1. Verdict
Definition of "success"
This portfolio uses an MRR ladder by month 24 ($1k, $3k, $5k, $10k, $20k, $50k), profitable at each level, plus two fixed tiers so ideas can be compared (../../README.md):
- Tier A (venture): $1M ARR within 36 months, ≥70% gross margin, ≥50% month-12 logo retention.
- Tier B (profitable business, the default target): $20k+ MRR (≈ $250k ARR) within 24 months, profitable, ≥80% gross margin, ≥70% month-3 retention.
Which wedge? Five versions compared
The idea says "the wedge is underserved formats: a region's banks, or document types the leader handles poorly". Here is each candidate, scored on the same ladder.
| Version | $1k | $5k | $10k | $20k | Why |
|---|---|---|---|---|---|
| As described: a generic converter for bank, payslip, brokerage and invoice PDFs, all regions | 20% | 5% | 2% | 1% | The most contested phrase in the category. QuickBooks does English statements natively (verified). DocuClipper starts at $20 a month (verified). The model cost is under 1¢ a page. You'd be the 30th clone, 5 years late to search. |
| India bank statements into Tally | 25% | 6% | 2% | <1% | A real, painful job, but Vyapar TaxOne (formerly Suvit) sells unlimited statements into Tally for ₹10,000 a year and claims 10,000+ practising firms (verified). Price ceiling ~$10 a month, so $10k MRR needs ~1,000 firms. |
| Non-English statements (one Spanish-speaking market, or one EU country) | 25% | 10% | 4% | 1–2% | QuickBooks' native import is English only (verified), and the comparable's audience is concentrated in English-speaking countries (verified, 2022). But EU banks export CAMT/MT940 files and have PSD2 feeds, Latin American prices are lower, and local competitors couldn't be checked in this run. Kept as Pivot B. |
| Extraction API for lenders and underwriters | 10% | 4% | 2% | 1% | Enterprise sales, security questionnaires and funded incumbents. Not reachable at limited hours per week. |
| Recommended: statement reconstruction for US accountants (forensic, divorce, tax resolution, catch-up, estates) | 45% / 60% | 20% / 30% | 10% / 18% | 4% / 8% | The job bank feeds can't reach (verified: Plaid gives up to 24 months, with consent). Buyers bill $150–400 an hour and must defend every number. It's won on proof and bulk, not on extraction. Fewer, higher-value customers: ~137 accounts for $10k. |
Figures in the last row are ~15 hrs/week / full-time. All figures are judgment estimates. The other rows assume full-time, because they need more effort, not less.
Probability summary (recommended version)
| Scenario | ~15 hrs/week | Full-time |
|---|---|---|
| $10k MRR by month 24 | 10% | 18% |
| Tier B ($20k+ MRR, profitable, ≥80% GM, ≥70% M3 retention) | ~3% | ~7% |
| Tier A ($1M ARR in 36 months) | <1% | ~1% |
| Base rate: no verified base rate for bootstrapped B2B micro-SaaS. Proxy: 4.6% of new subscription apps reach $10k a month in 2 years (Business category: 1.6%) |
Fundable in current form?
No. And it shouldn't try to be. A converter is a feature of every accounting platform (QuickBooks already shipped it). The recommended version is a small, sharp, profitable tool. The ceiling for a solo operator is roughly the comparable's: $300–500k a year. That's an excellent owner-run business, and a poor venture investment. See 06.
One-paragraph verdict
The demand is proven by one of the best-documented solo SaaS stories on the internet. Bank Statement Converter went from $3k to $318k a year in four years, on SEO and accuracy, with about $500 a month of costs. But what he sold in 2021 (turn a PDF into rows) is now a free feature of QuickBooks and a single prompt in any LLM chat. A 2026 entrant can't win the generic job. It can win a harder job next to it: taking a folder of hundreds of statements across accounts and years, proving they all balance, showing where every number came from, and handing over one workbook a professional can defend. The buyers are forensic accountants, divorce financial analysts, tax-resolution practitioners and catch-up bookkeepers. They exist because bank feeds and open banking can't reach the documents they work with. The market is small and episodic, so the odds are modest: ~10–18% for $10k MRR and ~4–8% for $20k. But the operating model suits limited hours, automation and low support. Build the reconstruction tool, or don't build this at all.