So I’m going to combine two very good articles I have read about Medium and work out my thoughts:
“There Will Be Money” by Edward Lichty and
“Images: All the Numbers” by Mike Sall.
The Importance of Images
Mike Sall in his post mentioned that he couldn’t answer the question on how Images affect the TTR (Total Time Reading) but he got to some pretty good conclusions:
1-If someone is willing to take the time to actually upload an image, they are more likely to be the type of person to write a post that performs well.
2-If they upload a really nice big image, they are even more likely to be that type of person.
3-If they upload lots of really nice big images and play around with different layouts and sizes, then they are that much more likely to be that type of person
So he concludes something which most of us would guess on our instinct but he has some data to show it, and this is that a story will do better on its TTR when it includes images. In his words:
“When more work seems to have gone into the images, that tends to correlate with more TTR. But this is likely a function of the storytelling efforts of the author, not the images specifically.”
Discover
We all know search is broken due to the amount of info we have added to the web. Scrolling through thousands of photo and images on the web is inefficient.
We have broken the feeling of discovering and we need to help users get it back.
Not everyone has on their computer that perfect image that their mind is creating while they write. And it’s so frustrating to search on the web for that image. (here is a pro tip from Medium staff)
E.g: If my title mentions Yosemite, the photos I should be offered would be from Yosemite National Park. As I write more (for example about snow) the algorithm should learn and offer more precise images when I want to add a second one.
Tricky note: the first picture is the most important one. If there is no title or written words how can you know what to offer? I haven’t found a solution to this that I would like to share (yet).
Carrousel of photos
Imagine if instead of having to search your stock of photos or the web for an image you would be given a carrousel of photos you can choose from.
You would ask what are the concepts behind this idea. Why would someone want to be able to choose an image? Why would Medium be interested in adding this feature for the authors? Let’s start by explaining the concepts of this feature.
Core concepts on the design
1- Medium Stock should be added as an option for the author so as not to create friction with the user but encourage discovery.
2- Looks of the Carrousel: it should respect medium etiquette and design. Beautiful and simple movement of images as if they were just floating in front of your eyes.
3-Personal algorithm: The algorithm should read from your words as you write and understand the main topic you are talking about. Because the save option is so powerful the algorithm will become more personalized and targeted as the author develops the story.
Build a stock
To be able to create the Carrousel we will need a huge stock of images. The good thing is that we have where to start! Here are some of the possible contributors to our data base:
1-Free images (here) (get API). See other free images stocks we could use.
2-Professional Photographers (for their own publicity — distribution) how should we give credit — automatic credit instead of friction of looking for who it is from. If medium is medium.
3-Magazines with good stock of pictures: the credit would be mandatory helping as free publicity and distribution of name. They will also be associated with the Medium brand.
Exchange of credit + distribution + association in exchange for plug-in to their database.
5-Users: there are some power users that are also photographers or who just love to take pictures that will be more than willing to contribute (I’m one of them, I have plenty of quasi-good photos that I could share).
6-Curation: so there are several filters or stages of curation depending on the root or origin of the picture. In my honest opinion the better the origin and filters, the better the image, the more the user will like and use it.
Curation levels: magazines (or photographers) > ranking or value or awards > algorithm > individual option = decision from the user.
Users as curators of images could empower the community around Medium.
Monetization
In this section we will mention some channels through which we believe the Carrousel and its images can be monetized and what this implies.
1- Branded content on the minimal unit of the product
Marriott — Gone it’s simply an amazing collection of stories with beautiful images. It just has it all: images, words, the story.
Great and beautiful idea but difficult to scale.
The normal writer on Medium wouldn’t be able to create such a beautiful story without any help. They (we) need and want help to create our stories.
When I mentioned the “Minimal unit of product” I’m referring to images. It can’t be words or phrases (maybe quotes) or paragraphs due to the fact that you can’t sell this or you would be a ghost writer for your own users. This is not Medium’s vision.
What if in the Carrousel we could find astonishing images that would be provided by brands (companies) to use by the writer for free. The only mention of the brand would be on the credits.
2- Psychological effect of images and subliminal messages on the users.
The idea is that there could be two kinds of sponsor images or ads (there actually could be many more but we are going to talk about two)
One is when you actually show the product on the image. These are the most problematic because there needs to be a direct connection between the image and the product and the article. Creating this three part connection is not easy.
We could go into a very deep analysis here but it’s not the point. We are just going to explain that there are three connections here:
Picture > Words / Brand > Picture / Brand > Words
The first one goes from the picture to the words. This is the most common one and the one people use without the need of any help. It’s the feeling that the picture expresses the idea of feelings in the words written or complements them (this is actually a larger topic in itself).
A branded picture (brands > picture) would be the connection between the actual image and what it’s describing and how this relates to the image. To give an example, the beer Coors makes a commercial where some climbers bring cold beer from a mountain. They are making a subliminal statement and giving a message: the beer is better cold, and it’s usually as cold as the top of a mountain can be, and it’s cool because climbers go all the way to get it just for you. So if I was writing a post about my experience in Nepal maybe I would use a beautiful image of the Himalayas. The problem is that if there is a beer on it I wouldn’t connect with it. But if it’s just the name of the beer on one corner then maybe I wouldn’t care because I like the quality of the image.
This example also explains the third connection. Brands to words. Some images could have the beer itself on the picture (example if I’m talking about going out to bars in Chicago) but they should be subtle enough so it doesn’t create friction in my article.
3-How to monetize?
So part of how to monetize we just covered with the example of Coors. But would it actually work? Well, the company obviously could pay $0.001 cents for every time someone sees the image on their carrousel. This monetization is based on the fact that users are already having contact with the brand. If they actually choose the picture and use it there should be an extra $0.001 because it will be public and distributed.
There is one more layer that I will just mention and it’s in case that post or story goes viral or makes it to the top list, so in that case maybe there should be a plus.
As all this would be “sold” with use of metrics (TTR), there should be a clear understanding of how the metric is related to the subliminal message communicated to the readers. Time spent on the post, but also time stopped on the image and numbers of readers.
Some other considerations (that are very important)
4-Don’t spam (overwhelmed)
One important thing that we can’t forget is that it needs to be clear that the pictures are ads. Let’s start first with ads that are spam.
This is actually something of concern. They can’t be visual spam. If I get 10 images to choose there can’t be 9 that are ads. There should be 1 at the most. This is a quantity filter.
In quality you can’t spam or “force” the user to use the branded images because he is writing about Thailand and there are 9 pictures of kittens and one of the Marriott in Thailand. NO. Or you can’t give 9 bad options of pictures and then the only one clear or actually good/acceptable one is the one that is branded. NO.
I believe its possible to create Ads (images) that are beautiful and that we would be happy to use.
4- It’s brand content!
There’s need to be a clear distinction on which image is branded. Obviously when a user selects an image we should automatically write the “rights” of the image and who it belongs to (credits). When showed on the carrousel it will need to have a yellow border around it that tells you clearly that this image is branded.
Overall concept: give options (with new feature) to simplify the creation of quality articles by providing excellent images.
Help people express themselves through the words and images that fill their lives everyday.
PS: If you got all the way here I want to thank you and tell you I really appreciate your time.
