Search Engine Optimization for Photos

SEO for photography websites is a big problem.  Search engines index text.  Images have no text except for any metadata which may be attached.  The camera provides a lot of this, but no one is likely to be searching by F stop or shutter speed, so this data has limited value. 

Stock photography websites optimize their home pages and advertising for terms like "Stock Photography" and then present visitors with good indexes and keyword search into their picture database.  This works for them, but an individual who wants to sell a few of his pictures online isn't going to be able to build a website that ranks for a term like stock photography.

That leaves "Long Tail" searches as a potential opportunity.  However, in the long tail of photography searches the photography entrepreneur is going to have to find ways to get site(s) ranked for these search terms.  That requires a way to quickly and cheaply add relevant metadata to the pictures. 

Our curiosity about this process has lead us to try a couple of things without a lot of success.  Picasa, the wonderful, free, image management program from Google has great keyword search ability.  If you create meaningful names, write captions, or enter keywords the program will quickly search through thousands of photographs to find the ones you need.  Unfortunately, not all this information is included when you export the picture for use on a website, so the search problem remains. 

We tried typing descriptions to include with the pictures but that gets old really fast.  We typically take a lot of scenery photos and may have five or six of the same tree or hillside.  It is very hard, and very boring to do that without repeating yourself.  It also takes more time than we have. 

We are going to see if talking to our computer is a viable approach.  Dragon Naturally Speaking is a program that converts speech to text.  We have read some good reviews and decided it was worth a try.  However, in the typical software upgrade manner, we will have to upgrade our computers to run the program. We hope to test it thoroughly during the first quarter of 2007.