As image collections have grown, finding the right image quickly has become an increasingly difficult problem for customers.
Back when the industry delivered 35mm transparencies, customers could call a picture agency, describe their need in detail and a researcher would delve through hundreds of images to find the few best. These selects were then shipped to the customer for final consideration. Internet search has changed all that.
While some customers like the more hands-on control provided by Internet search, many overworked art directors in today's pressure cooker environment want assistance in narrowing their search.
Agencies have tried to solve this problem by offering digital research services similar to those they provided with film. Customers call, describe their need and a researcher is assigned to go through the online collection and create a lightbox, which is emailed to the customer. Usually, a single researcher is assigned to a particular request. This can result in overlooking an image the customer would have liked to use because the researcher didn't like it or think it was appropriate. In addition, many customers have a problem with the lag time between making the request and getting images.
Better keywording has been offered as a possible solution, but when keywords are used to narrowly define a need, there is the risk that the specific keyword may not have been attached to a very appropriate image. Thus, that image will not be found. In the days of plastic sheets of images in file drawers, categories tended to be broad.
Given the need and the problems, some in the industry have given customers the option to order the images in the search return based on the number of times they have sold or been downloaded. Thus, customers can benefit from the research done by other professionals, or in Web 2.0 terminology, "the crowd." It's unclear which agency first introduced this system, but it was probably iStockphoto.
A disadvantage of using "search by download" exclusively is that new, quality images tend to get buried because they have not been downloaded previously. Thus, this system of ordering search returns should be one of several options available.
Traditional Sellers Ignore Benefits
The strangest aspect of using the crowd to supply research services is that traditional sellers have not adopted the idea. It is purely a microstock phenomenon. Traditionals acknowledge they have a serious problem with too many images in any given category. No customer has time to look at them all. But they refuse to give the customer the option of organizing the search return order to bring best sellers to the top.
Perhaps traditional sellers are embarrassed to let customers know how few times their images have been licensed, but they could still organize them by the number of times purchased and not include the actual numbers, as most of the microstock sites do.
Another possible option is to organize the images by the number of times a preview has been downloaded. The big problem with this strategy is that it could be easily "gamed" by photographers, who download their own previews many times. However, the problem could be diffused by using an algorithm that compares downloads with actual sales. Images with lots of preview downloads, but few or no actual sales, would be pushed down in the search return order, thus discouraging gaming.
As I pointed out earlier, "search by download" cannot be used exclusively. Customers must be given options. Different customers have different needs. But if the object is to serve customers microstock, sellers have a much better search strategy than traditionals.