I have seen a slight increase in sales for new images - but it might not be statistically valid, maybe I am just noticing it. I don't think I have changed my keywording strategy much if any.

true, no relatively recent images have seemed to get any long term traction, although I guess that would really take long term to see.

I've seen new uploads go to the very first position of the 'popular' search without getting downloads, maybe that's why you are getting some sales on new files.

How do you see this?

"Absolutely Bang On" Pretty much covers my recent experience as well. Some quick sales and then zip. I did have one kind of catch on, gets a DL a week now.

Since I ignored the OPs part about keywords, no I don't count them and no, I don't think the number has anything to do with it, unless of course they are lacking, which isn't a number things, it's "all the appropriate keywords that are necessary to properly describe the image"

Could too many be a problem/ In theory, yes, because of what the agencies say about relevance. I don't have any personal proof or observations that would be scientific or have and validity. So I don't know.

Here's what I know: "all the appropriate keywords that are necessary to properly describe the image" is the right number. 
 

Not unlike a cheap firework they pop and dazzle for a few moments and then fizzle away in the gloom.

I'd like to see a few at least get some real traction, just in case any of my bestsellers start to fade.

I have a few images which have between 40 and 50 keywords and they sold on the first day(s).

So there goes the theory

Ok that is good feedback. Maybe a tweak to their search. As of a few weeks ago NONE of my new stuff sold right away, only months later.