Google goes too far with new image search
A couple of days ago, Google released a new version of its image search. This turned out to be a huge slap in the face of content creators like me. When clicking on a thumbnail, the original image is hotlinked and embedded into Google’s result page. This costs bandwidth and the user has less incentive to visit the webpage of the original creator.
Here is a nice comment about the topic by a webmaster called EcoCatLady in response to Google’s blog post:
Well, it is easy to use – but it’s killing my biggest web site (a photography site) which is also my main source of income. In the past few days my page views have been cut in half while my bandwidth usage has increased – thanks to Google’s hotlinking of high resolution photos.
If this continues it will force me to either make some dramatic changes to the site (ie: removing all high resolution images and forcing the user to jump through a bunch of hoops to get them) or it will put me out of business all together.
I really don’t understand why Google insists upon harassing image publishers in this way… for all other types of content Google allows the user to find the content, but sends the user to the page to read the article or watch the video. But for image publishers it simply provides an easy way for people to access our content while circumventing our websites completely. The least you could do would be to disable right clicking on the hotlinked image and get rid of the “view original image” button so people would have to come to the site to download the content. It’s only fair…
I know user experience is paramount, and I’m all about share and share alike (I’ve even released all of my photos into the public domain.) But bandwidth costs money, and publishers do have to make a living, and for most of us that means we rely on page views and ad revenue. Is this groovy search feature really worthwhile if it puts the publishers out of business and ultimately means that quality images are being removed from the web because we simply can’t afford it anymore?
From a user’s perspective, I actually really like the change. From a content publisher’s perspective though I can see how it might be problematic…
If you are worried about Ad revenue, one thing you could do is make thumbnails for the RSS. I go though my Google Reader every day, and see lots of webcomics with their full comic image on there, there’s no reason for someone reading the RSS to click to the website.
You could also put advertisements in your RSS. Might be a better choice, actually.
Your core argument about hotlinking for the image search is good. It is bad internets manners to hotlink.
(here’s my 172 webcomics RSS if anyone is interested, no need to register to view- http://www.google.com/reader/bundle/user%2F04909202417962126659%2Fbundle%2FWebcomics_0 )
Sort of like ad blocking software…
huh… I thought that’s what google was doing already.
that sort of thing is exactly why most “professional” photographers do that sort of google-foiling already. Or let a site like shutterfly, etc. do it for them.
Ironically (or something), the whole reason why google is doing what it is doing.. is exactly because of overly heavy revenue generation tricks. ie:
1. use google to find a cool image.
2. click through to the site
3. be bombarded by 50 different revenue schemes, and redirects, just to attempt to see 1 funny image, or whatever it may be
So at the moment, google is probably attempting to make its users happy, rather than making providers, happy.
This “problem” could probably best be solved, by the CrazyCatLady getting better/cheaper hosting for her pics
@ Forest_GS: I’m guessing the reason Google chose to use hotlinks instead of hosting the images themselves is that they didn’t want to be get sued for copyright infringement. From a legal perspective hot-linking is fine, but copying the images to their own servers is not; even though the former practice is actually worse for content publishers than the latter.
@Ajedi32 – You are right it is not legal for google copy and host full resolution images it is ok for them to copy scale the resolution down and host and serve those though. At least according to the last few court cases I have seen.
You might have to double-check with Google about this – but if your thumbnails are in one directory, and your highest-resolution images are in another, then use robots.txt to disallow Google down the high-rez directory. The only real problem then I see is that users won’t necessarily know that you have higher resolution versions available.
If you e-mail Google and ask for help, possibly suggesting some routes you might take and requesting feedback on which is the best, you might get some good results. Referrer requirements might help, too – I dunno.
If you accuse them of trying to put you out of business, you’re likely to get nowhere. If you involve them in your concerns, subtly playing up the “I’m just a poor idiot and I’m asking help from the Net Gods” approach to flatter them, you might be surprised at just helpful they can be!
