Shutterstock investors often ask my opinion of stock photo industry’s future and the potential for Shutterstock’s growth. I tell them growth will slow significantly. Demand from customers willing to pay for the images they use will decline. Shutterstock has grabbed about all the customers they can from Getty so there is not much potential for growth there. Adobe will take a much bigger share of the market. Recently an investor asked me, “What would you do if you were Shutterstock?” Here’s what I told him.
Meeting Customer Needs
Picture buyers don’t need more images. There are already too many images. They need help in finding the best images for their purposes. They need collections that provide a more targeted selection of the kind of images they regularly use. They need to be able to review the entire collection quickly, identify the best image for their immediate need and get on with the rest of their day.
Easy to say, but hard to do because every buyer has different, unique needs. But they want smaller, tightly curated collections and are turning toward them when they can find them.
It is unlikely that there will be a significant growth in new customers. Thus, in order to grow revenue, sellers must find a way to charge more for at least some of the images they license. The good news is that many of Shutterstock’s best customers are happily paying more than Shutterstock charges when they can find the right image quickly.
Step 1 - Identifying In-Demand Subjects
I would identify the specific keywords that are used most frequently by Enterprise and Image-on-Demand customers to find images they actually purchase. I want to know the final set of words used that produced a set of images from which one is chosen and purchased for actual use.
I would not include subscription searches in this database. Subscription customers tend to download lots of images they never use. It costs them nothing more to download extra images. They grab a bunch of images and then work with them on their desktop to determine which is the right image for their particular project. Nobody, not Shutterstock, not anyone else, knows how many of the images downloaded actually get used in a project. About 90% of Shutterstock downloads are via subscriptions. There are reasons to believe that no more than one-in-six to one-in-ten of the images are actually used. Raw data alone can sometimes lead to incorrect conclusions.
Step 2 - Curated Collections
Next, I would build curated collections of the best images that can be found within Shutterstock’s 94 million image collection for each of the most in-demand set of keywords. Each collection should have a maximum of 500 images.
If there are more than 500 great images then an experienced editor should choose the 500 best giving weight to those most frequently downloaded by Enterprise or IOD customers. This editor would regularly review new images added to the main collection and determine if a new image should replace one in the existing collection of 500.
A second option for dealing with excess images would be to create a second collection, based on the same set of keywords, but to try to identify some additional sub-category that would distinguish the two collections.
To the degree possible the editors should be assigned categories of imagery to research such as: food, education, active seniors, recreation, wildlife, nature etc. It is unlikely that any single person will have the subject area knowledge necessary to adequately edit all types of imagery.
In some cases, the same image might appear in more then one curated collection. There could be an unlimited number of curated collections. There would be separate curated collections of photos, vectors and illustrations.
To understand the need for such curated collections, consider the number of returns Shutterstock delivers with the following searches.
Office, computer |
2,423,530 |
Office, computer, hands |
136,101 |
Office, meeting |
459,926 |
Office, meeting, multi-ethnic |
21,194 |
Office, meeting, computer |
104,760 |
Office, meeting, presentation |
80,739 |
|
|
Education children |
523,718 |
Education university |
679,124 |
Education teachers |
212,242 |
Education children classroom |
74,657 |
Education children recreation |
11,419 |
Education science |
394,567 |
Education children science |
21,092 |
Education concepts |
1,022,350 |
Education technology |
467,602 |
With such searches are the most appropriate images for any given customer included in the first 500 images delivered? Is there a better image further down? How far down? Typically, customers seldom look at more than 500 returns from a search before changing the search parameters.
It is worth noting that there is almost the same number of “Education Kids” images (523,699) as “Education children.” Do customers use the keyword “children” more than they use “kids” or do they use them about the same. If they use them about the same, and they use both a lot, then it may be useful to list the exact same collection under two different titles.
It is also useful to consider that in the “Education teachers” category above a lot of the first images shown in the search return have no teacher in the picture. This type of thing happens more and more as the collections get larger. Image creators attach every remotely applicable keyword to the images they submit in the hope that everyone looking for an image related to the large general category of the subject matter will see their image. This can lead to a lot of inappropriate images in the search return relative to what customers are actually seeking.
Step 3 - Pricing
Images in the curated collections would be offered at a higher price point. In addition, curated collection images would no longer be available to subscription customers.
There are a couple ways the pricing could be dealt with.
Shutterstock’s Image Pack prices could remain the same, but the cost of any image found in one of the curated collections could be equal to purchasing 2 images in the non-curated collection. This would be simple and very consistent with Shutterstock’s current pricing strategy.
I would prefer a system similar to Stocksy’s with 4 different prices based on file size. This allows web users who only need a small file to get the images they need for less that print users who need a higher resolution.?
However, the different file size model would also require some type of credit system pricing that would charge more for larger file sizes. For example, each credit could be worth $5 and the prices could be 3 credits for a Small file, 6 credits for Medium, 15 credits for Large and 25 for X-Large. These prices are adjustable and discounts could be offered when the customer purchases a larger package of credits.
Shutterstock already offers its higher priced curated collection Offset. But for a significant number of today’s users, I believe the prices for these images are too high. In addition, the curation strategy of Offset does not offer the number or variety of premium images in narrow subject categories that I believe my strategy would supply.
Step 4 - Image Search
When a customer enters a set of keywords the return will indicate the number of curated collections that have images that meet the criteria. Add more words to the search string and fewer curated collections will be shown. Simultaneously, the return would begin showing all the images that are not in curated collections.
If a customer wants to review the curated collections, she would click on that number and a list of all the collections available would be shown. Each collection would be defined by a specific set of keywords and all images in each collection would tightly relate to all the words in the collection title.
For example, the keywords entered might include combinations of the words like: Office, Computer, Hands, Meeting, Multi-Ethnic or Presentations. If it appears from the research that there is significant demand for any grouping of these words separate curated collections could be created for each grouping. It probably will turn out that the keyword “Office” alone is too broad to justify creating a curated collection of office pictures. Most likely customers regularly use additional words to narrow the focus of what they are seeking. The trick will be finding the most useful combinations of words that aid the customers without being so narrow that customers seldom use the words.
Technology can help a great deal in finding the right words, but a successful strategy will rely heavily on human editing skills.
Some images may be placed in several curated collections when it is deemed by the editors that the image is one of the 500 most likely to fulfill the needs of a customer using the keyword title of the image set. (The “children” and “kids” discussion above may be an example.)
In an effort to achieve maximum variety, similar images from a particular shoot will not be included in any particular curated collection. The editor would be charged with picking the best image from a shoot unless there are two substantially different images that meet the “best” requirement.
Customers will also be able to use the
Shutterstock Reverse Image Search to find similar images. This search will show images in other curated collections, as well as those in the general collection. This tool can help customers in at least two different ways.
(1) If the customer likes the general concept of an image, but would like to see different variations with a single reverse image search the customer can easily find all those variations.
(2) Subscription customers who want to keep their costs low may find that it is easier to start their searches in the curated collections. In that way they may find something they like quicker. Then they can use the reverse search to find similars that are available for download as part of a subscription package.
Downside
The principle downside to this strategy is that it will cost money. While technology can be very helpful in identifying the most requested subjects and the most downloaded images, in the final analysis a significant number of human editors with market knowledge would be required for this strategy to really be effective.