Score C · 5.70/10Free to read$100K Hackathon · November 16, 2026 → December 16, 2026Nov 16, 2026

Formula Forensics

A browser tool that scans a single uploaded .xlsx and returns a one-page PDF listing every hardcoded number, broken reference, and circular formula, for finance teams who inherited a spreadsheet they don't trust.

One product, one founder, thirty days. Everything below is built to be executed inside the contest window — the official rules still decide what counts.

Decide in 20 seconds

The facts that decide whether this is worth thirty days of yours.

Speculative5.70/10
MVP build (solo)
4 days
Launch price
$49
Sales to $1,000 revenue
21 customers
First paying customer
Direct LinkedIn outreach to controllers and finance managers whose posts or recent job changes signal an inherited spreadsheet model.
Who pays

Finance manager or controller at a 10-200 person company who has just inherited or acquired a budget, forecast or pricing model built by someone else.

Biggest execution risk

Monetization potential 3.8/10

Tie the price to a measurable money or time outcome and add a recurring or repeat-purchase path instead of a single low-value charge.

Why this could work in 30 days: one founder can ship the MVP in 4 days; the plan targets a first sale on day 9; 21 sales at $49 clears $1,000 of collected revenue; strongest signal is speed to first revenue at 8.3/10.

Score and projections come from the generated report and are not guarantees. The $100K Hackathon is run by Ship or Die — read the official rules before you build.

Your first 24 hours

10 concrete actions. Tick them off as you go — nothing is saved or sent.

0/10
Entry price
$49
Solo build
4 days
30d base revenue
$1,280
30d hackathon profit
$1,173
Section 1 of 8

Product opportunity and positioning

Target customer

Finance manager or controller at a 10-200 person company who has just inherited or acquired a budget, forecast or pricing model built by someone else.

Why now

Model handovers cluster at year-end close, acquisitions and ERP migrations, and every one of those moments forces a new owner to defend numbers they cannot personally vouch for.

Problem

Inherited spreadsheets hide hardcoded overrides, dead links and circular logic that silently corrupt numbers before they reach a board deck or a lender.

Solution

User uploads one .xlsx up to 20MB, pays $49 by card, and receives a timestamped PDF integrity report listing each finding by sheet and cell, emailed within 60 seconds.

Differentiation · Not an Excel add-in and not a $30k audit platform: one file in, one timestamped PDF out, no install, no account, no subscription, priced below an hour of analyst time.

Product opportunity and positioning

Formula Forensics sells one artifact: a timestamped PDF that tells a finance manager exactly what is wrong with a spreadsheet they did not build.

The trigger is a handover, and handovers are dated events. A controller starts at a 60-person company. A company closes an acquisition. A departing analyst leaves behind the forecast. An ERP migration forces old models into a new system. In each case the spreadsheet still feeds the board deck, the lender covenant test and the pricing approval, and nobody can say which cells are genuine formulas and which are typed-over overrides. The anxiety is not abstract; it is the fear of presenting a number that moved for no reason.

The existing options are bad in a specific way. Clicking through cells leaves no record. Asking an analyst burns an afternoon and also leaves no record. A full audit platform requires procurement and does not fit a 60-person company. Formula Forensics produces a shareable record in under a minute for the price of a lunch, and because the PDF carries sheet and cell addresses it is credible without a phone call.

Positioning is deliberately narrow: one .xlsx in, one PDF out, $49, no account, no add-in, no subscription. It is a receipt, not a platform.

The JEV score of 57/100 is a fair diagnosis. Speed to first revenue (8.3 normalized) and solo build feasibility (7.5) are why this wedge was chosen: a working product in days, sold to one anxious buyer who already holds a card. Market demand (4.1) and monetization potential (3.8) are the real weak points. The pain is genuine but episodic, and a single low-ticket charge has no natural expansion path.

The 30-day plan treats those two weaknesses as the actual work. Demand is addressed by narrowing further than the category: not 'spreadsheet review' but 'you inherited this file, prove it is sound.' Monetization is addressed by a Five-Pack at $199 offered inside the report email, and by a fractional-CFO channel where one buyer scans client files repeatedly. Both are testable inside the window with outreach volume the founder can already send.

