Thinking about Thumbnails


I had an idea on Wednesday, and thought I'd share how it came to be, and how I decided what would be in the MVP. Keeping it simple enough to be able to launch it tomorrow.


YouTube has been a popular platform for all kinds of users for a very long time. What started as a simple video sharing site is now home to media empires of folks like MrBeast and PewDiePie, but also for the old guard of media.

I've tried being a YouTuber, you can see my "One Hour SaaS" series from a few years ago, but find myself being more comfortable with the written word.

I do enjoy watching other people's videos, and have learned a lot over the years, from what has turned into a very valuable resource.

As the number of YouTubers has grown, there has been an increased attention to details for their channels.

I have wanted to build a tool for YouTubers for a while, but struggled to find an idea that would be a good fit for me.

For example, there are tools that use generative AI to create content for YouTube channels. Both the content and the thumbnails. As someone who doesn't really post a lot, knowing what is good generated content would be hard.

Similarly, there were a number of tools built that allowed you to test different thumbnails for your video. Without having a YouTube channel that gets a lot of views, it's hard to build and test something like this.

Partnering with existing YouTubers would be a good way to move forward, but that I've never quite managed to make a connection with the right person in the past.

Building tools like this can be risky, for example, YouTube recently built a way to do thumbnail A/B testing directly into their platform.

I still see the value in a tool like this, so I started to think...what if there was a way to test thumbnails before you published your video. This requires having access to people that are interested in YouTube videos. What might be a good way to get that kind of audience on a website or list.

What if there was a way for people to vote on which video they would watch based on just seeing a thumbnail. So the idea for thumbnail.vote was hatched.

MVP Features

Remember long term, we want to allow users to test thumbnails before they publish them on their own channels, but the in the short term, we are trying to build an audience of people interested in YouTube videos, that would be able to rate thumbnails.

We start by building a database of good, and potentially bad, thumbnails from existing videos published on YouTube. Allow people to tells us which ones are better. Allow people to see the results of the voting, so they can use that information when creating their own thumbnails.

Later, we have a few options. Train an AI with the results we've gathered to test new thumbnails. Or perhaps even better, we allow users to add thumbnails before they use them on YouTube, and let other users tells us which are better.

There are 3 parts to the MVP.

  • Collect existing video thumbnails from YouTube
  • Show existing thumbnails to users, and let them select with is better
  • Show the results for each thumbnail

That is the core features, we'll do things like allow people to subscribe to a mailing list, etc., but I don't count that as part of the core features.

I got most of this built on Thursday, and will GTFOL tomorrow.

Once that is done and the majority of the functionality is automated, I'll start thinking about how to attract people to the site to vote on the thumbnails...already have a few ideas for this.

Cheers,
Mubs

Sideproject MVP

Subscribe for tips and tricks for building your side project, from someone who's built over 100.

Read more from Sideproject MVP

Mubs A recent tweet by Pieter Levels reminded me of the importance of being a jack of all trades when working on side projects. It's not enough to be able to good build software, there is a lot more to building a good project than the quality of the software, in fact that's probably pretty low on the list. Wrote my thoughts on being a jack of all trades:Mastering the Art of Being a Jack of All Trades Continuing my learning for SEO skills, I recently completed a build challenge run by the team...

Mubs With over 120 side projects, a question I often get is: "How do ship so much?" Keeping on track with projects can be difficult when after the demands of a day job and personal commitments. I've shared some of the ways I stay on track with my projects in the most recent blog post: How to keep shipping? I'm focusing on SEO as a way to market (and grow) projects this year, and was excited to see that Danny Postma will be publishing a course on SEO soon. I've already learned a lot from...

Mubs This week, I've been thinking about who I'm building for. Yes, I'm building a side project for my own reasons (networking, revenue, whatever) but who is the actual project for? What is the Ideal Customer Profile for your project, and why is it helpful to think about that right when you get started. I dig a little deeper into the concept, and how I worked through the process for a previous project: Defining your ICP Featured Side Project I wanted to take a minute to highlight a side...