Google Images: Am I actually getting these hits?

[brokaddr]

I deleted a large amount of images recently. All of a sudden, I'm seeing a huge amount of traffic to them - I don't believe all of these are hotlinked.

The way Google & Bing's image search functions now, I'm curious if, when I actually see a referral from *images* - do I actually get a visit from a person, or is this simply a pre-load on the image results?

When I was searching recently, I noticed in the status bar, my url and a huge amount of competing url's loading... so the images a "pre-loaded" by my understanding.

If that's the case, then when the pages pointing to these images existed, that would appear as a 'hit' to the page, correct?

If so, this is greatly thwarting statistics, is there any way to decipher actual traffic vs someone viewing a hotlinkined image on Google or Bing?

[brokaddr]

After analyzing the activity of users for nearly 12 hours straight, I see some alarming results.

It appears that roughly 57% of the site's traffic for April to-date are 0-second visits. (I calculated this using only users with a referrer - those without I did not include in the calculation)

This particular site -used to- rank extremely well and has taken a nosedive as of late. I wonder if these 0 second visits have anything to do with it.

To troubleshoot:
- Checked the site in multiple computers/platforms
- Did a "Fetch as Googlebot" from WMT - sourcecode is clean of any unwanted links/javascript
- Some users interact with the site as expected (complete task, etc), so if something was broken, I would expect no users could complete the task
- If a user signs up for the task and does not complete, I send a followup mail to inquire about possible difficulty, so far, nothing out of the norm has been reported by any respondents
- Checked a few keywords, ranking is spot 3-6avg positioning on the first page for the tested keywords

Since the site is functioning fine for some users, albeit at a much lower percentage than in the site's ranking heyday, I am left puzzled and scratching my head.

Most puzzling of all: Should I factor the 0-second visits into the overall site stats, the traffic is normal (compared to the ranking heyday), which would be for both hits/visits from unique users.

Is there anything else I could do to try and pinpoint what is going on?

[lucy24]

Hey, I know the answer to this one. Well, part of it.

When the new image search first came out, I did some careful experimenting-- step by step through Image Search while concurrently asking for unrelated pages in a different browser so I could track each step in logs.

When a person using Image Search clicks on an image and gets the lightbox thing, google loads up your image and you see the request in logs-- but the human searcher doesn't see the image. They see a slightly scruffy jpg that google has already prepared and cached for the purpose.

If the user now proceeds to the next step, clicking on something within the lightbox, they then see your actual image.

You can save bandwidth by rewriting to a single-pix gif when the referer is "blank.html". Set its expiration to right away, and if the searcher really does take the next step, they'll get your real image.

[tantalus]

The Se's are free loaders taking you for a ride - a free one, I might add, at your expense.

Bing is worse than Google, sending visitors to your image url with no possibility of browsing through your site.

It really is time for webmasters to be highly selective about what they give to the engines.

I've banned Google and Bing from indexing my image folders a few weeks ago, and went further with Bing requesting they remove all my images from their index.

I am now watching my bottom line intently to see if it has had any effect whatsover. So far none, but it is eary days.

I don't have time at the moment, to see if there is a way to exploit image search for my site's benefit, but I suggest you read lucys thread about the topic, here: [webmasterworld.com...]

I'm curious too, if throttling is taking place, whether those image hits count towards your cap? Because I know what traffic I would prefer, but I would bow to other people's expertise here whether in fact this is or could be the case.

[brokaddr]

None of this sounds desirable.

I thought the image search was an inconvenience as a surfer (I found, I get less spammy results for an image search than regular search - but with this new format, it's hard as hell to get to the page that sourced that image! Often times I give up.)

As a webmaster, this is not good at all.

You can save bandwidth by rewriting to a single-pix gif when the referer is "blank.html". Set its expiration to right away, and if the searcher really does take the next step, they'll get your real image.

Is this in the thread tantalus linked? I've bookmarked it for a detailed read later tonight.
However, a large percentage of the regular traffic at the moment is referer-less, I'm only able to decipher this type of traffic from legit visits by the 0-second visit length.
So rewriting based on referer only solves a small percentage of the overall problem.

tantalus - Are you anticipating a rankings impact (both organic AND/or images)? I get some decent visits from images (both bing/google) - but nowadays that's few & far between, but they still trickle in. Guess a lot of people also have a hard time accessing the source page, too.

I really don't want to remove my images from Google & Bing, my images rank well on both engines - but if they are giving phantom traffic the rankings are essentially useless. My images aren't the type worth stealing, but rather, they grab attention and bring the user for the surrounding content.

I guess this is a ploy by Bing & Google to keep people on their site longer.

[brokaddr]

What a read that topic was, tantalus - thanks for the link!

As it stands, there doesn't appear to be any foolproof solution to this, outside of snipping those with referrers (ironically, back when my traffic was at it's peak, I blocked all referrers from Google Images/Yahoo Images/Bing Images, etc due to the scraper issue that is still rampant to this day). Traffic slowed slightly, so I unblocked. I almost wonder if I'd shot myself in the foot.


I am looking at my users currently active on the site through session tracking.
Most do not have the http://www.google.com/imgres?imgurl=http://www.* referrer - but some do.

I hope I'm wrong here - but 100% the phantom traffic active right now appears as though they're viewing the PAGE the image is displayed on. (ie. Simulated page view, without actually clicking the page)
It only shows up as the jpg/gif/png being requested, once that page is deleted and serves a 404 (both the image + page housing the image)

I just tried to recreate this myself, I was unable to - so I'm wondering if something else is at play.
Perhaps my issue is more related to this topic: [webmasterworld.com...]

[Str82u]

I don't rely on G images for any traffic but read this just now about sites losing lots of traffic to the new design [searchengineland.com...]

[brokaddr]

That's interesting Google is blaming phantom traffic for webmasters' drop in traffic. Did they talk to any ecommerce site owners? Surely, those folk would be able to decipher a phantom visit from a legit visitor through their sales statistics!

[tedster]

That's interesting Google is blaming phantom traffic

But they are not the one's who are saying that - rather it's the article writer's best guess on what Google might say.