Section 2 of 8

Smallest sellable MVP

A single upload page that parses one .xlsx, generates a findings PDF, and unlocks the download behind a $49 Stripe Checkout.

Must-have features

  • Upload page accepting one .xlsx up to 20MB
  • openpyxl parser building a formula and cell index
  • Numeric-literal-inside-formula detector (hardcodes)
  • #REF!/#VALUE!/#DIV/0! and broken cross-sheet link detector
  • Circular reference detector over the dependency graph
  • One-page branded PDF with sheet, cell, formula and severity
  • Stripe Checkout gating the report download
  • Resend email delivering a signed report link

Explicitly out of scope

  • No Google Sheets support
  • No live spreadsheet editing
  • No multi-file or portfolio dashboards
  • No team accounts, SSO or seats
  • No AI narrative commentary beyond a two-sentence summary
  • No recurring subscription at launch
  • No Excel add-in

Demo flow (what the buyer sees)

  1. Buyer opens the landing page and clicks Scan a file
  2. Uploads one .xlsx (20MB cap) and the parser runs in a serverless function
  3. Findings JSON is stored against a scan token
  4. Stripe Checkout collects $49
  5. Webhook flips the token to paid and the PDF renders
  6. Resend emails a signed download link
  7. Buyer opens a one-page PDF with cell-level findings and a UTC timestamp
Build plan

Technical implementation plan

Stack

Next.js on VercelPython serverless functionopenpyxlSupabase Postgres + StorageStripe Checkout + webhookResend for transactional emailRecharts-free PDF via server-side HTML-to-PDF

Architecture

Next.js serves the marketing page, the upload form and the report page. A Python serverless route streams the workbook through openpyxl, builds a formula index, and runs three detectors that emit finding objects into Supabase against a scan token. The PDF is rendered server-side, cached in Supabase Storage, and served through a signed URL. Stripe Checkout creates the payment and a webhook flips the scan token to paid, after which Resend emails the link. Nothing stateful runs on the founder's machine.

Build steps

  1. Create the Next.js repo, buy the domain, deploy a landing page with the $49 offer
  2. Wire Supabase project, uploads bucket and scans table
  3. Implement the upload endpoint with a 20MB cap and 60-second parse timeout
  4. Build the openpyxl formula index (sheet, cell, formula string, precedents)
  5. Ship the hardcode detector (numeric literals inside formulas)
  6. Ship the error-value and broken-reference detector
  7. Ship the circular reference detector using a dependency graph
  8. Render the findings JSON into a one-page PDF template
  9. Gate the PDF behind Stripe Checkout and a webhook
  10. Send the signed PDF link with Resend
  11. Dogfood on 20 real-world workbooks and fix the top failures
  12. Add graceful failure messaging that auto-refunds unparseable files

AI components

  • Optional: a single model call that writes the two-sentence summary at the top of the PDF from the findings JSON. It never generates or edits findings.
Hardest technical risk: openpyxl on messy real files: password protection, macro-enabled workbooks, external links, shared formulas and 50k-row sheets. A silent partial report is worse than a clear failure, so every unreadable file must fail loudly and refund automatically.
Section 3 of 8

Pricing and monetization

Single Scan

Entry offer
$49 one-time

For Finance manager with one inherited model

  • One .xlsx scan
  • One-page PDF integrity report
  • Email delivery

Five-Pack

$199 one-time

For Controller cleaning up several models

  • Five .xlsx scans
  • Five PDF integrity reports
  • Priority email delivery

Model: One-time per-file payment with a discounted five-scan pack; no subscription at launch.

First revenue offer: Free scan of one inherited .xlsx, delivered as a PDF within an hour; the full report with all findings unlocked costs $49.

Expected ACV: $64

Upsell path: Offer the $199 Five-Pack inside every report email to Single Scan buyers, then sell a $199 pre-close bundle to fractional CFOs who scan client files at month-end close.

Target customers
60
Revenue model
$2,940
Costs
$120
Profit model
$2,820
Section 4 of 8

First-customer acquisition

Primary channel

Direct LinkedIn outreach to controllers and finance managers whose posts or recent job changes signal an inherited spreadsheet model.

