Score C · 6.24/10Free to read$100K Hackathon · November 16, 2026 → December 16, 2026Nov 16, 2026

CleanupScope Diagnostic

CleanupScope turns a messy client accounting export into a branded cleanup scope report with quantified issues, estimated hours, and a fixed-fee quote for bookkeepers.

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.

Speculative6.24/10
MVP build (solo)
5 days
Launch price
$99
Sales to $1,000 revenue
11 customers
First paying customer
Direct LinkedIn outreach to solo bookkeepers and 1–5 person firms advertising catch-up bookkeeping, QuickBooks cleanup or Xero cleanup.
Who pays

Solo bookkeeper or 1–5 person bookkeeping firm in the US, UK, Canada, or Australia selling catch-up/cleanup engagements, triggered when a prospect sends QuickBooks Online or Xero exports and asks for a cleanup quote.

Biggest execution risk

Monetization potential 5.0/10

Why this could work in 30 days: one founder can ship the MVP in 5 days; the plan targets a first sale on day 11; 11 sales at $99 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

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

0/14
Entry price
$99
Solo build
5 days
30d base revenue
$1,287
30d hackathon profit
$1,235
Section 1 of 8

Product opportunity and positioning

Target customer

Solo bookkeeper or 1–5 person bookkeeping firm in the US, UK, Canada, or Australia selling catch-up/cleanup engagements, triggered when a prospect sends QuickBooks Online or Xero exports and asks for a cleanup quote.

Why now

The output attaches to a quoting moment that already exists, needs only CSV exports, and requires no accounting API, ledger write-back, tax filing or regulatory approval.

Problem

A bookkeeper must manually inspect a messy export to quote a cleanup; missed uncategorized accounts, duplicates, negative balances and opening-balance issues lead to underquoted fixed fees and scope creep.

Solution

Upload QBO/Xero CSV exports plus a bank CSV, map columns once, and a deterministic rule engine produces a branded Cleanup Scope Report PDF with an issue table, severity, estimated cleanup hours, a fixed-fee range, and an Issue Register CSV.

Differentiation · It is not a reconciliation or ledger product: it is the client-facing scoping artifact and fee range that a spreadsheet or a cleanup platform like Dext does not produce.

1. Product opportunity and positioning

CleanupScope Diagnostic sells into a moment that already has money attached: a bookkeeper has just been handed a messy QuickBooks Online or Xero export and has to say a number out loud. Today that number comes from a spreadsheet skim and a gut feel. The report changes what the bookkeeper hands back — a branded PDF with an issue register, severity per issue, estimated cleanup hours and a fixed-fee range.

The wedge is narrow on purpose. It is not reconciliation, not a ledger, not tax. Nothing is written back to the client's books, so there is no approval cycle and no accuracy liability beyond "review before you send". That is what makes it sellable inside 30 days by one person: the buyer evaluates it in one sitting, on their own data, and the output is immediately usable in a client conversation about a fee.

The positioning line is: the pre-engagement scoping report for cleanup work. The real substitutes are a blank spreadsheet, a senior's afternoon at their own hourly rate, or a capture platform such as Dext that solves receipt collection rather than quoting. The defensible angle is the artifact — branded PDF, issue register, fee range — plus the fact that the offer attaches to revenue the bookkeeper is already chasing. JEV scores this concept 62.4/100 (band C), with speed to first revenue at 8.32/10 and monetization at 5.01/10. The plan leans on the first and is built to survive the second: willingness to pay $99 for a generated report is unproven, so the 30 days run as a pricing and channel experiment with a working product underneath it, not as a build-then-hope launch.

2. Smallest sellable MVP

One upload page, three CSVs, one rule engine, one PDF, one Stripe link. If a bookkeeper can upload a QBO transaction list plus a bank CSV, get a branded PDF within a minute, and pay $99 to remove the watermark, the product is sellable. Everything else — credit dashboards, template editor, 10-pack branding — is a second-week convenience. Twelve deterministic detectors are the product's spine; one optional OpenAI summary sentence is decoration and must never be load-bearing. The MVP is deliberately CSV-only: no OAuth, no API sync, no write-back, which removes both the build delay and the trust objection.

