In the “
Goodbye Shutterstock” thread on
MicrostockGroup marthamarks said, “My older images still sell on Shutterstock, but newer ones die there.” Why would that be? One would expect newer images to sell better, particularly when agencies continue to ask for more and more images.
This does not seem to be insolated complaint, but one common to many long time Shutterstock contributors.
The reason older images continue to sell better than newer ones is because “search return order” (SRO) is based to a large extent on number of downloads and number of views.
Older images were downloaded a lot when there were fewer images to compete against them. If we go back just to Q2 2014 Shutterstock had 38.8 million images in its collection and 31.5 million downloads in that quarter. Go back to Q2 2011 and the ratio was even better (14.4 million DL and 15.3 million in the collection.) Historic download totals stick with the image. The new images have no downloads so they end up lower in the SRO.
Sure, the SRO is not totally based on downloads; there is a mix of new images with older images. But agencies keep a significant percentage of the images customers have used before near the top of the SRO because it has been proven that those are what, at least, some customers want to use.
Thus, if the new images don’t get downloaded quickly, and in significant numbers, they don’t have much chance of staying near enough the top of the search where they can be seen – particularly, if they are of commonly used subject matter.
The next factor is the number of images that customers tend to review before changing search parameters. It has been reported that most customers won’t review more than 500 images before they change their search parameters, no matter how many images were returned with the original search.
So consider. Suppose Shutterstock shows one new image for every image that has been previously downloaded; a 50/50 split. (My bet is they put more weight on images that have been downloaded, but they are not telling us how the algorithm works.) So, if someone searches for a keyword attached to an image just added to the collection, it has a chance of being seen, if it is included among the top 250 with that keyword.
But with Shutterstock uploading 800,000 new images a week, how many of those will have the same keywords? How long will an un-downloaded image stay within the top 250 -- a week, a day, a couple hours? It may be a great image, but if it isn’t exactly what one of the customers searching in the next few days wants it will quickly get so low in the SRO that it is never seen again.
Maybe the photographer shoots subject matter that is seldom requested, or uses keyword that are seldom requested. In those cases, the image may be found for a longer period of time. But, relying on obscure keywords that are only used by a small percentage of customers doesn’t solve the problem because the majority of customers still won’t see the image.
One way for photographers to determine if this might be part of their problem is to do searches for the keywords attached to most of their images. Note the number of returns. If there are tens-of-thousands of images with the same keyword or sets of keywords the odds that an image that has never been downloaded will be shown in the first 500 are slim.
Of course, if the customer uses several of the words the photographer attached to the image the odds get better, but the more words the customer has to attach the greater the chance that one or more of the words she uses will not be one of the photographer’s words and that will take her off in a different direction with a different set of images.
This is why sales via Offset and Stocky tend to be much better on a per-image basis. Not only is the revenue per image used much higher, but regardless of the keywords the customer chooses there is a very good chance the customer will actually see your image because there are fewer images of that subject matter in the collection.
As Shutterstock has loaded up new images in the last year or two the odds of an image being seen have changed dramatically. See Chart.
| |
Q2 2014 |
Q2 2015 |
Q2 2016 |
| Downloads |
31,500,000 |
35,900,000 |
43,400,000 |
| Images in Collection |
38,800,000 |
57,200,000 |
92,100,000 |
| Gross Revenue |
$80,200,000 |
$104,400,000 |
$124,400,000 |
| |
|
|
|
| Downloads Per Image in Collection |
0.81 |
0.63 |
0.47 |
| |
|
|
|
| Revenue Per Download |
$2.55 |
$2.91 |
$2.87 |
| |
|
|
|
| Revenue Per Image in Collection |
$2.07 |
$1.83 |
$1.35 |
Between the second quarter of 2014 and Q2 2016 the number of downloads per quarter increased 38%. But the number of images in the collection increase by 137%. The decline in downloads pre-image in the collection has been particularly dramatic in the last 12 months.
Shutterstock reports “Revenue per Download” quarterly and that improves as they sell more single images and make more Enterprise sales. But for photographers the “Revenue Per Image In the Collection” is a much more important number. That has declined from $2.07 in Q2 2014 to $1.35 in Q2 2016, a 35% decline in 2 years and a 26% decline in just the last 12 months.
For many photographers trying to beat the odds is a losing battle.
How Could Agencies Help Their Contributors?
Considering the overwhelming amounts of data agencies collect, it would seem that those interested in better servicing their customers would be able to supply photographers with image-by-image data of how frequently each specific image in a photographer’s collection has been downloaded and the gross revenue generated in the last 12 months and all time. With such information photographers could make more educated decisions about how to focus their time and energies when producing new images. For most agencies it would be a simple matter to create a simple database of each photographer’s imagery that showed a thumbnail of each image in the photographer’s collection with the following information beside each of those images.
| Internal Image ID |
|
|
|
| Original Photographer File Name |
|
Gross Royalty |
Gross Royalty |
| |
Downloads |
Previous 12 Months. |
Career |
| Credit downloads |
|
|
|
| Subscription DL |
|
|
|
| Single DL (fixed price) |
|
|
|
If these records could be indexed by the photographer based on total revenue generated by each image, it could be very helpful to those photographers hoping to support themselves from the revenue their images generate.
It would also be helpful to know which images haven’t been viewed in the last 12 to 24 months. Maybe the photographer could do something else with those images to generate some income, or, at the very least, not produce more of the same subject matter.