Where to find them

  • LinkedIn posts containing 'inherited model' or 'took over the forecast'
  • LinkedIn job changes to controller or finance manager at 10-200 person companies
  • Finance-ops and FP&A Slack and Discord communities
  • Fractional CFO and bookkeeping firm owners who inherit client models
  • Month-end close and post-acquisition integration threads

Exact outreach motions

  • Send a two-line message offering a free scan of one file, no meeting request
  • Deliver the free PDF within an hour and reply in the same thread
  • Ask for $49 only after the PDF is in their hands
  • Follow up once at 48 hours with a sample report attached
  • Post an anonymized 'common hardcode patterns' note in two finance-ops communities

Message template

Hi {first name} - you mentioned taking over a model someone else built. I run Formula Forensics: it scans one .xlsx and returns a one-page PDF listing every hardcoded number, broken reference and circular formula with the exact cell. Want me to run it on your file free? Takes an hour and you keep the PDF either way.
Week-one goal: 30 LinkedIn messages sent, 5 free scans delivered, 1 paid $49 report.
Section 5 of 8

30-day execution roadmap

MVP deadline: day 5First sale target: day 930 days planned
Week 1 · Build & validate · 7 days
Day 1Setup

Ship the shell

  • Create the Next.js repo, buy the domain, deploy the landing page with the $49 offer.
  • Create Supabase project, uploads bucket and scans table.
  • Open Stripe, connect the account to TrustMRR, create the $49 and $199 products.

Deliverable: Live landing page, Stripe products, deployed empty app.

Day 2Parser

File in, findings out

  • Build the .xlsx upload endpoint with a 20MB cap and 60s timeout.
  • Parse with openpyxl into a formula index of sheet, cell and formula string.
  • Emit hardcode findings as JSON.

Deliverable: JSON findings for hardcodes on a test workbook.

Day 3Checks

Complete the detector set

  • Add #REF!/#VALUE!/#DIV/0! and broken cross-sheet link detection.
  • Add circular reference detection via a dependency graph.
  • Run all four classes against 10 sample workbooks.

Deliverable: All finding classes working end to end.

Day 4Report

Make the PDF credible

  • Render a branded one-page PDF with sheet, cell, formula and severity tags.
  • Add a UTC timestamp and a findings count header.

Deliverable: Downloadable PDF from a preview route.

Day 5CommerceMVP shipped

Turn on payment

  • Gate the PDF behind Stripe Checkout and a webhook that flips the scan token to paid.
  • Send the signed report link with Resend.

Deliverable: End-to-end paid scan live (MVP deadline).

Day 6Dogfood

Survive real files

  • Run 15 real-world .xlsx files through the parser and log crashes and false positives.
  • Fix the five worst failures.

Deliverable: Parser survives 15 of 15 files.

Day 7List

Build the prospect list

  • Assemble 150 controllers and finance managers from LinkedIn posts and job changes.
  • Record source URL and trigger phrase per prospect.

Deliverable: CSV of 150 named prospects.

Week 2 · Launch & first sale · 7 days
Day 8Outreach

Wave one

  • Send 30 LinkedIn messages offering a free scan of one file.
  • Post a short hardcode-patterns note in two finance-ops communities.

Deliverable: 30 sends logged, 2 community posts live.

Day 9First saleFirst sale target

Convert a free scan

  • Run every free scan personally and email the PDF within the hour.
  • Ask for $49 in the same thread after delivery.

Deliverable: First paid Stripe charge.

Day 10Quality

Review every report

  • Manually check each delivered PDF and log missed findings.
  • Add a context column showing the hardcoded constant beside its formula.

Deliverable: Report template v2.

Day 11Outreach

Wave two

  • Send 40 messages with a second opener angle.
  • Log reply rate per opener.

Deliverable: 70 cumulative sends.

Day 12Upsell

Test the Five-Pack

  • Offer the $199 Five-Pack to anyone holding more than two files.
  • Record accept and reject reasons.

Deliverable: First Five-Pack result documented.

Day 13Content

A public proof page

  • Publish an 'Inherited Model Checklist' page with a sample report embedded.

Deliverable: Indexable page live.