3. Technical implementation plan

Next.js and TypeScript on Vercel, Supabase for Postgres/Auth/Storage, a pure TypeScript rule engine as a separate module so it can be unit-tested without the UI, React-PDF for the report, Stripe Checkout for money, Resend for delivery. Uploads are parsed to normalised rows in Postgres; the rule engine emits an issue register JSON; the scorer converts issue counts and severity into hours using the firm's hourly rate and markup; the PDF renderer consumes only the issue register and settings so it can be regenerated. OpenAI is called once per report for a three-sentence narrative, wrapped in a timeout so a model outage never blocks a paid report. The hard part is parser variance: QBO and Xero export formats differ by edition, locale and report type, so column-mapping presets plus a manual mapping fallback matter more than any detector. Build order: auth and credits, upload and mapping, detectors, scoring, PDF, checkout gating, delivery link.

4. Pricing and monetization

Two one-time SKUs: $99 Single Report (one PDF, one Issue Register CSV, 30-day link) and $299 Firm 10-Pack (ten credits, logo and colours, default hourly rate and markup). One-time pricing is correct here because the trigger is episodic — a bookkeeper quotes cleanups when a prospect asks, not every month. The 10-pack exists to convert a single satisfied quote into a repeat buying habit and to lift blended ARPU from $99 to roughly $143–$172. The first revenue offer is deliberately the $99 report, not the pack: a cheaper first yes validates the artifact, and the pack is offered only after a delivered report. Post-hackathon growth path is a monthly cleanup-quote subscription once repeat usage is visible.

5. First-customer acquisition

One channel: direct LinkedIn outreach to bookkeepers and 1–5 person firms advertising catch-up bookkeeping, QuickBooks cleanup or Xero cleanup in the US, UK, Canada and Australia. The message is not a pitch, it is an offer to run one free diagnostic on an anonymised export so the buyer can compare it against their own manual quote. Volume targets: 50 prospects listed by day 7, 20 DMs on day 8, 60 cumulative by day 12, and roughly 25 DMs per push day thereafter. Secondary channels are free and organic only — bookkeeping Facebook and Slack communities, cleanup-specialist bookkeepers on Upwork, and LinkedIn posts showing a sample report. Paid ads are out of scope for all 30 days, which is why no scenario carries ad spend.

6. 30-day execution roadmap

Days 1–5 are a hard build gate: shipped, buyable, self-serve by end of day 5. Days 6–7 are QA on real export formats plus the prospect list; nothing is sold before the sample PDF looks like something a firm would put its logo on. Days 8–11 are the free-diagnostic sprint and the first paid ask, with day 11 as the first-sale target — the free runs exist to generate three quotes worth chasing, not to be a service. Days 12–14 convert non-responders, ship firm branding and publish the $299 pack. Days 15–21 are volume: 30 DMs, three community posts, Upwork DMs, parser hardening and a trust page, ending in a decision on which SKU to push. Days 22–28 add a referral line, an FAQ built from real objections, free directory listings, a $149 rush-price test and a credit dashboard. Days 29–30 are a close-out push and reconciliation before the 13:00 UTC cut-off on December 16. Each day has one deliverable; if a day's deliverable slips, the following day absorbs it and outreach volume is protected, because outreach is the only thing that produces revenue.

7. Revenue and profit scenarios

All three scenarios below are projections built from stated assumptions, not verified revenue, customers or traction. Conservative: 2 Single Reports, $198 gross, $8 fees, $20 operating — hackathon profit $190, business profit $170. Base: 9 buyers (7 singles, 2 packs), $1,287 gross at $143 blended ARPU, $52 fees, $45 operating — hackathon profit $1,235, business profit $1,190. Upside: 22 buyers (14 singles, 8 packs), $3,784 gross at $172 blended ARPU, $99 refund, $159 fees, $95 operating — hackathon profit $3,526, business profit $3,431. Every scenario assumes zero paid marketing because the concept forbids ads in the first 30 days, and every scenario keeps a non-zero operating line for hosting, email and per-report model calls. The original concept's $14,950 target assumes 50 pack buyers in 30 days from cold LinkedIn DMs; that is treated here as an unvalidated ceiling, not a plan. Refunds and chargebacks are modelled at zero in conservative and base because there is no history to base a rate on — if they occur, both profit figures fall dollar for dollar.

