BS: Duplicate Images = Duplicate Content


Since Bing and the Google started filtering "duplicate" images from the Image Search, there has been a steady rumor that Google will start penalizing similar images like it is penalizing similar text content now (well, not "penalizing" but let's say "filtering out"); both in- and cross-domain.

That would presumably mean, everyone who uses images from iStock or similar sites are going to be penalized...

Jill was pretty much clear when she was asked that question in comments:

Paul Schlegel said: I am now hearing rumors that Google has started penalizing sites with duplicate images, such as the one you've used in this post. Jill Whalen said: @Paul, that sounds kinda silly to me therefore I wouldn't believe it. Paul Schlegel said: OK. It was from a post by Roy Reyer - I don't know anything about him other than he did some trainings with Jerry West. I could be misinterpreting the post, too - or like you said, it could be nonsense. Let me know if you want a link to his article. Jill Whalen said: Found the article. He seems to be talking about product images on eCommerce sites. But I still think it's total BS.

But I heard very experienced SEO say "when it comes to Google, I will believe anything now".

Now, I have my own opinion, but I would love to hear yours!


I'll call bullshit

No, I don't see that idea having any basis in fact or even logic. If you've ever used Google Image search to try to identify an image, then you know that their image matching is as likely to return a puppy as a motorcycle.

And even if they could accurately match images, as you say, Ann, iStock images, as well as affiliate shopping sites, would be a major problem.

Just to clarify...

I am in no way trying to support this rumor. On the contrary, I needed a place (to link to) to bust the myth when someone emails me about that...


Ann, I really doubt you'll be able to find any sort of authoritative source addressing that question.l've certainly never seen anything on it.

The only time I've even seen anyone at Google speak in any detail about their image matching capability was someone saying that while the technology exists to do finite matching, it just doesn't scale for their purposes.

Good One

That's Good ONe!!




Makes perfect sense to me to filter them out because why should image scrapers get credit for the original author's image?

The argument about stock images from image banks is lame. You don't own it in the first place unless you bought exclusive rights to it. Besides, any slight modifications, including a simple resizing, may at this time be enough to get Google to think it's original based on what I've seen in the current image gallery so at a minimum I'd make everything a few pixels different than the stock version and perhaps slap a little watermark on it which would also differentiate it.

The only places where this gets really ugly is using stock images in very large ecommerce catalogs as everyone will get filtered out and the guy taking his own images will have a definite advantage over the rest.

The real problem I see is for people that used to use image search to catch people stealing your images, which I've done a few times, unless they provide an option to disable filtering and show all occurances found.


They might be able to scale detecting 100% duplicates across the entire interwebs — i.e. same size, same pixels, same compression algorithm yielding the same bitmap, but doing this if even _1_ pixel changes becomes more computationally expensive.  Resize an image and it's even more expensive.  It's just not worth it.  The algorithms to detect spun or lightly paraphrased copy might be scalable and mature, but this just isn't there.  Even detecting edges in robotics is pretty computationally expensive and difficult to parallelize as compared to, say, a massively parallelized inverted index lookup (the most basic part of Google).

So they don't want to do this for the entire internet, and I doubt it works all the time.  To use it as a ranking factor would be enormously expensive.

Bogus.  On top of that I don't see it as a valid ranking factor.  They might want to cluster away the duplication to some degree, but who's to say a larger and a smaller image don't serve a purpose for a searcher sometimes anyway?

So Ann, what's your opinion?

Now, I have my own opinion, but I would love to hear yours!



I started it with BS... I am pretty sure that rumor is lame, but apparently not everyone thinks so!

If Google ever had any

If Google ever had any contact with reality they've completely lost it again.

So I really wouldn't be surprised if the rumour proved true, with the caveat that sites with a certain "authority" would be able to escape the filter.

The result would be that Eric Schmidt gets to keep Google focused on brands, Google PR dismisses complaints by calling it "anti-spam" measure to "enhance user experience", and small businesses who lose out are told they should never be reliant on Google for anything. For ultra tin hat, in the long-run Google can then insert its own affiate codes into all organic redirects, including normal search as well as image search.

Simples. :)

You know you want to believe it's not true, but you also know it's more than possible. :)

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.