Every publicly-visible idea on Starter Story's problems database, scored through your constraints — non-coder, minimal capital, real-estate/commercial edge, a demanding day job — then compared across the three markets you'd realistically operate in. The winners are not the ones with the biggest revenue. They're the ones you could start in weeks with almost no cash.
Four things the raw list won't tell you, and they matter more than any single idea below.
You have capital access, a network, and an oversight role that already fills the week. Every idea here dies the same death: you don't give it 15–20 focused hours. That's the real filter, so I weighted bandwidth and buildability heavily.
Most entries are horizontal SaaS that need an engineering team and years. For a non-coder, the honest wedge is productized services you can sell now and maybe turn into a tool later — not SaaS from scratch.
Bill in USD, deliver from Hanoi. The same "get-found" or advisory work earns roughly 3–5× more from US/expat clients than VN SMEs, while your cost base stays Vietnamese. That geography trade beats any product pick.
Services throw off cash in weeks at near-zero capital. Use that cash to fund the one product bet — not your savings. Asset-light, risk-tranched, exactly the structure you already favor. Pick one, not four.
You logged in, so I opened all 22 linked case studies and read the real numbers. The full list stayed paywalled, but the detail pages didn't. Here's what actually moved.
The businesses you can run solo-to-tiny cluster at the top: consulting (3 people), chefs (2), dashp (1), stackero (1), myzia (2). The ones I told you to skip run on armies — perfect.live 100, OSINT 70, chatfuel 38, plurio 37, brümate 57. The model held.
Its founder shipped the commission tracker solo on Retool (no-code) for $0. That kills my "needs an $8k build" assumption — you'd do the same on Retool or Claude Code. Fit 73.6 → 82.6, and the card economics are corrected (upfront ~$1k, not $8k).
The list and the founder stories disagree by 2–5× (rankings.io: list $2.5M/mo vs story "$3.6M/year"; modernproducers: $233K vs "$85K/month"). Rank on fit and team size, not on their revenue.
rankings.io's $2.5M/mo runs on 168 employees. Still a top pick as a model, but that ceiling is a staffed agency, not a solo play — you'd start alone and it scales only with headcount.
Sort by any column. Fit is your ability to execute (rubric at the bottom). US / Asia / VN are market attractiveness for that business, 1–5. Reported revenue is self-reported and survivorship-biased — treat it as noise, not signal.
| # | Idea & your angle | Archetype | Build | Team F/E | Fit ▾ | US | Asia | VN | Rev/mo* |
|---|
*Reported by the founder to Starter Story; unverified, survivorship-biased. A high-revenue row is not a recommendation.
Five different business models, ordered by fit. Two are "get-found" cousins pointed at different markets (US vs VN) — that split is deliberate, it answers your market question. Each card: the play, the money, how it dies, how it wins, and the minimal-capital steps. Numbers are computed, not guessed; every one is labeled.
A four-slide Instagram recut of a Dec 2025 interview with Louis-David Paul-Hus (GlowUp, AI makeup try-on, Canada). What the slides say, and what the sources add. Verified = source opened this session; Claimed = said on camera, unverified.
| Claim on the post | What the sources say |
|---|---|
| “$800,000 in revenue” | Verified: ~$800K over ~15 months (launch Oct 2024). By the interview it was already $15K/month. |
| 100K users in 3 days · #1 Lifestyle · #4 free overall | Consistent across the case study and the transcript. One video: ~48M views, 60K downloads, $12.5K sales in a day (Claimed). |
| What the post leaves out | Verified: TikTok reach stopped → revenue fell → he sold GlowUp for $150K. The app now lists under Zipo Apps Ltd. (13.4K ratings, 4.4★). His new portfolio makes $3–4K/month. His own post-mortem: an undiversified channel tanked the business. |
Post about a product that does not exist and see whether strangers ask for it. A $0 stranger-demand test — the exact test the council record says you have never run.
A Terms-of-Service workaround. Works until it does not; multi-account fleets are the first thing TikTok purges.
A volume job, not a 5–10 h/week activity. It is also the part that died and took the business with it.
Sound. $9.99/week at a ~90% margin made the year — and is why the app was worth only $150K once the traffic went.
youtube.com/@starterstory — 879K subscribers; the 30 most recent videos run 36K–366K views each. 20 transcripts chosen across playbooks plus the two most-viewed of the quarter. Six more refused transcripts after rate-limiting and are not covered. Every founder number is self-reported on camera.
| Founder | Business | How a non-coder did it | Fit |
|---|---|---|---|
| Rashid — AngelMatch / InvestorHunt | 3 paid B2B directories, $32K MRR (Claimed) | Hired devs and six writers; hand-built 40K rows; SEO + Product Hunt | High — data you can assemble; SEO compounds without you |
| Bo — Savvy Nomad | Productised US state-tax domicile change, $1.7M/yr (Claimed) | Entire product on Bubble; sells a computed saving in onboarding | High on model, conflict on subject (option 5) |
| Tomer — Israeli Cal AI clone | Localised a proven US app, $81K/mo (Claimed) | Cursor MVP in 2 weeks; cheap local CPMs; influencer rev-share | High — a marketer’s play; VN CPMs are cheap |
| “Lots” — app flipper | 4 apps built to $10–20K MRR and sold, $500K total (Claimed) | Upwork engineer; UGC + Meta ads; MicroAcquire exit at 2–4× | High — fix-and-flip applied to apps |
| Roman — Career Hound | Job board, ~$25K/mo (Claimed) | Tech co-founder; he runs $20 creators via WhatsApp and a format sheet | Medium — needs a builder |
| Sarah — Stella | Manifesting app, $340K/mo in 60 days (Claimed) | Vibe-coded in 2 days on a 4M-follower personal brand | No — it is her face |
The constraint that binds every option, from your own ledger: a stranger has paid, or committed in writing. Nothing you have built has passed it, so each option is written so the first milestone is a stranger paying, cheaply, within weeks, before anything is built. The App Intel Engine KILL (22/09) still stands — none of these mines reviews and none is a build factory.