Day 14Review

Week two numbers

  • Tally sends, replies, scans and revenue.
  • Kill the worst-performing opener.

Deliverable: One-page metrics snapshot.

Week 3 · Repeat & systematize · 7 days
Day 15Hardening

Fail loudly

  • Handle password-protected, macro-enabled and 5k-formula files with a clear error.
  • Auto-refund any file that cannot be parsed.

Deliverable: Graceful failure paths shipped.

Day 16Outreach

Wave three

  • Send 40 messages aimed at post-acquisition and ERP-migration posts.

Deliverable: 110 cumulative sends.

Day 17Referral

Borrow the buyer's network

  • Add a share line to the report PDF and a $20 credit code in Stripe.

Deliverable: Referral code live.

Day 18Content

Anonymized findings

  • Publish aggregate hardcode statistics from delivered scans, no client data.

Deliverable: Community post two live.

Day 19Follow-up

Close the loop

  • Send the 48-hour follow-up with a sample PDF to every non-responder.

Deliverable: Follow-ups sent to 100% of non-responders.

Day 20Conversion

Landing page rewrite

  • Rewrite the hero around the inherited-model trigger.
  • Add a downloadable sample PDF.

Deliverable: Updated landing page deployed.

Day 21Review

Channel decision

  • Compute paid conversions per 100 messages.
  • Decide whether one sponsored slot is worth testing.

Deliverable: Decision logged.

Week 4 · Scale what works · 6 days
Day 22Outreach

Wave four

  • Send 40 messages to fractional CFOs and bookkeeping firm owners.

Deliverable: 150 cumulative sends.

Day 23Speed

Under a minute

  • Cache the parser warm path and trim PDF render time.

Deliverable: p95 report delivery under 60 seconds.

Day 24Partners

Fractional CFO angle

  • Email 15 fractional CFOs offering three free client scans.
  • Log replies and any reseller interest.

Deliverable: 15 partner emails sent.

Day 25Trust

Refund guard

  • Add a confidence note and auto-refund when a file over 200 formulas returns zero findings.

Deliverable: Refund policy shipped in-product.

Day 26Outreach

Wave five

  • Send the final 50 messages and close every warm thread.

Deliverable: 200 cumulative sends.

Day 27Upsell

Mine existing buyers

  • Offer the $199 Five-Pack to every Single Scan buyer from days 9 to 26.

Deliverable: Upsell emails sent to all buyers.

Days 29-30 · Consolidate · 3 days
Day 28Content

Final push

  • Post the 'five hardcodes that killed a forecast' write-up in three communities.

Deliverable: 3 posts live.

Day 29Finance

Reconcile

  • Match Stripe payouts against TrustMRR records including refunds and fees.

Deliverable: Reconciled P&L sheet.

Day 30Close

Publish and archive

  • Publish the 30-day numbers and archive all delivered reports.
  • Back up the Supabase scans table.

Deliverable: Final metrics page and backup.

Section 6 of 8

Revenue and profit scenarios

How the numbers are defined

Hackathon profit — Collected revenue − refunds − chargebacks − payment processing fees − paid marketing spend.

Business profit — Hackathon profit − operating costs (hosting, domains, tooling, AI/API delivery costs).

The contest ranks entrants on hackathon profit only. Hosting, tooling and AI/API delivery costs are not deducted by the organisers, but each sale still has to cover them — so we show both. Definitions verified on 2026-11-15 against the official rules.

Stated assumptions

  • All three scenarios are projections from stated assumptions, not verified revenue.
  • Base case: 200 LinkedIn messages, ~15% reply, ~25% free-scan-to-paid conversion, supported by 200-message volume across five waves.
  • Blended ARPU reflects the mix of $49 Single Scan and $199 Five-Pack: Conservative $55 (~4% Five-Pack), Base $64 (~10%), Upside $78 (~19%).
  • Payment processing assumed at Stripe 2.9% + $0.30, modelled as roughly 4.5-5.2% of collected revenue because fixed fees weigh heavier on $49 tickets.
  • Chargebacks assumed zero across all scenarios; no disputes expected on card purchases of this size in 30 days.
  • Ad spend is zero in Conservative and Base because acquisition is organic outreach; Upside includes a $120 sponsored-post test.
  • Operating costs cover Vercel, Supabase, Resend, domain, PDF storage and optional model calls; hosting is excluded from hackathon profit but included in business profit.
  • No existing customers, partnerships or traction are assumed; day 9 is the first-sale target, not a booked sale.

