12 - Ike

When my client, Josie Lifts Things, posted a video meeting people who live in the Parisian catacombs, it underperformed. The initial thumbnail was constructed from a bunch of stock photos and Generative AI to follow a pre-determined formula - one that was popularized by other creators, such as Tyler Oliveira.

The video doing poorly was a big let-down, as Josie had stayed in Paris for a long time to try and get access to this secret society. She had a very unique video, but the thumbnail was a barrier stopping her from reaching her audience.

In her email to me, she wrote: "My channel is based on interesting cultures and people around the world. Realism is very important to me, I do not want cartoon, childlike looking thumbnails as my content is based on authenticity."

My favorite part of thumbnail design is usually coming up with unique concepts. I actually don't look at my clients' competition, I don't try to copy what others do on the platform. I have a gut feeling about what could work as a thumbnail for a video, and I'm fortunate to have clients who let me experiment with them. I've had some great successes, and some disappointing flops, but that's what you get when you try to predict the human mind! I use YouTube daily and watch hours of content, so I've built up a catalogue of concepts in my mind that I can apply to all kinds of designs. It's fun to test brand new things on YouTube and see if it catches on. Honorary mention for my second favorite thing: putting a Camera RAW Filter on the finished thumbnail and making it pop 10% more. Chef's kiss.

I'm thinking about it. It's not something I've offered before, because I have fun taking those concepts out of my head and onto the canvas. I have the vision for my own stuff, in a way that's sometimes hard to explain to others, you know? Giving my ideas for someone else to execute on would be a new step, but I'm not opposed to it - I just gotta figure out the logistics, pricing and all that. If you want to hire me for concepts on your channel, hit me up, we'll work something out.

The ideal client is and always has been someone who will take custom high quality asset photos for me and play along with my unique thumbnail ideas. Some of my best work came from equally incredible photos taken by my clients, who perfectly match the sketch I've provided them and allowed me to put my concept to the test exactly as I'd imagined it. Bonus points if the video performs well! Ah, and clients who pay my rates ahead of time, with no attempts at negotiating. There is nothing quite as empowering as a client who responds to a price with "Sure! When can we start?" - it proves to me that they understand the value of the work I make.

I'm personally a designer that uses very minimal AI in my process. I prefer not to rely on these tools for my work, and practice my skills independently from them. As such, the only use of AI in this thumbnail I've shown today is to extend the wall on the right of the image and to add a hi-vis yellow vest on Josie using Photoshop's built-in Generative Fill. By not using AI in my work, I preserve a much stronger sense of realism, which is what got me clients like Josie in the first place: people who want to tell real human stories don't want their content to be represented by made-up places and people. I advise designers to only rely on GenAI as a last-ditch resort for their assets, rather than as a crutch for their style, as these technologies are unreliable and ever-changing. Working on your fundamental skills will never be a bad idea.

It's not easy being a newcomer in the current landscape, I'd imagine. I was already an established thumbnail designer by the time Generative AI became better at making images, so I didn't have to compete with it, but rather choose to adopt it or not in my workflow. If you're a beginner, you're directly competing with services like Midjourney for the thumbnails of smaller channels and clients with lower budgets... I'd say it's honestly not a big loss. Focus on improving your own skills and make thumbnails for fun, as a hobby. Ask YouTubers you like for some assets to practice with, make some designs for their channel and don't demand anything of them. Don't ask to be hired, just post your work online and keep getting better. If you're playing the social media game right, you'll get the attention of people who won't try to replace you with a machine. That's when you can actually turn this into a full-time job: when your clients respect your expertise, when you're more than a tool to them. In short, practice for fun and post your work online. Slowly build up your portfolio and you'll get clients passively. Don't run after low-budget jobs.