I don't know if anyone covered that subject but here is my few cents. In stock photography we are familiar with the term "keywording the images". Pretty hard task to do. Sometimes even hard to figure ten descent keywords. With time we become skilled, we learn how to use synonyms, dictionaries, we research through other images, we learn how to answer on who, what, where, when, why etc. And at the moment we can say that we are skilled in keywording our images. But they don't sell! Why?
Good keywords are one of the main factors to give exposure to our image. If they don't sell, there is something wrong. Lets see how can we make some diagnostics on what happens.
Every image, aside other data has two important values: views and downloads. Each of them are saying something to us. If the "views" number is wrong it means either our keywords are wrong or to weak, or customers doesn't search for that words because no one needs images that subject. If "views" are high but "downloads" are low, then probably keywords are good but image quality or realization is bad, customers don't like it. We can improve the image on the same keywords which shows great interest and try to shoot several more shoots on this subject. This leads me to the title of this article - what about imaging the keywords instead of opposite?
If you have weak portfolio (just like me) you don't have enough data to get some experience about what sells and what doesn't. The solution is to browse the DT image database to see what is going on.
First interesting data is on front page. You can see latest downloads, and latest searches there. Unfortunately we don't have any aggregate data or search cloud to point which keywords are popular. But, latest searches can give an idea what people are looking for.
Next, You can do some test by searching the database yourself for a certain keywords. If your search returns thousands of images it is probably not worth go for this subject as it is overloaded. My habit is that I always sort the search results by "downloads - descending" so I can see the most downloaded images first. Most downloaded images shows what people buy. Look at the subject and write down few keywords, at least most important ones. For example, try to search on "car crash". You will find less then 40 images on that phrase. I am getting to the point.
I ask myself, can I create an image on "car crash" (for example) that will met the DT criteria. Not much downloads on those images, but for start it may make me happy.
I am intensively browsing DT everyday and I have descent list of keywords that obviously sells. I will try to do some photos on that subjects. I know that it will not make me rich, but it is nice to see what competitors did and try to do the same or even better. As an inexperienced beginner I spent lots of time on images that didn't pass, and those which passed are not so interesting to the nation.
Conclusion: If you are sure you can do great shot but you just don't have idea, do research on DT. You are looking for subjects that sells well but doesn't have too much images on the subject. Collect your keywords and do some descent job imaging them. Does this make sense?