Conservative

6 customers
$330
collected revenue · avg payment $55
Refunds
$0
Chargebacks
$0
Payment fees
$17
Paid marketing
$0
Hackathon profit
$313
Operating costs
$18
Business profit
$295

Base case

20 customers
$1,280
collected revenue · avg payment $64
Refunds
$49
Chargebacks
$0
Payment fees
$58
Paid marketing
$0
Hackathon profit
$1,173
Operating costs
$45
Business profit
$1,128

Optimistic

45 customers
$3,510
collected revenue · avg payment $78
Refunds
$98
Chargebacks
$0
Payment fees
$158
Paid marketing
$120
Hackathon profit
$3,134
Operating costs
$95
Business profit
$3,039

Model your own numbers

Runs in your browser · nothing is saved
Collected revenue
$980
Payment fees (3.9%)
$38
Paid marketing
$0
Hackathon profit
$942
Business profit
$897

Hypothetical arithmetic, not a forecast. The fee rate is a flat 3.9% assumption, refunds and chargebacks are left at zero here, and none of this is tracked data.

Scenario models from stated assumptions, not verified revenue. No customers, revenue or traction exist at launch.

Section 7 of 8

Main risks and validation experiments

Risk 1The openpyxl parser fails on complex, macro-enabled or password-protected workbooks and produces a partial report.

Mitigation

Hard 20MB cap, 60-second timeout, explicit unreadable-file message, automatic refund when zero findings are returned on a large file.

Validation experiment

Run 20 real workbooks collected from public templates and personal files through the parser on day 6 and count hard failures.

Kill criteria

If more than 30% of files fail to parse cleanly by day 10, restrict the offer to standard .xlsx only or rebuild the parser on a commercial library.

Risk 2Free-scan outreach gets ignored and the LinkedIn channel produces no paying buyers.

Mitigation

Five waves of 30-50 messages with two opener angles; pivot the same day's budget to community posts if reply rate falls below 5%.

Validation experiment

Measure reply rate at 30, 70 and 110 cumulative sends and compare opener variants.

Kill criteria

If reply rate stays under 5% after 110 sends with both openers, switch the primary channel to finance-ops communities and fractional CFO partners.

Risk 3$49 is perceived as too expensive for a one-time PDF, and conversion from free scan to paid collapses.

Mitigation

Always deliver a real free PDF first so value is visible before the ask; anchor the price against an analyst afternoon.

Validation experiment

A/B the landing page at $49 versus $99 during week two and record checkout starts and completions.

Kill criteria

If fewer than 2 of 20 free-scan recipients convert in week two, drop to a $29 entry price or bundle two scans for $49.

Risk 4Buyers expect ongoing monitoring rather than a one-time scan, producing refunds and no repeat revenue.

Mitigation

Sell the Five-Pack inside every report email and position the PDF as a point-in-time receipt, not a monitor.

Validation experiment

Ask every buyer one question at delivery: how many other files do you not trust?

Kill criteria

If fewer than 1 in 8 buyers takes the Five-Pack by day 27, accept a one-time product rather than building monitoring.

Risk 5Reports miss an obvious error and the buyer demands a refund, damaging credibility in a referral-driven channel.

Mitigation

Hand-review every report until day 21, add a findings-confidence note, and refund without argument when a miss is shown.

Validation experiment

Log every buyer reply mentioning a missed finding and count misses per 10 reports.

Kill criteria

If reported misses exceed 2 per 10 reports by day 20, stop selling and fix the detectors before continuing outreach.

Risk 6The pain is episodic, so 30-day volume does not repeat and the product has no second act.

Mitigation

Open the fractional-CFO channel on day 24 where one buyer scans client files repeatedly.

Validation experiment

Email 15 fractional CFOs offering three free client scans and log interest.

Kill criteria

If neither the Five-Pack nor the CFO channel lifts blended ARPU above $70 by day 25, treat this as a $2-3k one-time product and do not invest further in it.

