I'm relaunching an old newsletter of 7,679 readers into Sunday Lab: 20 side hustle ideas every Sunday. Here's everything: what I built, the tools, the costs, the detours, and the rule that decides whether I keep it.
UpdatesThe story so farKeep or kill
New updates land here as the test runs: numbers, decisions and what I learned. Newest first.
Loading…
The story so far
I picked the test with the shortest path to money: one built on things I already own.
That's the whole idea behind a complementary build: you don't start from zero, you start next to something you already have. The question for test #01 is simple: can I relaunch this and actually make money from it this time?
From the first idea bot to the first email, in the order it happened.
| Date | What I built or decided |
|---|---|
| 21 Oct | Started automating the Startup Library Instagram: quote carousels approved in Telegram and posted automatically. |
| 22 Oct | A bot starts sending me new ideas every morning, pulled from 17 newsletters. 232 ideas so far. |
| 31 Oct | Picked the name Sunday Lab after checking about 80 domain names. Bought sundaylab.co and sidefit.co. One-page site live. Moved the list to Kit and cleaned it: 1,712 inactive readers out, 7,679 left. Set the goal: from $5k once to $5k a month. |
| 1 Nov | Idea library live. Paid reports per idea ($2.97) through Stripe. Signup tracking per link. Designed 7 ads (never ran them). |
| 2 Nov | Side Hustle Starter Kit v2 as the free signup gift. Simplified the idea bot. |
| 3 Nov | Free accounts with one-tap login from the email. Rewrote the homepage after a friend said it was unclear. New ideas now go live only when the Sunday email goes out. Rebranded a second Instagram page I own to Sunday Lab. |
| 4 Nov | Content machine for Sunday Lab: carousel templates, one-tap approval in Telegram, automatic posting and scoring. Chose ManyChat for comment-to-DM. Built a job-ad finder (side project). First carousel with a clear winner: "Never start a business until you know these 7 numbers". |
| 5 Nov | Planned the next tests (a car-rental price watcher). Meta blocked my app's access to post stats until the next evening. |
| 6 Nov | First subscriber stories and a feed post on both Instagram pages, each with its own tracking link. Split my personal take from each idea card. Sorted six Eden posts into a content playbook. Started testing cover colours on carousels. |
Most of it runs on free plans. The two I pay for monthly are the AI (Claude) and the content research (Eden).
| Tool | What it does for me | Why this one | Plan |
|---|---|---|---|
| Claude | Builds everything with me: the site, the bots, the database, these words. Its API writes the idea cards, reports and draft posts. | I describe what I want in plain language and it builds it. I'm not a full-time developer. | Paid monthly |
| Supabase | The database and the engine room: 78 tables, 32 scheduled jobs, about 28 small server functions. | Database, scheduled jobs and server code in one place, with a generous free plan. | Free |
| Vercel | Hosts this site and the tool that turns slide designs into images. I bought the domains here too. | Deploys in seconds, and the domain and DNS sit in the same place. | Free |
| Kit | Sends the newsletter and holds the 7,679 readers. | Free up to 10,000 subscribers, and good deliverability. | Free |
| Eden | Shows posts that beat their creator's usual reach, so I copy formats that work, not words. | Saves me from guessing what content works. | Paid monthly |
| ManyChat | Someone comments a keyword, they get a link by DM. | Already approved by Meta, so I didn't need to go through Meta's app review myself. | Free |
| Telegram | My control panel: I approve ideas, posts and reports from my phone. | It's where I already am all day. Works from the beach. | Free |
| Stripe | Takes payment for the $2.97 idea reports. | Simple payment links, no shop needed. | Fee per sale |
| Resend | Emails buyers their report. | Simple to send email from code. | Free |
| Forward Email | Forwards my sundaylab.co address to my normal inbox. | A professional address without paying for a mailbox. | Free |
One-off: $59.98, half of it wasted. Monthly: $129, plus about $17 of AI usage so far. Revenue from real buyers: $0.
| Item | One-off | Per month | Note |
|---|---|---|---|
| sundaylab.co domain | $29.99 | Renews yearly | |
| sidefit.co domain | $29.99 | Not needed yet. I bought it on day one for a later tool: an interactive dashboard to find an idea and work on it. That can live on sundaylab.co too. | |
| Claude (Max plan) | $100 | The subscription I build everything with | |
| Eden | $29 | ||
| AI usage beyond the credits | About $17 so far | Mostly spent while building, not while running | |
| Instagram shoutouts | $0 | Posted on my own pages | |
| Two test purchases of a $2.97 report | Both by me, to test the checkout. Not revenue. | ||
| Everything else | $0 | $0 | Free plans: Supabase, Vercel, Kit, ManyChat, Resend, Telegram |
My challenge budget is $5,200 for 52 weeks. The two domains are booked against it, so $5,140.02 is left. The monthly tools aren't in that budget, because I'd pay for them anyway. Counting them and the AI usage, the first month costs about $206.
Each one felt like progress. None of them was the email. (The lesson I took from it.)
None of this is wasted. The content machine already posts, the idea library is the signup magnet, and the checkout will sell reports one day. But I built it in the wrong order. Step one was always the same: send the email, see who opens it, ask what they want.
Free channels first, paid only once I know what a subscriber is worth. Every link has its own tracking code, so I can see which one actually brings in signups.
A ladder from free to paid. The newsletter stays free; each step up only gets built once the step below proves people want it.
| Step | What you get | Price | Status |
|---|---|---|---|
| 1 | The newsletter, a free account and the Starter Kit | Free | Live |
| 2 | A full report on one idea from the library | $2.97 | Live, no real buyers yet |
| 3 | A playbook with real numbers from one of my tests | $27–47 | After the first test with results |
| 4 | A replication kit: everything to rebuild a tested idea in your own market | $97–197 | After a test proves itself |
| 5 | Builds of the tools I use myself, starting with the content machine, for small businesses | $100–500 a month, or a one-off build | First paid build maybe soon, through my own network |
| 6 | A tested side project, sold ready-made | Around €2,500 | When one works but doesn't fit us |
Every test gets a rule before it starts, so I can't talk myself into it afterwards.
I'm not judging this one on sales yet. I'm judging it on whether people find it useful. By 22 November, two weeks after the first email:
Hit them: I keep going and start testing what people will pay for. Miss them: I pause and rethink, out loud, in the newsletter and in the updates above.
Every Sunday: 20 side hustle ideas, plus what happened in the lab that week.
Get the Sunday email →