8. Risks and validation experiments

Four risks matter. First, bookkeepers may scope a cleanup in fifteen minutes and refuse to pay $99; the test is 20 free diagnostics, and if fewer than 3 of the first 40 convert to paid, the price drops to $19 pay-per-report or the offer becomes a quote-defence bundle. Second, parser variance can produce broken uploads and refund requests; the test is five real anonymised exports before day 7, and if more than 30% need founder intervention by day 15 the product narrows to one format. Third, LinkedIn DM response may be too thin; the test is reply rate over the first 100 DMs, and below 5% the channel moves to bookkeeping communities and Upwork. Fourth, false positives destroy trust faster than a missing detector; the signal is how many flagged issues the bookkeeper deletes before sending, and above 40% deletion the noisiest detectors get cut rather than tuned. Free diagnostics are capped at three per day so support never eats selling time.

9. First 24-hour action plan

Hours 1–2: domain, Vercel, Supabase, repo, first deploy. Hours 3–5: upload page and QBO column-mapping preset. Hours 6–8: first six detectors plus an issue table on screen. Hours 9–11: PDF renderer with placeholder branding and one real sample output. Hours 12–14: Stripe products, checkout and the credit-issuing webhook, tested in test mode. Hours 15–18: 25-prospect LinkedIn list, DM script v1, profile headline rewritten. Hours 19–22: three synthetic exports through the pipeline, crash fixes, landing page copy published with the sample report. The day ends with a deployed app, a sample PDF and a prospect list — the three things that make day 8's outreach possible.

Section 2 of 8

Smallest sellable MVP

A CSV upload page, twelve deterministic issue detectors, an effort-to-hours scorer, and a branded Cleanup Scope Report PDF plus Issue Register CSV unlocked by a $99 Stripe payment.

Must-have features

  • CSV upload for QuickBooks Online and Xero transaction lists, trial balance and bank statement
  • Column-mapping presets for QBO and Xero plus manual mapping fallback
  • Twelve deterministic issue detectors (uncategorized accounts, duplicates, negative AR/AP, missing opening balances, unbalanced trial balance, stale bank items, unusual vendor spikes, and similar)
  • Effort-scoring model converting issue counts and severity into estimated cleanup hours
  • Fixed-fee range calculation using the firm's hourly rate and markup
  • Branded PDF report with firm logo, client name, issue table, severity column and next steps
  • Issue Register CSV export
  • Stripe Checkout for $99 Single Report and $299 Firm 10-Pack with report-credit gating
  • 30-day signed report link delivered by email

Explicitly out of scope

  • No QuickBooks or Xero API sync; CSV exports only
  • No ledger write-back, reconciliation or month-end close
  • No tax filing, tax advice or audit wording
  • No receipt chasing or invoice collection
  • No client portal or multi-user roles and permissions
  • No paid advertising or marketplace listings
  • No accounting-firm white-label beyond logo, colours, hourly rate and markup

Demo flow (what the buyer sees)

  1. Bookkeeper uploads a QBO transaction-list CSV and a bank CSV
  2. Maps columns using the QBO preset and confirms the preview
  3. Engine runs twelve detectors and returns an issue register with severity
  4. Scorer converts issues into estimated hours and a fixed-fee range
  5. Bookkeeper previews the branded PDF and pays $99 via Stripe
  6. Report credit is released and the PDF plus Issue Register CSV are emailed with a 30-day link
Build plan

Technical implementation plan

Stack

Next.js + TypeScript on VercelSupabase Postgres, Auth and StorageNode/TypeScript rule engine moduleReact-PDF for report renderingStripe Checkout and webhooksResend for transactional emailOpenAI API for one optional narrative summary callVitest for detector fixture tests

Architecture

Uploads are parsed into normalised transaction, balance and bank rows stored in Supabase. A pure TypeScript rule engine reads those rows, emits an issue register JSON with severity codes, and never touches the database directly. A separate scorer converts the register into hours and a fee range using firm settings, and the PDF renderer consumes only the register plus settings so any report can be regenerated identically. Stripe webhooks write report credits; a credit is consumed when a report is generated, and Resend delivers a signed 30-day URL. The optional OpenAI call sits behind a three-second timeout and is skipped silently on failure.

