I’ve worked on countless side projects over the years. Most of them never saw the light of day, but they were still great learning experiences, while others failed miserably. Only a few of the ones that I completed are what I’d call successful. And it really depends on how you define success.
Over this time, I’ve gathered some useful tips on how to pick the right project to work on.
For me, the process starts before you have even thought about project ideas. The first step is to answer the question ‘Why are you building it?’.
Let’s say you are building an AI-powered research assistant using the OpenAI Assistants API and proprietary information.
Are you looking to learn a new tool in the process? Do you plan to release this research tool to the public? Is it going to be paid? Are you looking to potentially turn it into a business? Or are you building it for fun?
Those are all very important questions that you must ask yourself before moving forward.
This is when you can start brainstorming ideas if you don’t have one yet. List everything you want to eventually build and start ranking the ideas. How much does each idea your goal of building? If you want to create a paid micro-SaaS, which idea is best for this monetization strategy? Or if you want to learn new tools, which idea will get you working with the most technologies you haven’t worked with yet?
For me, building things is always about scratching my own itch. I love tinkering with software and building small apps. The fact that we now live in the age of AI and no-code builders is great for me, as I can now build small web apps spending a few hours here and there. Compare that to 10 years ago when it took much longer to build simple apps, and you see how powerful no-code is for tinkers like me.
Next step would be to set realistic expectations for your side project. If it’s a side project, then you probably have work or school to think of before spending time working on anything else. Think about a realistic timeline, one that you can adhere to throughout the project. Write down the major milestones and the days/weeks it will take to finish each one. Then double those times. Trust me, double them, and you’ll be much closer to what will probably happen in reality.
Okay, so you have your perfect idea and your realistic timeline. Now commit to them. Do not compare yourself to other makers out there and do not strive to finish it faster at the expense of sleep and time for social interactions. It’s not really worth it.
Next, brainstorm all the features you need to build. In the earlier example, this might look like:
Compilation of proprietary information to use
Prompt for the Assistants API (write, test, iterate)
Web interface with sign-up/login
Landing page
Page to interact with assistant
Page to review previous interactions
Option to regenerate responses
Option to ask follow-up questions
Extracting summaries from conversations with the assistant
Sharing responses and summaries
Chatbot integration with the assistant
Option to provide feedback on generated responses
Zapier/Make integration
Integration with social media (e.g., responding to chat messages, etc.)
This list can be quite long, I’ve only illustrated the notion above.
Now comes the most painful step. Look through the list and mark the barebones functionalities that have to be there for the project to work and perform its core feature. Here’s what that looks like:
Compilation of proprietary information to use
Prompt for the Assistants API (write, test, iterate)
Web interface with sign-up/login
Landing pagePage to interact with assistant
Page to review previous interactions
Option to regenerate responsesOption to ask follow-up questionsExtracting summaries from conversations with the assistantSharing responses and summariesChatbot integration with the assistantOption to provide feedback on generated responsesZapier/Make integrationIntegration with social media (e.g., responding to chat messages, etc.)
Strike everything else and put it in a drawer for some other day when your project has been successful, and you want to expand it. In rare occasions you may pick more than one core feature, but in my experience going over 3 core features to start with can be daunting for a solo-builder.
Roll up your sleeves and get to work. Try to get people to try it out ASAP and adjust the functionality, UX and UI accordingly. However, be very careful not to completely rehaul your project based on every single interaction with a user.
Always remember why you are building it!
So, pick the right idea for your current circumstances and goals, stick with it, and remember to have fun while you’re at it. At the end of the day, it’s probably not going to pay your bills, so apart from knowledge, fun is the most important thing you can get out of it.
Thanks for reading this far and for sharing the newsletter with your friends and colleagues. It’s the main way I grow my audience and am able to reach more people who might find value in what I have to share.
Till next week.
Best,
Dobri 🍃