Beauty dominates TikTok Shop Vietnam (>$10B GMV in 2025, second only to Shopee — Verified, industry reports). Staff film faceless before/after and process clips; hashtag-mine the VN hair tags exactly as Louis did in #makeup; anchor link → booking or the products the salon already sells. You approve once a week. Gate already passed (strangers pay the salon), faceless so no visibility problem, zero cost, and it makes the one upside line in your portfolio produce data. First milestone: 20 clips in 2 weeks, measure traceable bookings. Kill: none in 30 days.
The flipper builds to $10–20K MRR and sells at 2–4× profit. Invert it: buy a $1–5K MRR consumer app on acquire.com with a working paywall and dead marketing, add the distribution the seller never did ($20 UGC, VN-localised ads), hold 6–12 months, relist. An acquired app passes Gate 1 on day one. Capital is the input, not code or hours — fix-and-flip applied to apps. First milestone: a 20-listing spreadsheet, one evening, $0. Risks: US developer and ad accounts in your name; you are buying rented reach unless you diversify it. Council before any LOI.
Vietnamese solo developers already sit in today’s VN Lifestyle top-grossing chart: two tử vi apps at #4 and #9 (4,691 and 3,989 ratings), a Vietnamese dating app at #13 with 62,862 ratings (Verified). Pick a huge, US-validated category (calorie scan, height, glow-up, habit lock), ship a Vietnamese version with FlutterFlow or Claude Code, price at VN levels (49–99K₫/month, Assumption), acquire with cheap VN CPMs and one KOL on rev-share after traction. Test first: 10 faceless Vietnamese TikToks about the not-yet-existing app; if comments do not ask for it, do not build. Ceiling: a $10–30K/month business, one Cursor session from being cloned.
You already run pipelines that collect structured commercial data weekly. A niche database strangers pay $49–97/month for — verified township-scale retail and F&B operators, or foreign-brand franchise holders in Vietnam. Manual first 2,000 rows, programmatic SEO, Stripe. Test first: three pre-orders from LinkedIn/Zalo contacts outside Omni before assembling row one. Why fourth: Rashid’s SEO ramp was 6–8 months, and the subject sits close to your employer’s data.
Savvy Nomad’s model is the best fit to your expertise (compliance, licensing, internal control) and this page already scores “fractional commercial strategist” at 83/100. But it puts your name on a paid service in your exact professional domain while you hold the KSNB/QTRR seat. The Vitals kill was about visibility; this is more visible, not less. Park it until the employment question is decided, or run it only for a non-Vietnamese buyer.
Rebuild the GlowUp engine literally (US SIM, VPN, seven phones) — it is the part that failed, it breaks the platform’s terms, and it is a full-time job. Anything that starts with a build and ends with “then we market it”. Convening the Dream Team on all five — pick one; the chain manufactures queue when run on a list.
Sequence: two hours this week on option 1 (format sheet in Vietnamese from the hashtags that rank); one evening on option 2’s listing spreadsheet; then a council sitting only on whichever shows a stranger paying. Full brief with sources: 93 KB/Claude outputs/Starter Story/260925_SS_Brief_YouTube-Lessons-and-Options_v1.md.
Each idea scored 1–5 on eight dimensions, weighted:
| 18 | Buildability without a dev team | 18 | Capital efficiency (low upfront) | 18 | Your edge (domain + network) |
| 14 | Distribution access | 9 | Velocity to first revenue | 9 | Bandwidth fit (part-time) |
| 9 | Scale ceiling (worth your time) | 5 | Defensibility |
Verified AEO/GEO 2026 retainer bands ($1–2.5k entry, $3–8k mid, $10–25k+ enterprise) and that the category is real but crowded in the US (a $7.3B-by-2031, ~34% CAGR projection). All fit and market scores are my judgment on a fixed rubric, recomputed in code. Assumption every unit-economics figure on the cards — client counts, retainers and margins are planning cases, not forecasts; VN retainers are marked Likely (VN agency rates run well below US).
Only 24 of the claimed ~99 problems load even with a free account — the full list is a paid Starter Story feature. Logged in as the user, I opened all 22 linked case studies and read team size (founders / employees) and headline revenue from each; the team figures are Verified and populate the Team column, and they drove the dashp re-rating (73.6 → 82.6). Self-reported revenue is internally inconsistent across the list and the stories, so I did not weight it. Counter-case on the top pick: consulting scores highest partly because it's easy and instant, but it caps at your hours and collides with the Ecopark day job — which is why the shortlist also carries scalable (agency, micro-SaaS) and passive (media) options. Working on the memory floor (kernel not mounted); the palette here is a Bright-Professional floor, not your locked kernel hex.