Build steps

  1. Scaffold Next.js + TypeScript repo, Supabase project and Vercel deploy
  2. Add Supabase Auth magic link and users/report_credits tables
  3. Build CSV upload to Supabase Storage and a parser normalising rows into Postgres
  4. Ship QBO and Xero column-mapping presets with a manual mapping fallback
  5. Implement the twelve detectors as pure functions with fixture tests
  6. Implement severity weighting and the hours/fee-range scorer
  7. Build the React-PDF report template with firm logo, colours and issue table
  8. Add Issue Register CSV generation
  9. Wire Stripe products, Checkout and the credit-issuing webhook
  10. Gate report generation on credits and add the 30-day signed link
  11. Add Resend delivery email and a regeneration endpoint
  12. Add the optional OpenAI narrative summary with timeout and fallback
  13. Deploy to production and run one paid-mode test purchase end to end

AI components

  • OpenAI chat completion generating a three-sentence narrative summary and recommended next steps from the issue register, capped at one call per report
  • No LLM is used for issue detection or effort scoring; both stay deterministic so results are reproducible
Hardest technical risk: QuickBooks Online and Xero export formats vary by edition, locale, date format and report type, so the column-mapping and normalisation layer will break on real files even when synthetic fixtures pass.
Section 3 of 8

Pricing and monetization

Single Report

Entry offer
$99 one-time

For Solo bookkeeper quoting one cleanup engagement

  • One branded Cleanup Scope Report PDF
  • Issue Register CSV
  • 30-day report link
  • Email delivery

Firm 10-Pack

$299 one-time

For Small bookkeeping firm quoting multiple cleanups

  • 10 report credits
  • Firm logo and colours
  • Default hourly rate and fee markup settings
  • Issue Register CSV exports
  • 30-day report links per report

Model: One-time report credits: $99 for a single branded report, $299 for a ten-report firm pack with branding and rate defaults.

First revenue offer: Run it free on one anonymised client export and I will send back a branded cleanup scope report with the issue list, estimated hours and a fee range within 24 hours — if it beats your manual quote, the PDF without watermark is $99.

Expected ACV: $143

Upsell path: Single Report buyers are offered the $299 Firm 10-Pack after a delivered report, and pack buyers with repeat usage are offered a monthly cleanup-quote subscription after day 30.

Target customers
50
Revenue model
$14,950
Costs
$550
Profit model
$14,400
Section 4 of 8

First-customer acquisition

Primary channel

Direct LinkedIn outreach to solo bookkeepers and 1–5 person firms advertising catch-up bookkeeping, QuickBooks cleanup or Xero cleanup.

Where to find them

  • LinkedIn search: 'catch-up bookkeeping' filtered to US, UK, Canada, Australia
  • LinkedIn search: 'QuickBooks cleanup' and 'Xero cleanup' in profile headlines
  • Upwork bookkeepers whose profiles list cleanup or catch-up projects
  • Bookkeeping Facebook groups and Slack communities for firm owners
  • LinkedIn posts tagged #bookkeeping #cleanup by solo practitioners

Exact outreach motions

  • Send 20 DMs offering a free diagnostic on one anonymised export
  • Follow up non-repliers after 48 hours with a one-line value restatement
  • Post the sample report publicly as a LinkedIn post with the free-diagnostic offer
  • DM ten Upwork cleanup specialists offering the same free run
  • Post the offer in three bookkeeping communities after day 15
  • Send the anonymised case note to twenty new prospects on day 19

Message template

Hi {name} — saw you run catch-up/cleanup work for clients. I built a tool that turns a messy QBO or Xero export into a cleanup scope report: issue list, estimated hours, and a fee range you can actually send. Want me to run it free on one anonymised client export so you can compare it to your manual quote? No signup needed.
Week-one goal: 50 prospects listed by day 7, 20 DMs sent on day 8, 5 free-diagnostic acceptances and 1 sample report published.
Section 5 of 8

30-day execution roadmap

MVP deadline: day 5First sale target: day 1130 days planned
Week 1 · Build & validate · 7 days
Day 1Foundation — Nov 16

Stand up the stack and money rails

  • Scaffold Next.js + TypeScript repo, Supabase project, deploy to Vercel.
  • Add Supabase Auth magic link and users/report_credits tables.
  • Wire Stripe Checkout in test mode for $99 and $299.

Deliverable: Deployed skeleton with login and test-mode checkout.

