Clicky

X

Subscribe to our newsletter

Get the State of Digital Newsletter
Join an elite group of marketers receiving the best content in their mailbox
* = required field
Daily Updates

Google Image Search traffic is ruining my SEO reports

22 August 2011 BY

In case you missed it, a few weeks ago Google seems to have made a quiet change to how Google Analytics records visits coming from Image Search results.

Whereas previously Image Search showed up in your Google Analytics as a referral visit – and thus allowed you to easily filter it from your reports – now it seems Image Search visits are part of the organic search traffic in your Google Analytics reports.

See for example these two graphs for one of our client sites:

Google Image Search traffic

Google Image Search traffic

Google organic search traffic

Google organic search traffic

As you can see the drop in referral traffic from Image Search coincides with a rise in Google organic search traffic.

Shouldn’t Google Image Search traffic be part of organic?

Now you may think that Image Search should be part of the organic search traffic measured in Google Analytics. But you’d be wrong. Because Image Search traffic doesn’t behave like regular search traffic. At all.

Image Search traffic has an incredibly high bounce rate. Users search for an image, look at it and perhaps save it to their harddrive, and then click the back button or close the window. Either way it is extremely rare for Image Search traffic to look beyond the page where the image is embedded. In fact most of the time, thanks to how Image Search works, users don’t even properly see the page where the image resides.

This makes Image Search traffic all but useless for the vast majority of websites. So the fact that it wasn’t part of your organic traffic reports in Google Analytics, and the ability to easily filter it from your reports, were very handy and much-used.

But now that Image Search traffic is part of the organic search report in Google Analytics, things have changed drastically. It totally pollutes organic search metrics, skewing the bounce rate, throwing keywords in there that have no real value for your website, and generally messing up your SEO reports.

And the real problem is that you can’t filter these image searches from your reports. Not without some serious Google Analytics wizardry. In fact, if anyone knows how to filter image searches from your organic search reports, please do let us know.

I have no idea why Google has chosen to pollute organic search traffic reports with this useless Image Search traffic, but one thing is sure: it’s having a severe negative impact on our ability to report accurately on our SEO efforts.

Now for every keyword Google Analytics reports organic search traffic for, I find myself wondering if this is actually useful traffic that ends up on a landing page and that we can potentially monetise, or if it’s utterly useless Image Search traffic that is nearly impossible to extract value from.

AUTHORED BY:
h

Barry Adams is one of the editors of State of Digital and is a freelance SEO consultant based in Belfast, delivering specialised SEO services to clients across Europe.
  • http://richardfergie.com Richard Fergie

    It looks like image search referrals will have an “&tbm=isch” parameter.

    You could try filtering on that

    • http://www.barryadams.co.uk/ Barry Adams

      Thanks Rich, will definitely check that out. Time to build *yet another* filter in to my Google Analytics profiles.

  • http://www.e-difference.nl/ Jeroen van Eck

    My guess is they are combining it into organic search traffic (because that’s what it is) now, to split it up into the new SEO reports for Google Analytics later: http://www.seroundtable.com/google-analytics-seo-13639.html. A strange choice to do this in phases instead of at once though.

    • http://www.barryadams.co.uk/ Barry Adams

      Yes that was what I was thinking as well. If so, this is horrendously bad timing – or a deliberate and cynical move from Google to make GA less useful for SEO reporting, thus making their SEO reports easier to sell. (Yes, I think Google will start charging for these specialised reports.)

      • http://www.jeroenvaneck.com Jeroen van Eck

        Do you have any reason to suspect that? I highly doubt they will charge for specialised reports. Haven’t seen any sign in that direction.

        • http://www.barryadams.co.uk/ Barry Adams

          There have been rumours for a while now:
          http://www.mediaweek.co.uk/news/1053142/Google-launch-paid-for-analytics-package/

          We’ll see. GA is not making Google any money right now, and I think they’re seriously considering a free basic version and a paid advanced version. Wouldn’t surprise me at all if such specialised SEO reports would be part of the latter.

          • http://www.basvandenbeld.com Bas van den Beld

            That would go against everything which Google has done in the past decade. I cannot imagine a move that drastic to come out of Larry Page his CEO-ship to be honest…

  • http://www.londonweddingdresses.com Karl Hindle

    I have been scratching my head over this very same issue – how to extract and identify the image traffic from “ordinary” search traffic. My principal adwords site enjoys around 1,500 hits a day and half is through images. For me, image traffic is very valuable because customers are searching for images first in most instances as they try to get an idea for what they want to wear (buying a wedding dress online is not unheard of but rare). The big issue for me is not how to identify and filter it out but how to get more, but I am finding image ranking to be an ephemeral process.

