StatementKit

Three most likely causes of death, ranked by probability. They compound: #1 makes #2 worse, because a commodity product can only grow through search, and #2 leaves the founder chasing every layout alone, which leads to #3.

---

Cause #1: Commoditised before a search position was won (probability ~45% of failure cases)

What happened. The founder built a solid converter in 3 months. By then:

• QuickBooks' native PDF import had spread to more regions and languages;
• Xero shipped its own version;
• every LLM chat app handled "turn this statement into a table" with drag-and-drop.

Search results for "[bank] statement to Excel" filled with AI answers, the comparable's 5-year-old pages, and 30 clones. The product was pitched as "more accurate", which nobody could check. Signups came, converted one file for free, and left. Paid users were mostly one-off catch-up jobs that cancelled after a month. By month 12, MRR was stuck at $1.5k.

Early warning signs

• Month 3: most signups arrive via "PDF to Excel"-type searches, not reconstruction terms.
• Median pages per paying account under 100 a month (one-off users, not case users).
• Free-to-paid conversion below 2%, and month-1 paid churn above 20%.
• Users describe you as "a converter", not "the tool for my cases".
• A platform (QuickBooks, Xero, a big LLM app) announces bulk PDF statement import.

Fixable? Yes, if it's done early. Pick the reconstruction buyer on day 1, build the proof and the case workflow first, and write the search pages for use cases rather than generic conversion. After 12 months as a generic converter, it's too late to re-position cheaply.

---

Cause #2: Search never compounded, and there was no second channel (probability ~35%)

What happened. The founder followed the comparable's playbook: bank-specific pages. But the comparable, DocuClipper and dozens of clones already held those results. AI search answers took more of the clicks, and a new domain took 9–12 months to rank at all. With limited hours per week, the publishing routine slipped from 3 pages a week to 1. Nobody in the forensic or tax-resolution communities knew the product existed, because the founder never showed up there. Paid search was tried and lost money, just as it did for the comparable ($3,908 spent for about $988 of revenue).

Early warning signs

• Month 6: fewer than 30% of signups from organic search, and fewer than 500 organic visits a week.
• Fewer than 150 search pages published by month 6.
• No talks, guest articles or community posts by month 9.
• More than 50% of revenue from the founder's personal network (it doesn't scale).

Fixable? Partly.

• Use-case pages ("bank deposits analysis for unfiled returns") are less contested than bank pages, so start there.
• Professional communities (forensic, divorce-analyst and tax-resolution associations) are small and trusting. One speaking slot can outperform a year of posts.
• The design partner's network is your first 20 customers.

You can't fix it by buying ads. The comparable proved that.

---

Cause #3: An accuracy or trust event in a small professional community (probability ~20%)

What happened. A divorce financial analyst used a StatementKit workbook in a deposition exhibit. A scanned statement had a row the model silently dropped. The balance check caught it, but the user exported anyway, because the red badge was easy to miss. Opposing counsel found the gap. The analyst posted about it in a professional group. Or a different version: a customer asked whether their discovery documents had been sent to a model provider that keeps data, and nobody could answer quickly. Either way, in a community of a few thousand practitioners, one story ended new signups.

Early warning signs

• Balance-check pass rate below 97% on new uploads (not just the test corpus).
• Users exporting statements that failed the check (it should never be silent).
• Support tickets about missing or duplicated rows, over 1 per 50 cases.
• Security questions you can't answer in writing within 24 hours.
• Growing layout-fix backlog (more than 2 weeks old).

Fixable? Yes, and it's prevented cheaply.

• Block or loudly flag export of failed statements, and require a typed acknowledgment.
• Put the proof status on the workbook itself.
• Publish the security summary.
• Use zero-retention model terms.
• Keep the layout backlog short with the "report a bad row" loop.

---

Honourable mentions (less likely, still real)

• Episodic churn. Case users cancel between cases. Monthly churn runs at 10–12%, and $10k becomes unreachable. The fix is annual plans, case passes, and catch-up work as the between-cases use.
• DocuClipper moves down into cases. It already has 10,000+ finance teams, SOC 2 and brokerage support. If it adds case folders and a balance proof, your edge narrows to price and search position.
• Founder hours. At ~15 hrs/week, the search routine, layout fixes and support compete for the same hours. Something slips, and it's usually the publishing, which is the channel.
• Security incident. One leaked case folder of someone else's bank statements ends the business and may create liability. Encrypt, minimize retention, and get cyber insurance once you have Firm accounts.