Day 2CSV ingestion

Upload and column mapping

  • Build upload wizard accepting transaction, trial-balance and bank CSVs.
  • Ship QBO and Xero column-mapping presets with manual fallback.
  • Persist normalised rows to Postgres and the raw file to Storage.

Deliverable: Upload and mapping working on two anonymised exports.

Day 3Rule engine

Twelve detectors plus scoring

  • Implement twelve deterministic detectors with severity codes.
  • Add effort scoring and fee-range maths using hourly rate and markup.
  • Write fixture tests for every detector.

Deliverable: Issue register JSON and fee range produced from a test export.

Day 4Report output

PDF and CSV artifacts

  • Build React-PDF report with logo, issue table, hours and fee range.
  • Add Issue Register CSV download.
  • Add optional OpenAI summary behind a three-second timeout.

Deliverable: End-to-end PDF and CSV from a single upload.

Day 5MVP liveMVP shipped

Buyable and self-serve

  • Gate report generation on Stripe-issued report credits.
  • Add 30-day signed link and Resend delivery email.
  • Deploy to production domain and run one paid-mode test purchase.

Deliverable: CleanupScope Diagnostic live and buyable. MVP deadline.

Day 6QA on real formats

Harden the parser

  • Run five synthetic and two anonymised exports through the pipeline.
  • Fix multi-currency, blank-row and date-format parse failures.
  • Produce one polished public sample report.

Deliverable: Public sample PDF plus a known-issues list.

Day 7Prospect list

Build the 50-name list

  • Rewrite LinkedIn headline around cleanup scope reports.
  • List 50 bookkeepers/firms advertising catch-up or cleanup work in US/UK/CA/AU.
  • Write three DM variants and a follow-up cadence.

Deliverable: 50-row prospect sheet and three DM scripts.

Week 2 · Launch & first sale · 7 days
Day 8First outreach

20 DMs and a public post

  • Send 20 DMs offering a free diagnostic on one anonymised export.
  • Publish the sample report on LinkedIn with the offer.
  • Log every reply and non-reply in the prospect sheet.

Deliverable: 20 DMs sent, one LinkedIn post live.

Day 9Free diagnostics

Deliver within 24 hours

  • Deliver up to three free diagnostics within 24 hours of receiving exports.
  • Record each false positive and manual correction.
  • Send 20 more DMs.

Deliverable: Three delivered PDFs and a detector bug list.

Day 10Fix and follow up

Convert free runs to interest

  • Ship the top three detector fixes from the free runs.
  • Follow up every non-replier with a one-line value restatement.
  • Ask each free user for a two-sentence testimonial.

Deliverable: Patched engine and all follow-ups sent.

Day 11First paid saleFirst sale target

Ask for money

  • Send the $99 Stripe link to every free user with a 48-hour window.
  • Offer $99 credited toward the $299 10-Pack within seven days.
  • Confirm payment lands in the TrustMRR-connected account.

Deliverable: First paid report (target).

Day 12Second touch

Widen the funnel

  • Send the sample PDF to 20 non-responders asking them to compare it to a manual quote.
  • Send 20 new DMs.
  • Publish a short before/after quote comparison post.

Deliverable: 60 cumulative DMs and two posts live.

Day 13Firm 10-Pack

Ship the higher-tier SKU

  • Ship firm settings: logo, colours, default hourly rate, fee markup.
  • Publish the $299 10-Pack checkout page.
  • Email every free user the pack offer.

Deliverable: Branded 10-Pack flow live.

Day 14Week 2 review

Measure the funnel

  • Count DMs, replies, free runs and paid reports; compute per-step conversion.
  • Kill the worst DM variant and scale the best.
  • Set a target of 10 paid reports by day 21.

Deliverable: One-page funnel metrics snapshot.

Week 3 · Repeat & systematize · 7 days
Day 15Volume outreach

30 DMs in one day

  • Send 30 DMs using the winning variant.
  • Publish a cleanup-scoping checklist as a lead magnet.
  • Reply to three bookkeeping community threads.

Deliverable: 30 DMs sent and checklist post live.

Day 16Community channels

Add organic distribution

  • Post the free-diagnostic offer in three bookkeeping Facebook or Slack communities.
  • DM ten Upwork cleanup-specialist bookkeepers.
  • Clear the queued diagnostics.

