Will more images in an agency collection grow revenue? Is more choice always better?
Shutterstock is adding 1,608,350 new images a week to its collection. That’s 229,764 new images a day. The average customer reviews 500 or fewer returns before changing search parameters. So how do they make it possible for customers to review all those new images?
Consider a few searches:
| |
Image Returns |
| Barbecue |
1,184,479 |
| Barbecue party |
153,188 |
| Family cookout |
3,220 |
| Soccer |
1,158,849 |
| American Football |
145,006 |
| New York City Skyline |
154,773 |
| Woman in office, computer |
696,016 |
| Family recreation |
214,412 |
| High school sports |
29,221 |
There is no human editing of accepted images. The theory is that customers will do the editing for them. With technology and artificial intelligence, the agency is able to track what customers see and pass over, as well as what they purchase. Then they are able to show the next customers what others have liked.
But, at some point that no longer works. The agency doesn’t want to just show images that have been previously purchased in the first 500 returns because it wouldn’t be long before there would be enough images selected at least once that new images would never be seen.
On the other hand, they don’t want to show just the newest images that have never been purchased because none of the first 500 might be what the customer wants and they must assume that most customers will quit looking if they haven’t found what they want after reviewing 500 or fewer images.
Thus, if the purpose is to maximize sales, they will probably want a mixture of images that are brand new and have never been used along with older images that have been good sellers.
Let’s suppose that within the first 500 images shown in any keyword search at least 250 have been used at least once and 250 are images that have never been used. The first customer to do a search chooses one of those that has been previously used by someone else, or she passes over everything and doesn’t find anything she wants to use.
Presumably, the next searcher will be shown a different group of unused images. With 229,764 new images a day to go through there as a lot of other new images that need to be shown at least once before they re-show to a second customeran image that was shown before and rejected. It would take 919 searches to get through all of one day’s new images at least once.
The next big question is how many times do you allow an image to be passed over before it is no longer within the first 500 shown. Shutterstock says they currently have six searches a second so they could go through 919 searches in about 2.5 minutes. Thus, it would be possible for the same image to be passed over 576 times in one day.
Obviously, not all the 279,764 will have all the same keywords. Some will be images of very popular subjects and others will be images that are seldom requested. Let’s assume they allow a new image to remain somewhere with the top 500 search returns and be passed over 250 times before it is pushed below 500 in the search return in order. Thus, new images of a “barbecue” or “soccer” might stay high enough to be seen for a little over 12 hours if one of the first 250 people to look at it doesn’t license its use.
Other images of subjects that are requested less frequently might stay where they can be seen for a few days, or a few weeks. But, of course these are of subjects that are not in very high demand and the subject is not purchased frequently.
If you still think more is better, consider these comparison figures between October 2015 and October 2018.
| |
2015 - October |
2018 - October |
| Shutterstock |
63.7 million images |
233 million images |
| |
38.1 million DL |
43.9 million DL |
| |
69% downloads |
18% downloads |
| |
|
|
| Adobe |
40 million images |
100 million images |
Between October 2015 and October 2018 the number of images in the Shutterstock collection jumped from
63.7 million to 233 million. Downloads during the same quarters only grew from 38.1 to 43.9. So, if each image was downloaded only once and no image downloaded multiple times 69% of the images in the collection were downloaded in 2015 and only 18% in 2018. Of course, we know that many of the images used were downloaded multiple times during the quarter so a much smaller percentage of unique images were downloaded in each case.
Adobe’s collection grew at a much slower pace. We don’t know how many downloads they had in either case. However, we do know that in 2018 Shutterstock’s overall revenue only grew 12% while the revenue of
Adobe Stock grew 25%.