Let me introduce you to a side-project I’ve been heads down on my own for the past 6 months. Tbh finished the MVP 2-3 months ago but hadn’t touched it since, so I thought might as well write about it now to keep the context alive.
What is Social St
I’m creating a social web app to host and join all sorts of activities around the area. Similar to meetup.com, eventbrite, radius.to, lu.ma (this I just recently discovered, and it looks so good 🫢). But unlike any, I’d like Social St to capture interactions that are less structural and more spontaneous like:
- finding companions to get coffee
- going out on a walk, co-working
- a table tennis partner
- boba picnic at the park
- grab dinner
The opportunities to meet people are endless! Social St empowers anyone to host any type of activities at any time during the day, or join one in their community.
Demo
Learn more @ www.socialst.club/learn
The beta @ www.socialst.club/beta
Feel free to play around with the app and create your own activity
Costs
One of the most important lessons I had through the experience was knowing how to spin up a service with least 💰. Running Redis on Fly.io is $0.20 per 100k requests, but I was shocked to see my bills at $20 from only letting Sidekiq run idle. It appears the heartbeat Sidekiq makes to Redis every 10 sec plus the job polls increases the request count by A LOT. I think the solution here is to use Solid queue / Delayed Job which uses MySQL instead of Redis. (a TODO)
Another source of cost is DB. I thought Railway’s trial credits would last me a long time, but turns out it only took 2 months for it to run out. I would need to pay $5/month for hobby plan. At that point, I would just buy a Digital Ocean VM for $6/month and run a MySQL server there. And I did!
The domain name is another fixed cost - $5 first year and $25 per year thereafter.
Last but most definitely not the least, my time and energy is another cost 🥹.
I knew for sure I needed a way to monetize this to keep it up alive. But first I must validate the idea. Who would pay for this?
It’s time to validate the idea + get feedbacks
I figure that a good way to validate my idea was to see if anyone would pay for it. So I went ahead and build a subscription plan called ‘Guild St’ and try to pre-sell it. My idea is that Guild St is a plan that allows hosts to create private / exclusive communities, custom badges that’ll spark community engagement, and get paid as hosts.
This is just an idea and obviously didn’t bother building it unless I can find a customer for it - which is something I probably should do early on before building the whole MVP this far 🙂↕️. But the MVP is here, and I’ll take any type of donations / fundings for those who’d want to see Social St come to life. 👀🧎♂️➡️
That’s the stage I’m at. A newbie tech sales trying to land my first customer 👨💼🧳.
I could also start collecting feedbacks. I’ve tried posting on Indiehackers but unfortunately saw no responses. Will perhaps try out Reddit, Product Hunt, or maybe targetting certain clubs / hobby groups. I could perhaps dogfood my own app by helping to promote other existing events / clubs.
What could I have done better
I do not regret spending time to make this - I learned tons of stuff (can’t wait to share them in a separate post)! But I had to admit that it wasn’t the best, and wouldn’t recommend it for anyone. If I were to redo it, perhaps I should begin by pitching my idea and collecting feedback regularly before attempting to create an MVP, so that I wouldn’t just work in a vacumm and over-deliver.
Building it for months on my own
Building it is a pretty lonely journey, but I had much fun. A few things really helped me:
- Trello board: it kept me focused on the current task, 1 step at a time, while separating the good-to-haves from the essentials.
- Rails! 💪
- Finding time everyday. Small progress >> no progress. Try to keep tasks small and easy so I don’t get too bored stuck on one.
Below are notable features / learnings while building:
- Setting up Rails / Tailwind / Stimulus / Redis / Sidekiq stack
- Deploying to Fly.io with Docker
- Desiging the UI. Ensuring it looks good in mobile.
- Using Device gem to setup Authentication routes, and Google OAUTH for Sign-in.
- Pundit gem to authorize which user can perform an action (ie. only hosts can update an activity)
- Designing DB models to represent relationships (ie. Activity, Memberships, Users)
- Tagging system with acts-as-taggable-on gem, and slim-select library for multi-select UI and allowing users to dynamically add new activity types. You can search activities by tag.
- sidekiq-cron gem to setup reminder and activity status update scheduler.
- Dealing with Timezones: both client-side (inferred from browser) and server-side (user settings)
- Broadcast realtime notification (activity venue changes / someone joined or leaves)
- Infinite scroll with turboframe
- List joined / hosted activities and live stats
- Participants can rate a host after the activity completes. Average rating is calculated on write.
- mapkick.rb gem to display interactive maps
- Google’s Place Autocomplete API to validate user address input.
- Filter activities by location and distance range. A button to get the current device location, and using reverse geocoding API to turn the latitude/longitude into a place name.
- Share to socials buttons, attaching an invitation image of current activity as the OpenGraph preview. Generating the image dynamically requires a headless chrome like Puppeteer and deploying it through Docker was a challenge 💦
- Creating the landing page
- Launching Digital Ocean VM to self-host MySQL server