Deliverable: Three community posts and ten Upwork DMs.

Day 17Trust work

Reduce false-positive risk

  • Publish an assumptions-and-limits page stating what the report does not claim.
  • Add a one-line explanation per issue in the PDF.
  • Re-run three old reports to confirm output consistency.

Deliverable: PDF v1.1 with per-issue explanations.

Day 18Second sales push

Hours estimate versus manual quote

  • Follow up every free user with their hours estimate next to their manual quote.
  • Offer a 48-hour $99 single-report price to ten warm prospects.
  • Deliver all outstanding diagnostics.

Deliverable: Warm pipeline with a stage per prospect.

Day 19Content proof

One anonymised case note

  • Publish an anonymised case note: export to issue count to fee range.
  • Share it in two LinkedIn bookkeeping groups.
  • Send 20 DMs referencing the case note.

Deliverable: Case note live and 20 DMs sent.

Day 20Parser hardening

Cover more export types

  • Add native QBO transaction-list and Xero aged-receivables presets.
  • Add a clear 'format not supported' fallback with manual mapping.
  • Run the full fixture regression suite.

Deliverable: Parser covering five export types.

Day 21Week 3 review

Pick the offer to push

  • Recalculate funnel and the 10-Pack share of revenue.
  • Decide whether the final push sells singles or packs.
  • Send 25 DMs on the chosen offer.

Deliverable: Offer decision recorded and 25 DMs sent.

Week 4 · Scale what works · 7 days
Day 22Referral loop

Turn buyers into a channel

  • Add a referral line to every delivered report email.
  • Give one free credit per referred paying firm.
  • Email past buyers the referral offer.

Deliverable: Referral mechanic live.

Day 23Objection handling

Answer what buyers actually said

  • Collect the top five objections from 21 days of replies.
  • Publish a short FAQ answering each one.
  • Update the DM and landing page copy with those answers.

Deliverable: FAQ page and revised copy live.

Day 24Free distribution

Listings and groups

  • Submit to three free bookkeeping directories or newsletters.
  • Post the sample report in two LinkedIn groups.
  • Send 25 DMs.

Deliverable: Three free listings submitted, 25 DMs sent.

Day 25Second cohort

A fresh free-diagnostic round

  • Run five new free diagnostics for new prospects.
  • Target two paid conversions within 48 hours.
  • Log founder minutes per free run to protect selling time.

Deliverable: Five new free runs and a time-cost baseline.

Day 26Pricing test

Test a rush price

  • Offer a $149 single report with 12-hour rush delivery to ten prospects.
  • Keep $99 as the standard offer for everyone else.
  • Compare conversion and uptake.

Deliverable: Pricing test results logged.

Day 27Credits and history

Make packs worth buying

  • Show remaining credits and report history in the dashboard.
  • Email firms with unused credits a reminder.
  • Add one-click re-run on a new export.

Deliverable: Dashboard with credits and history shipped.

Day 28Week 4 review

Actual versus projection

  • Compare actual revenue to the base case and update assumptions.
  • List the three highest-converting message angles.
  • Plan the final 48-hour close push.

Deliverable: Projection-versus-actual table.

Days 29-30 · Consolidate · 2 days
Day 29Closing push

Last-call outreach

  • DM every warm prospect with a 24-hour close offer.
  • Email all free users a last-call 10-Pack price.
  • Deliver every outstanding report the same day.

Deliverable: Final outreach batch sent.

Day 30Lock and report — Dec 16

Reconcile before 13:00 UTC

  • Stop selling before the 13:00 UTC cutoff and reconcile payments, refunds and fees.
  • Export the funnel and revenue log for TrustMRR.
  • Write a post-launch note with the next 30-day priorities.

Deliverable: Final revenue and funnel report submitted.