Section 8 of 8

Hour-by-hour: the first 24 hours

Hours 1-2

  • Buy the domain, create the Next.js repo, deploy a landing page with the $49 offer and a sample PDF placeholder.
  • Open Stripe and create the $49 and $199 products.

Output: Live landing page URL and two live Stripe products.

Hours 3-5

  • Build the .xlsx upload endpoint with the 20MB cap.
  • Parse a test workbook with openpyxl and return a formula index.

Output: Working parse function returning sheet, cell and formula rows.

Hours 6-8

  • Implement the numeric-literal-in-formula detector.
  • Implement the #REF! and error-value scan.

Output: Findings JSON for one real file.

Hours 9-12

  • Render a plain one-page PDF from the findings JSON.
  • Email it to yourself through Resend.

Output: First end-to-end report PDF in an inbox.

Hours 13-16

  • Build a 40-name LinkedIn prospect list with trigger phrases.
  • Write and save the free-scan message template.

Output: 40-row prospect CSV and a saved message template.

Complete report

Full 30-day launch report

Read the full generated report

Formula Forensics - 30-Day Launch Report

1. Product opportunity and positioning

The buyer is a finance manager or controller at a 10-200 person company who has just taken ownership of a budget, forecast or pricing model built by somebody else. The trigger is a handover: a new job, an acquisition, a departing analyst, an ERP migration. The model still feeds a board deck and a lender covenant test, and nobody can say which cells are genuine formulas and which are typed-over numbers.

Formula Forensics answers that in one pass. The user drops one .xlsx into a browser page, pays $49, and receives a one-page PDF listing every hardcoded numeric literal inside a formula, every #REF! and broken sheet link, every circular reference, each with sheet name and cell address. It is not an Excel add-in, not a subscription, not a review platform. It is a receipt for one file at one moment.

That narrowness is the point. The alternative today is clicking through 4,000 cells or asking an analyst for an afternoon, and neither produces a shareable artifact. A timestamped PDF does, which is why it gets attached to handover emails and pre-close checklists and spreads inside the buyer's own team.

The JEV score of 57/100 is a fair read. Speed to first revenue (8.3) and solo build feasibility (7.5) are the strengths: the wedge is days of work and the buyer already holds a card. Market demand (4.1) and monetization (3.8) are the weak spots, because the pain is real but episodic and a $49 one-time charge has no natural expansion. The plan below treats those two weaknesses as the primary work, not the build.

2. Smallest sellable MVP

One upload page, one parser, one PDF, one Stripe button. Nothing else ships on day 5.

What must be true at launch: a 20MB .xlsx uploads; the parser returns findings with sheet and cell; the PDF renders those findings with severity tags and a UTC timestamp; Stripe Checkout unlocks the download; the link lands by email within 60 seconds. No accounts, no dashboard, no Google Sheets, no live editing, no AI narrative.

3. Technical implementation plan

Next.js on Vercel serves the upload page and the report page. A serverless Python route streams the workbook through openpyxl, builds a formula index, and runs three detectors: numeric literals inside formulas, error values and broken cross-sheet links, and cycles in the dependency graph. Findings are stored in Supabase against a scan token, and the PDF is rendered server-side and cached in Supabase Storage. Stripe Checkout takes the payment; a webhook flips the token to paid; Resend emails a signed URL.

The hardest technical risk is openpyxl on messy real files: password protection, macros, external links, shared formulas, 50k-row sheets. Handle it with a hard 60-second parse timeout, a 20MB cap, and an explicit unreadable-file message that triggers an automatic refund rather than a silent partial report.

No LLM is required for the findings. If one is used, it writes the two-sentence summary at the top of the PDF from the findings JSON and never touches the findings themselves.

4. Pricing and monetization

Two one-time products: Single Scan at $49 and Five-Pack at $199, both through Stripe Checkout at 2.9% + $0.30. No subscription at launch, because the buyer's pain is episodic and a recurring charge would kill conversion.

Blended ARPU is projected at about $64, implying roughly one Five-Pack per nine Single Scans. The Five-Pack exists to answer the monetization weakness directly: a controller who has paid for one report almost always has four more files. It is offered inside the report email, not on the landing page.

