One of the major problems faced by the stock photo industry is “too many images.” Customers don’t need more images they need to be able to
find an image that works quickly.
More and more art directors dread the endless hours spent scrolling for the perfect image for their next creative project. Couple that with the pressure art directors are under to produce a greater and greater output of projects per day than a few years ago and many stock photo buyers are not happy campers.
Some art directors say they are being expected to produce as much as 10 times the number of completed projects as 8 to 10 years ago.
To offer some additional search options Adobe has introduced some AI-powered
aesthetic filters that will enable art directors to search more naturally and intuitively.
In addition to being able to choose a color pallet customers can now use a Vivid Colors slider to show only those images with certain prominent or dominant colors. There is also a Depth of Field slider that lets the customer view only those images with a lot of depth front to back, or images with a narrow range of focus.
New to Adobe's Sensei AI-powered search is a feature that lets designers
search for photos by using other photos. They can drag any image they’ve seen onto a target to find similar ones.
Organizing Search Returns
While all the above are useful additions to improving search some changes in some of the standard search options might also be helpful. For example, Adobe provides 5 different ways for customer to organize their search returns – Relevance, Most Recent, Popularity, Downloads or Undiscovered. But, it is unclear exactly what some of these terms mean.
What is “Relevance?” Has Adobe determined what is relevant for this particular customer based on the searches the customer has done in the past? What if the customer is new? What if the customer is looking for something totally different than they’ve ever searched for before? (Often the case.) Relevant to who? It it only relevant for Adobe? It would be helpful it they explained how they define Relevant, but I can’t find that anywhere.
It seems that “Most Recent” organizes all the images based on when they were added to the collection. That doesn’t necessarily mean when the image was created. It might have been created years earlier and just uploaded recently. We also have no idea how recent is recent. Was the image uploaded in the last week, 3 months, or the last year. At the rate they are adding new images all the first 500 images found use a particular search term might have been uploaded in the last week.
It would be helpful to have the option to specify a period of time (3 months, year) and be able to review just the images that have been uploaded in that time period. Adobe (or no one else as far as I can tell) lets customers do that.
The big mystery to me is the difference between “Popularity” and “Downloads?” Actual downloads are pretty obvious. Although the number of times is not revealed, presumably if an image has been downloaded 10 times it would be shown ahead of one that had been downloaded 9 times. We also have no idea how many times a given image has been downloaded. We do know that 7 or 8 years ago some of the images on microstock sites were being downloaded hundreds and thousands of times. That information might be helpful to the customer, but it is no longer supplied.
It would be helpful to know the number of images of a particular subject that had been downloaded 10 times or more, or 5 times or more. The search return delivers every image in the collection regardless of whether it was ever downloaded. There is no way to determine where the “Undiscovered” images start. If the customer doesn’t want an image that was too popular it might be helpful to know which images had been downloaded more than once, but fewer than 5 or 10 times. That information is not being supplied.
The “Undiscovered” category is useful if you’re looking for an image that has never been used. I did a search for “People” and found 9,090,864 photos (not counting illustrations). These images represent about 13% of Adobe’s total collection. About 2% of the 9,090,864 are unreleased images for editorial use only. The rest are model released. The interesting thing is that about 72% of the people images are Undiscovered. Presumably none of them have ever been used. Many of them have been in the collection for many years. I suspect the percentage of unused images in other less popular subject areas is even higher.
It would be very interesting to know how many of the 28% of “Discovered” images have been used in the last year or two. I suspect a very small percentage. Most of the ones that have been used recently have probably been used many times. They end up at the top of the search-return-order and get seen and used frequently. The rest die.
There is a great rush to get more images despite the fact that no one has looked at most of the images in the collection for years. How many of the images no one is seeing might be exactly the image a customer might want to use if the customer could only find it.
The changes Adobe has made are helpful. There is still much more they could do in the future.