    Glad I’m not the only one going bald over this though ;)

    • http://www.barryadams.co.uk/ Barry Adams

      Karl, if your aim is to rank in Image Search, you might want to check out our guide to image search optimisation:
      http://www.stateofdigital.com/the-definitive-guide-to-image-search-optimisation/

      • http://www.londonweddingdresses.com Karl Hindle

        Thanks Barry – I’m optimizing images pretty much along the lines outlined and I’m enjoying the results, however they are just so “ephemeral” – much more than textual organic – yet for many of my clients (and myself) image search is so important. Thanks again for raising this subject matey – much appreciated.

  • http://www.burtonseo.co.uk adrian

    That graph and comparison figures dont make sense.

    Your image referrals were are around 90 per day at the most.
    Yet, when this apparent changed happened, your figures jump to 600 a day, from a previous 300. How do you account for the extra 200?

    This throws doubt onto your theory.

  • http://www.ppcni.com Jordan McClements

    If it is of any use to anyone – in relation to AdSense earnings – I have found in the past (although it’s a while since I checked an by the looks of things I’ll not be bothering:-) that non Image Search traffic is worth 3.4 times more than Image search traffic (at least if you use a javascript frame buster so people actually get to see the ads on your page)..

    So yes, Image search traffic is definitely worth less, but I don’t agree that it is totally useless, especially if you can get a lot of it…

  • http://fendorf.dk Jackie Frandsen

    Of course image search is part of SEO – it is SEO. You can optimize the images and our studies show that 40% do search for pictures. I even believe that optimizing your images make you rank better for the KW SEO.
    I’m sure that bounce rate are higher – the user search for a picture and find it. Why should I be here? Call it branding – just like a magazine.
    Pictures can of course convert, just make sure to have the right images for the right wording – let’s call this optimization ;-)
    Anyway nonpaid images are free traffic so the only issue I could find should be the tracking part.

  • Pingback: searchmetrics – nyheter | seo & sem

  • Pingback: Simsalaseo - Bildersuche SEO Contest - DoFollow.de

  • http://www.esterling.co.uk eSterling

    Really interesting article, you’re right that a Google Images search does not perform the same function as organic search, and should therefore not be regarded in the same way at all. I still agree with some of the other comments that it is worth optimising the titles of images though.

  • Pingback: 3 Google Analytics Changes That Will Affect Your Traffic

  • Pingback: » Pandia SEM Wrap-up September 4

  • http://www.wanshee.com/ robert

    Hello i have the same issue, this is killing my earning, my traffic and my website too! Have you found a solution>?

  • Pingback: Pandia SEM Wrap-up September 4 | WORDPRESS PORTAL

  • http://www.ppcni.com Jordan McClements

    So what is the (easiest) solution to segmenting your search and image search traffic with Google Analytics?

  • http://www.cogent-design.com Chris Hellowell

    Good article – thank you. My question – From other SEO sites I have looked at they all say they high bounce rate lowers page rank. ( not sure if that is true) We have an image on our site called tube map ( we are in London all you subway people) and it creates lots of visits and gives our contact us page a high bounce rate 69% . I am wondering if I should rename it to something less popular?

    • http://www.barryadams.co.uk/ Barry Adams

      Hi Chris,

      That bounce rate urban myth is exactly that – an urban myth. Google doesn’t look at a page’s bounce rate as a ranking signal.

      This myth likely originates from a misunderstanding of what Google does do, which is analysing click behaviour. If a SERP gets clicked on, then the user comes back to Google and clicks on another SERP, this is an indication that the page first clicked on does not contain what the user is looking for. And if that happens often enough, that may mean Google decides to rank that page lower in its results for the same query at some stage.

      As long as you make sure that your website delivers what users are finding it on, you won’t need to worry about bounce rates on individual pages – not for search engine rankings anyway. From a usability and monetisation point of view, high bounce rates are of course very much worth investigating.

  • http://www.everythingpeacock.com Ryan

    Sure enough, if you set “Google Property” up as a second dimension in the SEO > Queries report you can filter image searches out.

  • http://www.ppcni.com Jordan McClements

    So what is the (easiest) solution to segmenting your search and image search traffic with Google Analytics?

    Can anyone actually answer this?

    Do I *have* to modify the Analytics code on my web site?

    • http://www.barryadams.co.uk/ Barry Adams

      Jordan, this post on YouMoz might help: http://www.seomoz.org/ugc/tracking-google-shopping-traffic-with-google-analytics-14244

      It’s primarily aimed at segmenting Google Shopping traffic but it also tells you what to do for image search.

      • http://twitter.com/PPCNI Jordan McClements

         I *think* I implemented it correctly according to that article. I am getting some “Google Images / Organic” traffic showing, but it is definitely being *massively* under-reported. For last 30 days I got 25,000 “(direct) / (none)” visits and only 85 from Google Images (It should be the other way round – God how I would love 25,000 people a month to be going directly to my photo web site)…
        mmm…
        Strange.
         

  • Pingback: Is Google Image Search Worth a Marketing Look? - Content Muse, Anthony Pensabene Greenlane SEO

Watch our free webinar about blogging now!