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 scenarios: one-time report sales only, no subscriptions inside the 30 days (applies to Conservative, Base, Upside).
  • Conservative: 2 Single Reports at $99, no pack sales, zero refunds and zero chargebacks.
  • Base: 9 buyers — 7 Single Reports at $99 and 2 Firm 10-Packs at $299 — giving a blended ARPU of $143.
  • Upside: 22 buyers — 14 Single Reports and 8 Firm 10-Packs — giving a blended ARPU of $172.
  • Payment fees modelled at 4.0–4.2% blended, assuming Stripe at 2.9% + $0.30 plus international card and currency-conversion cost on non-US buyers.
  • ad_spend_usd is 0 in every scenario: paid ads, marketplace listings and partnership integrations are out of scope for all 30 days.
  • Upside includes one voluntary $99 refund; Conservative and Base model zero refunds because there is no history to estimate a rate.
  • Operating costs cover Vercel, Supabase, Resend, domain and per-report OpenAI summary calls: $20 at 2 reports, $45 at 9, $95 at 22.
  • The concept's $14,950 / 50-customer target is treated as an unvalidated ceiling, not a forecast.
  • All figures are projections from stated assumptions, not verified revenue, customers or traction.

Conservative

2 customers
$198
collected revenue · avg payment $99
Refunds
$0
Chargebacks
$0
Payment fees
$8
Paid marketing
$0
Hackathon profit
$190
Operating costs
$20
Business profit
$170

Base case

9 customers
$1,287
collected revenue · avg payment $143
Refunds
$0
Chargebacks
$0
Payment fees
$52
Paid marketing
$0
Hackathon profit
$1,235
Operating costs
$45
Business profit
$1,190

Optimistic

22 customers
$3,784
collected revenue · avg payment $172
Refunds
$99
Chargebacks
$0
Payment fees
$159
Paid marketing
$0
Hackathon profit
$3,526
Operating costs
$95
Business profit
$3,431

Model your own numbers

Runs in your browser · nothing is saved
Collected revenue
$891
Payment fees (3.9%)
$35
Paid marketing
$0
Hackathon profit
$856
Business profit
$811

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.

These are projections built from stated assumptions, not verified revenue, customers or traction. Actual results may differ materially; refunds and chargebacks reduce profit dollar for dollar.

Section 7 of 8

Main risks and validation experiments

Risk 1Bookkeepers scope a cleanup in fifteen minutes and refuse to pay $99 for a generated report.

Mitigation

Position the output as a client-facing artifact and fee defence, not as analysis, and show estimated hours beside the bookkeeper's manual quote.

Validation experiment

Run 20 free diagnostics in week 2 and ask each user for a paid $99 conversion within 48 hours.

Kill criteria

If fewer than 3 of the first 40 free diagnostics convert to paid, drop to a $19 pay-per-report price or reposition as a quote-defence bundle.

Risk 2QuickBooks and Xero export formats vary by edition and locale, so real uploads break the parser and trigger refund requests.

Mitigation

Ship manual column mapping as a fallback and keep a fixture regression suite for every supported export type.

Validation experiment

Push five real anonymised exports through the pipeline before day 7 and log every failure.

Kill criteria

If more than 30% of uploads need founder intervention by day 15, narrow the product to the single QuickBooks Online transaction-list format.

Risk 3LinkedIn DM response rates are too low to produce paying customers in 30 days.

Mitigation

Run three channels in parallel — LinkedIn DMs, bookkeeping communities and Upwork cleanup specialists — and kill the weakest quickly.

Validation experiment

Measure reply rate over the first 100 DMs sent between days 8 and 14.

Kill criteria

Below a 5% reply rate after 100 DMs, move primary effort to bookkeeping Facebook/Slack communities and Upwork outreach.

Risk 4False positives erode trust and the report is seen as noise rather than a quote asset.

Mitigation

Label severity per issue, add a one-line explanation per issue, and state clearly that the report is a draft for review, not an opinion.

Validation experiment

Track how many flagged issues bookkeepers delete before sending each report to a client.

Kill criteria

If bookkeepers delete more than 40% of flagged issues across the first ten reports, cut the noisiest detectors instead of tuning them.

Risk 5Free diagnostics consume founder hours and crowd out selling time.

Mitigation

Cap free runs at three per day with a 24-hour SLA and automate delivery end to end.

Validation experiment

Log founder minutes per free run from day 9 onward.

Kill criteria

If free runs exceed 90 minutes of founder time per day, replace the free tier with a $19 prepaid sample report.

Section 8 of 8

Hour-by-hour: the first 24 hours

Hours 1-2

  • Register domain, create Vercel and Supabase projects, push the repo.
  • Deploy the default Next.js app to production.

Output: Live deployed URL with a placeholder landing page.

Hours 3-5

  • Build the CSV upload page with Supabase Storage.
  • Ship the QuickBooks Online transaction-list column-mapping preset and preview table.

