Prove three things before calling it a business.
People must pay for a bank-connected analysis. The analysis must save more value than it costs. The service must run without buying the founder a job.
The API bill is small. Human time is not.
The prior brief claimed 74–93% margins while omitting support labor. That made the weakest cost look like the strongest business.
$20 analysis
Contribution before shared infrastructure. Includes twelve support minutes. P is the undisclosed Plaid price per Item-month.
$5 rerun
Contribution before shared infrastructure, assuming no new Plaid Item cost. Stripe takes 8.9%. Sell credits or packs instead.
| Offer | Revenue | Payment | Model | Support | Contribution |
|---|---|---|---|---|---|
| Analysis | $20 | $0.88 | $0.12 | $8.00 | $11 − 2P |
| $10 rerun | $10 | $0.59 | $0.02 | $2.67 | $6.72* |
| Monthly Pro | $5 | $0.48 | $0.16 | $1.33–3.33 | $3.03 − 2P to $1.03 − 2P |
Measured: Stripe public US pricing; OpenAI public pricing; Supabase public pricing. Assumption: support at $40/hour. *No incremental Plaid Item.
Conversion is a hypothesis, not a benchmark.
No credible public benchmark ties a general visitor to a paid consumer-fintech analysis after bank connection. Plaid publishes Link optimization claims, not this full funnel. The original 0.4–1.0% range was unsupported.
| Assumption | Downside | Working | Upside |
|---|---|---|---|
| Visitor → analysis | 0.15% | 0.35% | 0.70% |
| Analysis → annual | 8% | 15% | 25% |
| Rerun attach | 5% | 12% | 20% |
| Year-one revenue | $572 | $7,347 | $64,878 |
| After modeled cost | −$93 | $2,888 | $26,409 |
Assumption: revised downside-first rates. Measured: model output from evidence/sol-r51-model.mjs. Traffic targets are unchanged assumptions. Model includes support and $30/month infrastructure.
A buyer pays for profit that survives handoff.
Acquire.com reports a 3.9× median profit multiple for SaaS deals in both 2024 and 2025. Small buyers usually anchor on verified profit. Applying 2.5–4× ARR at month 18 without churn history overstates the case.
Premium only after proof
- Clean cohort retention and low refunds
- Verified profit after support
- Assignable contracts and vendor accounts
- Documented operations with low owner hours
- Defensible accuracy and audit trail
Romance until proof
- “Verified-data pipeline” without lift data
- Partner relationship without assignment rights
- Plaid approval tied to the seller entity
- Automation logs without outage metrics
- Audit volume without customer outcomes
Use 3–4× trailing annual profit as the base case. Treat strategic premium as zero until a named buyer validates it.
Acquirer reality
| Candidate | Evidence | Ruling |
|---|---|---|
| Acquire.com buyer | Marketplace publishes micro-SaaS profit multiples | Most credible BATNA |
| AwardWallet | Existing reported data relationship | Conversation worth earning; no acquisition precedent supplied |
| CardPointers, MaxRewards, point.me, Roame | Product adjacency only | Hypotheses, not buyer list |
| Intuit, Red Ventures, Gen Digital | Have acquired fintech/media assets | Real acquirers, unrealistic before material scale |
| Goldman Sachs | Acquired Clarity Money, later folded it | Precedent warns that acquisition can erase the product |
Reported: Acquire.com 2026 multiples report. Reported precedents: Intuit/Credit Karma, Goldman Sachs/Clarity Money, Zip/Pocketbook, Gen Digital/MoneyLion. These large transactions do not validate a micro-acquisition premium.
The data room begins before revenue.
- Entity owns code, domains, marks, data, and vendor accounts
- Founder and contractor invention assignment
- AwardWallet contract assignment and change-of-control terms
- Plaid, Stripe, Supabase, OpenAI, Resend, Cloudflare terms and DPAs
- Privacy notice names model-assisted transcription and subprocessors
- Consent, deletion, retention, refund, and incident evidence
- MRR, gross profit, churn, refunds, cohorts, CAC, and channel attribution
- Owner-hours and support-time ledger
- Backup restore and disaster recovery evidence
- Security inventory, access review, secrets rotation, dependency bill
- Tax, chargeback, complaints, sanctions, and litigation schedule
- Transition plan and customer communication rights
Measured from Plaid docs: a client ID uniquely identifies the team; an access token works only with credentials from the same developer account. Inferred: transfer to another buyer account is not automatic. Contract assignment and acquisition procedure require direct Plaid confirmation.
Earn the right to sell.
Pilot
100 paid analyses. Prove coverage, delivery, refunds, support time, and repeat demand.
Operate
Six months of cohort data. Reach positive contribution after labor. Keep founder work below a declared ceiling.
Market test
At $2K MRR, request three indications. Compare 3–4× trailing profit against the value of another year.
What this revision changed
It added support labor, replaced unsupported conversion rates with downside sensitivities, cut projected working revenue from $14.6K to $7.3K, changed exit math from ARR multiples to profit multiples, demoted adjacent companies from “likely acquirers” to hypotheses, and added transfer, privacy, contract, security, tax, and operational diligence.