After the first sale the path is Five-Pack for repeat buyers, then a $199 pre-close bundle sold to fractional CFOs who scan client files at month-end close.

5. First-customer acquisition

One channel: direct LinkedIn outreach to controllers and finance managers, filtered by posts and job changes that signal an inherited model. Search phrases: inherited model, month-end cleanup, new controller, post-acquisition integration, took over the forecast.

The offer is a free scan of one file, with no meeting request. Deliver the PDF first, then ask for $49 for the full unlocked report. Follow up once at 48 hours with a sample attached. Volume target: 200 messages across five waves of 30, 40, 40, 40 and 50.

Secondary and organic only: two finance-ops communities plus a public inherited-model checklist page. No paid ads at launch.

Week-one goal: 30 messages sent, 5 free scans delivered, 1 paid report.

6. 30-day execution roadmap

Days 1-5 build and deploy: repo and landing page on day 1, upload and parser on day 2, reference and circular checks on day 3, PDF on day 4, Stripe and email delivery on day 5. The product is live and takeable on day 5, and no feature work happens after that unless a buyer asks for it.

Days 6-9 are dogfooding, list building and the first outreach wave. First sale target is day 9, from a free scan converted personally.

Days 10-21 run the acquisition engine: message waves, a public checklist page, a referral code, and a report template hardened by manual review of every early delivery. Every report is checked by hand until day 21.

Days 22-30 split between a fourth and fifth wave, a fractional-CFO partner angle with three free client scans, the Five-Pack upsell to every Single Scan buyer, reconciliation of Stripe payouts against TrustMRR records, and a published metrics page on day 30.

The roadmap assumes zero paid marketing. If two consecutive waves return under 5% reply rate, the next day's effort moves to community posts instead of more messages.

7. Revenue and profit scenarios

All three scenarios are projections built from stated assumptions, not verified revenue.

Conservative: 6 customers at $55 blended ARPU, $330 gross. Base: 20 customers at $64, $1,280. Upside: 45 customers at $78, $3,510. Processing fees are modelled at roughly 4.5-5.2% because the $0.30 fixed fee weighs heavier on $49 tickets. Refunds are $0, $49 and $98. Paid ad spend is $0, $0 and $120. Operating costs covering Vercel, Supabase, Resend, domain, PDF storage and occasional model calls are $18, $45 and $95.

Hackathon profit deducts refunds, chargebacks, processing fees and ad spend but not hosting. Business profit also deducts operating costs. In the base case that is roughly $1,173 hackathon profit and roughly $1,128 business profit. Every sale is priced far above delivery cost: one scan costs cents of compute.

The concept's own 60-customer target is not used here. A base of 20 is what the outreach math supports at 200 messages, 15% reply and 25% free-to-paid conversion.

8. Main risks and validation experiments

Parser failure on real files is the first risk; test 20 messy workbooks before day 6 and remove the silent-partial-report path immediately. Outreach fatigue is the second; measure reply rate at 30, 70 and 110 sends and switch channel below 5%. Perceived value at $49 is the third; A/B $49 against $99 on the landing page in week two. Refunds from missed errors are the fourth; auto-refund when a file over 200 formulas returns zero findings and hand-review everything until day 21. Episodic rather than recurring demand is the fifth and structural, which is exactly why the Five-Pack and the fractional-CFO channel exist. If neither lifts blended ARPU above $70 by day 25, the honest conclusion is a $2-3k one-time product, not a business.

9. First 24-hour action plan

Hours 1-2: domain, repo, deployed landing page with the $49 offer. Hours 3-5: upload endpoint and openpyxl parse returning a formula index. Hours 6-8: hardcode and #REF! detectors emitting findings JSON. Hours 9-12: PDF render, emailed to yourself. Hours 13-16: 40-name prospect list from LinkedIn plus a drafted free-scan message.

This brief was generated from public data about established startups, then narrowed and scored for a solo 30-day launch. Revenue and profit figures are scenario models, not verified results, guarantees or evidence of existing customers. Validate demand yourself before you spend a month on it — and check the contest rules for what counts toward your score.

Related ideas

All ideas →