Google liking incidental photos on my site unrelated to main topic

[MikeNoLastName]

Okay, this has been bugging me for months now. According to the WMT "Search queries" report (with the WEB filter on) my top query (with 2500 impressions in a time period) is for an incidental PHOTO linked on one of my pages of, lets say, a "fuzzy puppy". So I get 2500 impressions for the term "fuzzy puppy" (which has next to nothing to do with my, let's say, "gardening" site), with an avg position of 3 (confirmed photo result under web search in this case) and a click thru ratio of 0%!

The second highest query (under text search) is right-on for a secondary key search term, but only gets 2000 impressions in the same time period with an avg position of 4.3 and a CTR of 20%, followed by other similar terms. Adwords stats show it should be getting far higher impressions than this.

If I include ALL searches (web and image) I now get similarly unrelated photos getting as much as 4500 impressions in the period with avg position of 22-30 and CTR of <2%.

It seems I am getting a lot of exposure but for the wrong terms (primarily related to photos). So I get visitors, but none are interested in the primary topic of my site, so high bounce rate and few ad clickthrus. G has pretty much turned off 95% of my previous on-topic keywords since the early days of Panda.

My question is, has anyone had this situation and tried just getting rid of the misfounded query topics by say deleting the photo (in my case maybe 3 or 4 biggies and 5-10 lower but substantial ones out of thousands) and seen an increase in impressions for their main keywords? Or should I assume that ANY exposure is good exposure and keep them all? Are these few 'odd' query terms possibly skewing G from realizing the true subject of my site, or are they just a free side traffic benefit? Why does G keep sending me 1000's of impressions for a very minor term which no one clicks on, while not trying others that they more than likely would click?

I would be interested in hearing others' actual experiences.
.

[edited by: Robert_Charlton at 7:07 pm (utc) on Nov 10, 2012]
[edit reason] fixed formatting - added spaces between paragraphs [/edit]

[TheMadScientist]

Hmmm ... That's 2500 impressions, some clicks, some site exposure, some 'site/brand recognition', etc. for free you wouldn't normally get ... I don't know how many people search for 'fuzzy puppies' and are also involved in whatever type of 'widgeting' your site happens to be about, but personally, I'd not worry about 'correcting' the free exposure you're getting, because most people I know have a variety of interests and free exposure is free exposure.

What I'd do instead is work on getting the lost rankings back for the other keywords, but I'd definitely be leaving well enough alone with the photo that's giving some free exposure and clicks while I was doing it.

IOW: Think billboard ... Not everyone who sees them is a buyer today, but tomorrow? Maybe...

[lucy24]

You have to fine-tooth-comb all text within, say, half a mile of the image tag to figure out why on earth the search engine thinks your page about leaf-cutter ants has something to do with teddy bears. And bend over backward to devise a boring-but-accurate alt for each of those images.

Frankly most of the time I just look at the numbers, say wtf and ignore it. Though it is exasperating when you constrain your gwt results to "web" and it still shows jpgs among the results. You're left wondering what it is about your utterly unrelated image that made g### think it's closer to the search term than several million other people's full pages. Or possibly even one of your own pages-- but not the one the image lives on. (And what is the point of setting up a separate Image search if you're still going to throw images at people in the regular search? But that's another thread.)

[MikeNoLastName]

Good point billboard MadScientist. You've convinced me. Although actual clickthru is rather small compared to the impressions according to WMT. And since it is a photo totally unrelated to the site... Maybe I could add our website URL REAL BIG across the image... :)

Exactly lucy24. Though I thought G was a little smarter to eventually stop showing the link when no one is clicking on it. They seem to stop showing GOOD keyword links right away when the CTR drops. Must not have gotten there with images yet.

[TheMadScientist]

Maybe I could add our website URL REAL BIG across the image... :)

I Like It!

(Okay, maybe not real big, but I really like the idea of getting it on there for people to see.)