I think you can block the Image search crawler with a robots.txt file. It could mitigate the problem until a better solution arise. It would block Google Image Search while keeping the site as it is now.
I know! I went to search for something the other day and had a really hard time visiting the site. Another downside from the user perspective is that the images don’t ‘pop-out’ when you hover over them, so you can’t get a better view without clicking their embedded link. I’m very grumpy with the change.
You can easily disable hotlinking be checking the referer. You could just send a small smileyface icon to Google hotlinked images instead or no image at all. Or maybe a different preview/teaser pic for users to click your website.
Google is basically just doing hotlinking because they can – and it can easily be prevented.
The new image search is definitely not an improvement. It’s giving content-creators more hassle for fewer rewards, and it’s also inconvenient to use, at least for me. I agree that it is going too far.
My initial response, as a user, was positive, since I was getting sick of sites that use code to break out of the search and show me things unrelated to what I was looking for, but I can certainly understand and agree about how it can hard the content creators. Maybe there’s some sort of middle ground, but I don’t know the code well enough to say.
You could put a handler on your web application that throws an exception or drastically reduce image size for images that are not requested directly from your own html client. There might also be some sort of micro formatting that will tell google search to avoid crawling certain content. Google has one of the most open API’s on the interwebs. Certainly there is a way to protect content.
I really like the change and the new UI is very very nice. Overall a big win for consumers. The hotlinking sucks for providers because it does drain their bandwidth. However, for Google to maintain a full-resolution copy of the original image would put it into very shaky legal territory with regards to copyright. If Google hotlinks to an image they’re fine legally because displaying thumbnails are fair use – but caching the full image would invite a barrage of lawsuits from content creators (many more lawsuits than they’ll get for simply hotlinking).
In the new UI there is also a prominent “Visit Page” button and clicking the picture itself also takes you directly to the site.
@ Philip: Even cheep hosting is going to cost if Google is hotlinking all the images on your site. Musings has over 28000 images and each one has 3 different resolutions. All those take up over 2 gigs of server space. Can you imaging all those images being viewed an not once going to the website to see the written content that goes with it nor see the ads (which are minimal) which help to pay for the bandwidth and hosting?
I am seriously thinking of killing all access to images for the three search engines just to be able to maybe be able to keep the site up. Screen shots is what brings many people to the site and many of them stay to look around. If they don’t have to go to the site then why should I even have the site up to begin with?
@ Emanuel: They are not using thumbnails but actually hot linking the full size high resolution image to display to the searcher so that they never even have to leave google at all. Google gets all the benefits of the content the the creator sees their cost go up and their traffic fall downward.
I foresee possible lawsuits from image heavy site owners against Google.
This is bad for webcomic artists. Since most of them feed from purely ad revenues.
Can’t one serve low resolution derivatives of pictures under the same URL, when an http request originates from Google? I know that’s extra work to setup in the content pipeline, and it’s only a possibility when confronted with easily identifiable and very limited numbers of blacklisted origins. But if it’s about Google alone, could do the trick.
Isn’t there a way to prevent google from linking images? Now I know that you don’t want to not be able to get googled but maybe you can prevent linking most of the images via a robot.txt or something. So if someone googles sandraandwoo in the image search they still get the first, let’s say, 15 comics and half a dozen of the more recent ones. But to actually read the comic and not only those 20 pages they have to visit the website.
I don’t host a website or anything so I’m probably at least partly wrong about what’s possible and what not.
So that’s what it does… By the strangest of incandescence when they switched over the new format i started using bing images search you know why … I really really hate the new search one Google put up I find it really inconvenient to use.
glade i did really I know these things cost money and if they get a little bit of ad revenue to help em out good on them.
My biggest complaint with the new image search has been the crappy lag I get on that page – I suspect poorly tested JavaScript.
The old version loaded the original page behind the image, so it counted as a page view for advertising purposes.
I suspect if Google gets a large number of complaints/threats from content authors, they will adjust the system before getting repeatedly blocked.