, or your site is down, that attempt fails and Google still shows their copy.

This broke webmasters being able to redirect visitors to their sites whenever Google tried to load the image, "if image is not on my domain then redirect visitor to my domain" no longer works. I thought it was because my page couldn't be loaded as a preview behind the image for a while but when I dug in and tried to force the same on images it was clear they simply abort the load attempt if you try to use this type of code.

The reason for the drop, however, is that any incentive to come visit your site is lost when the image is already displayed in Google search and can be enlarged directly on google's page. Google did add a couple of links that help a visitor reach your site but, in my experience, those were rarely used. Again, the person looking up images already saw what they wanted to see.

Three popular "fixes" were used by many webmasters
#1 - Do nothing, it is what it is. I think most webmasters left things as is and moved on.

#2 - Create a "click on image to see full size" semi transparent watermark that gets loaded over the image in Google search image results. This involved some code to detect the referrer (which google later removed causing a cat and mouse game with webmasters) but has long since been made into a plugin and extensively described online. This option encourages clicks from people who want to see the image but the bounce rate is extremely high.

#3 - The approach I ultimately ended up using, a robots.txt entry prohibiting google's image bot from indexing images. Unlike with web pages where Google will still index a page with a "the contents can not be shown due to robots.txt" as description, Google image simply does not index the image at all.

Why I went with #3. In the new layout I found that not only was traffic reduced by 90% the quality of that traffic also took a nosedive. I was left with 10% of my original image traffic which was also twice as likely to bounce back to Google to see more images. Not only that but I was still left with 100% of the scrapers of my images which had the effect of making them all duplicate anyway, and so wrongly credited to others. Scraping became easy and I couldn't block it because the scraper didn't even need to visit my site anymore leaving me no way to stop them.

My metrics in the Google dashboard, then known as webmaster tools, instantly and very dramatically changed when I blocked Google from my images. CTR skyrocketed because all data was now web based and not skewed by image data which had low CTR. All of the data in my account also became extremely useful and easy to understand because it was 100% web data and not an image/web mix. The settings for "web only" or "image only" did not filter 100% of the other, that also became obvious after the full image block.

In short I lost no benefits, 10% of the original traffic with a high bounce rate wasn't worth it, but I gained a lot in terms of improved metrics across the board. For sites who relied heavily on image traffic, and have a site that does well with a visual crowd, social sites like Pinterest replaced the lost traffic over the next year or so.

I now 100% block Google's image search and, unless they change their layout again to send more traffic, doubt I'll ever enble it again. I sincerely hoped they would revist their new layout when the effects were clear for webmasters, and I even joined the discussion to ask them for exactly that, but nothing has changed since. I guess that since it's been rolled out to other countries just recently it's not likely to change soon either. Maybe some day...

[keyplyr]

I've noticed no change in daily image search referrals.

I'm niche based so visitors following through image search usually read other pages.Yes, much less than when I used a script breaker, but still a steady flow.

I pay little attention to GSC metrics so no loss there. I learned a long time ago to diversify and not rely solely on Google traffic.

[MrSavage]

It's good to see confirmation of the negative effect of the image search. Drastic enough to notice and it's a reminder of what's important and what's not. The comments here affirm when this first rolled how much it destroyed image traffic. I believe this was part of the transition. My image traffic dependent site still exists but I've spent around 30 minutes on it over the past years since this user friendly Google image search interface took over. Let's not forget about the Getty images lawsuit though.

[lucy24]

a robots.txt entry prohibiting google's image bot from indexing images. Unlike with web pages where Google will still index a page with a "the contents can not be shown due to robots.txt" as description

Careful there: crawling and indexing are different things. Denying a crawler in robots.txt means it can't see the content. Sending a "noindex" header (or using a noindex meta on a page) means it can't be shown in searches, whether or not the search engine knows what the material looks like.

In the case of a roboted-out page, the URL may still come up in SERPs if vast numbers of good sites link to it. With a roboted-out image, there would be no point, because there's nothing at all to show in the SERP.

So when you deny a non-page file in robots.txt, you are technically not saying anything at all about indexing.

Food-for-thought question: Is a search engine's* crawl budget for your site determined by how many files exist, or how many files are accessible to them? That is, if significant numbers of images are roboted-out, will they crawl other stuff more often, or will there be a net reduction in the amount of bandwidth the search engine uses?


* Yes, I realize this is the Google subforum. But most-though-not-all behaviors are shared by all search engines.