Six Side Hustles That Actually Worked This Year (And Three That Didn't)

I've been tracking nine builders for the last twelve months. Not students of mine. Not interview subjects. Friends and acquaintances who agreed to share a quarterly P&L in exchange for me not naming them. Here's what the data says.

The six that worked

Ranked by 12-month margin, not revenue.

1. **Verticalised B2B newsletter.** $40K MRR by month 18 (covered separately on this site). Margin >95%.

2. **WordPress maintenance services for restaurants in one US metro.** Boring. Profitable. $12K MRR off 38 clients at $279-$549/mo. The operator works 22 hours a week. Margin 78% after subcontracted dev help.

3. **Niche SaaS — software that schedules pickup runs for one specific type of equipment.** $7,400 MRR, 31 customers, $239/seat. Built in 4 months, took 14 months to find product-market fit. The PMF moment was three customers in a row asking for the same Stripe integration. Margin >90% at this scale.

4. **Etsy printables in a wedding-adjacent niche.** $3,800/mo on average, peaks at $7,200 in May/June. The operator runs ads at a 4.2x ROAS and reinvests aggressively. Margin around 60% after ad spend.

5. **YouTube Shorts adapted from a single boring B2B niche.** 411K subscribers in 9 months. AdSense ~$2,800/mo (Shorts pays poorly). The real revenue is the email list — 22,000 — and the lead-gen for a $497 cohort the operator runs twice a year. Total business: $190K in year one.

6. **Fiverr Pro illustration commissions in one art style.** $4,100/mo, mostly book covers. Specialised, highly defended by quality, hard to scale and the operator is fine with that. Time invested: 25 hours a week. Margin 92%.

The three that didn't

1. **Crypto trading bot.** Operator put $14K of capital and 6 months of dev time into a system that backtested beautifully and lost 11% in live trading. Closed it down in October. The honest version of this story is: backtests lie, fees and slippage are real, and the survivorship bias on crypto-bot operators you see online is brutal.

2. **YouTube faceless channel about productivity.** Saturated niche, generic content, lost a thumbnail war with five better-funded channels. 8,200 subscribers after a year. AdSense barely covers the editor's invoice. Operator pivoted in February.

3. **Print-on-demand t-shirts via TikTok.** Three viral designs, total revenue $11K, total ad spend and fulfilment cost $9,600. Profit on a year's work: $1,400. Operator describes it, accurately, as 'the most expensive lesson I've bought myself'.

What separates the two lists

Three things show up consistently. The hustles that worked were verticalised — narrow audience, specific pain, named customers. The hustles that failed were horizontal — broad market, generic content, anonymous demand. The successful operators could draw their customer in detail by month two. The failed operators couldn't tell you who their customer was at month twelve.

Second pattern: the ones that worked had real distribution before they had product. The newsletter operator built the LinkedIn audience first. The SaaS operator interviewed 40 prospects before writing code. The Shorts channel had a 22-day content calendar planned before the first video. The ones that failed assumed the algorithm would distribute for them.

Third: cost of acquisition was always knowable. Always. If you can't compute CAC by month three, the hustle is leisure dressed as business.

What I'd start in 2026 if I were starting again

A maintenance-services business in a vertical I had domain expertise in. $279-$549 month-over-month, ten customers in the first six months, twenty by month twelve. It is the most boring item on this list. It is also the thing with the cleanest path from zero to $10K MRR I have seen all year. The boredom is feature not bug.