There was talk a month back that Google were beginning a test of running graphical advertisements next to image search results.
Link - http://www.searchengineworld.com/google-search/3457669.htm
I haven't seen any evidence of this in AU - but it made me think ... if Google DID choose to roll this out what sort of revenue COULD they achieve.
Well - lets consider a $3cpm for the purposes of this hypothetical.
Google Image Search has 2.6m AU users - which makes it a top 20 site usage wise.
These 2.6m AU users on average visit 3.18 times per month, viewing 59 pages. So, collectively, they account for 153,896,000 page impressions. (to put this in perspective the ninemsn homepage does approx. 80m page impressions)
So - based on a 40% sell through (lets be realistic here), they would sell ads on 61,558,000 pages - which at a $3cpm works out at monthly revenue of $184,674. Not bad but I doubt worth Google's time.
But lets say within 18-24 months with better technology in terms of matching ads to search queries the CPM could increase to around $10 - and the sell through to around 80%
Then monthly revenues would be around $1.23m. Over the course of a year this works out at $14.77m.
These figures are far from definitive - but they are based on real usage data and realistic cpms. However, with minimal info on the ad model it's hard to go beyond anecdotal predictions.
The potential is there. I guess the key is getting data on what images people are searching for and whether there's advertiser demand around these areas. It could definitely work for commerce but would it be relevant to the areas that seem to be driving industry growth - finance, auto, IT, real estate etc? Plus, does Google want to experiment with something that isn't broken for what is a relatively a small amount of incremental revenue?