Is Knowing Which Images Get The Most Downloads Enough?

I have been told that there are at least 650 million images available for easy licensing in various databases around the world. This doesn’t include all the images that can be found by searching Google, Bing and the social media web sites where other images can be found, but not easily licensed.

I believe there were about 240 million unique licenses of images worldwide in 2015. A significant number of the same images were licensed multiple times so the actual number of unique images licensed was much less than 240 million.

Of that 240 million about 185 million were licensed via a subscription. I’ll explain later why subscription downloads should not be considered in the same way as when a customer pays a separate fee for each image downloaded. These numbers leave a maximum of 55 million images downloaded where the customer was confident enough that the specific image was the right one for the customer’s project that they were willing to pay a specific fee for the right to use it.



In the case of many of these 55 million downloads, the image chosen was the same as had been downloaded by many other people. My guess is that the actual number of unique images downloaded in 2015 was less than 25 million. Thus, less than 4% of all the 650 million were individual paid downloads, not via a subscription. This figure is probably high because I know that some major agencies license use of less than 1% of the images in their collection annually.

Thus, a huge percentage of the images in today’s stock photo collections are never used by anyone. Being able to distinguish the few that are actually being used from all the rest is of huge importance to image producers.

Knowing What’s Selling  

Knowing what’s selling would be tremendously helpful, but it isn’t anywhere near enough to be an effective guide for production. Given the wide range of pricing for various uses the gross revenue an image has generated in its lifetime is an important factor.

If the average cost to produce an image is $25 and the average amount you earn from an image in your collection is $2.00 or $3.00 per year with the useful life of an image being two years, that’s not a very good business investment. Of course, there may be other non-monetary value in producing an image that cannot be measured in dollars such as joy, happiness and recreation.



(One simple way to determine an average cost of producing an image is to total all your business expenses for a year and divide that number by the number of new images you made available for marketing by actually placing them with stock agencies during the year. Also, don’t forget to consider your time invested and what else you might have done with that time.)

While having some idea of the images in any given category that generated the most revenue for the producer would be of immense value for photographers planning new shoots, the simple fact of just knowing if an image has sold at all – and being able to distinguish sellers from non-sellers – would still be a major improvement over shooting in the dark.
 
Shutterstock allows customers to organize searches based on “Popular,” iStock on “Most Popular” and AdobeStock on “Popularity” and “Downloads.” With AdobeStock the customers get different results depending on whether they organize search returns by “Popularity” or “Downloads,” but it is unclear why. In any event, to be useful, the organization of returns needs to go much further. None of the traditional agencies offer any information about the images that are most popular, or have been downloaded most frequently.

There needs to be a way to look at the most popular in the last year compared to all time popularity and maybe 2 years and 5 years. It may be useful to list the most popular in the last month or 6 months for subjects where styles and trends change rapidly. This would be a very simple algorithm to build.

Popularity should be based, not just on number of times an image has been downloaded, but on gross revenue generated. This is particularly true for images that may have been in the collection for a long time and were downloaded frequently at much lower prices than are charged today. It wouldn’t be necessary to reveal exactly how much revenue an image has generated in its lifetime, but simply to know that the first image shown had generated more revenue during the period of search (1 year, 2 years, etc.) than the second image shown, and so on for all images shown.

In a search for most popular images, only images that have sold at least once should be shown. It should then be possible to toggle to see all the additional images in the collection that had never sold.
 
iStock used to let you organize “Most Popular” by number of downloads and tell you how many times each image had been downloaded. Now you have no idea how many times a given image has been downloaded and whether they are still delivered based on the most downloaded first.

Subscription Downloads

When trying to determine the popularity of an image based on number of downloads or revenue generated, there is a big question as to whether subscription downloads should be counted at all.

The problem with subscriptions is that many customers download huge numbers of images for reference purposes that they never actually use, or intend to use, in a project. Since it costs them no more to download extra images they tend to download a lot. No one knows how many of the images downloaded via subscription actually get used in a project.

Counting such downloads equally with those where the customer must pay an additional fee for each image downloaded can lead to some grossly incorrect assumptions of which images are really popular.

Consider the following. Suppose a photographer has produced a shoot that has 6 to 10 slight variations or angles of the same situation. Ten subscription customers download 8 of the images with the intent of playing around with them on their desktop and later deciding which single image to actually use. The single image customers come in, review all the images offered and purchase just one each that they intend to use. No one knows how many images downloaded via subscriptions actually end up getting used.

Out of this group 20 images will probably be used, but 90 images have been purchased. Which of the 10 variations is most popular? The single image customers might have all chosen the same image from the sequence, but unless you carefully look at single image sales no one will be aware of that fact.

On top of this the photographer who sees 90 sales of this situation, with every image getting used to some degree, will think this was a very popular situation, that he should duplicate on his next shoot. In fact, only 20 people used an image from this situation and the majority may have chosen to use one particular shot. The photographer has no idea which particular angle was most useful for customers and which ones eventually got rejected. The photographer really has no idea as to where to focus his efforts on a future shoot.    

If the revenue generated from each download is worked into the calculation that might solve some of the problem. For example, the average revenue from a single Shutterstock subscription download is about $1.25. If the Sutterstock customer pays for each single image downloaded the cost per images is about $10.00 or higher.

In my opinion it would be more useful to image creators if subscription downloads left out of the calculation entirely and only the downloads where the customer pays an additional fee for each image downloaded were used to consider popularity.