Output: Working upload and mapping page on a real export.

Hours 6-8

  • Implement the first six detectors as pure functions.
  • Render the issue table with severity and counts on screen.

Output: Issue register visible in the UI from a test CSV.

Hours 9-11

  • Build the React-PDF report with placeholder branding.
  • Wire the hours and fee-range scorer into the template.

Output: First sample Cleanup Scope Report PDF.

Hours 12-14

  • Create $99 and $299 Stripe products and Checkout links.
  • Add the webhook that issues report credits and test a purchase in test mode.

Output: Test-mode payment that unlocks a report credit.

Hours 15-18

  • Rewrite the LinkedIn headline and build a 25-prospect list.
  • Draft DM variant v1 and the 48-hour follow-up line.

Output: 25-row prospect sheet and DM script.

Hours 19-22

  • Run three synthetic exports through the pipeline and fix crashes.
  • Publish landing page copy with the sample PDF embedded.

Output: Landing page live with a downloadable sample report.

Complete report

Full 30-day launch report

Read the full generated report

1. Product opportunity and positioning

CleanupScope Diagnostic sells into a moment that already has money attached: a bookkeeper has just been handed a messy QuickBooks Online or Xero export and has to say a number out loud. Today that number comes from a spreadsheet skim and a gut feel. The report changes what the bookkeeper hands back — a branded PDF with an issue register, severity per issue, estimated cleanup hours and a fixed-fee range.

The wedge is narrow on purpose. It is not reconciliation, not a ledger, not tax. Nothing is written back to the client's books, so there is no approval cycle and no accuracy liability beyond "review before you send". That is what makes it sellable inside 30 days by one person: the buyer evaluates it in one sitting, on their own data, and the output is immediately usable in a client conversation about a fee.

The positioning line is: the pre-engagement scoping report for cleanup work. The real substitutes are a blank spreadsheet, a senior's afternoon at their own hourly rate, or a capture platform such as Dext that solves receipt collection rather than quoting. The defensible angle is the artifact — branded PDF, issue register, fee range — plus the fact that the offer attaches to revenue the bookkeeper is already chasing. JEV scores this concept 62.4/100 in band C, with speed to first revenue at 8.32/10 and monetization at 5.01/10. The plan leans on the first and is built to survive the second: willingness to pay $99 for a generated report is unproven, so the 30 days run as a pricing and channel experiment with a working product underneath it, not as a build-then-hope launch.

2. Smallest sellable MVP

One upload page, three CSVs, one rule engine, one PDF, one Stripe link. If a bookkeeper can upload a QBO transaction list plus a bank CSV, get a branded PDF in under a minute, and pay $99 to remove the watermark, the product is sellable. Everything else — credit dashboards, template editor, pack branding — is a second-week convenience. Twelve deterministic detectors are the product's spine; the optional model-written summary sentence is decoration and must never be load-bearing. The MVP is deliberately CSV-only: no OAuth, no API sync, no write-back, which removes both the build delay and the trust objection in one decision.

3. Technical implementation plan

Next.js and TypeScript on Vercel, Supabase for Postgres, Auth and Storage, a pure TypeScript rule engine as a separate module so it can be tested without the UI, React-PDF for the report, Stripe Checkout for money, Resend for delivery. Uploads are parsed into normalised rows in Postgres; the rule engine emits an issue register; the scorer converts issue counts and severity into hours using the firm's hourly rate and markup; the PDF renderer consumes only the register and settings so any report can be regenerated byte-for-byte. OpenAI is called at most once per report for a three-sentence narrative, wrapped in a timeout so a model outage never blocks a paid report. The hard part is parser variance: QBO and Xero exports differ by edition, locale and report type, so column-mapping presets and a manual fallback matter more than any detector. Build order: auth and credits, upload and mapping, detectors, scoring, PDF, checkout gating, delivery link.

4. Pricing and monetization

Two one-time SKUs: $99 Single Report (PDF, Issue Register CSV, 30-day link) and $299 Firm 10-Pack (ten credits, logo and colours, default hourly rate and markup). One-time pricing is correct because the trigger is episodic — a bookkeeper quotes cleanups when a prospect asks, not every week. The pack exists to convert one satisfied quote into a repeat buying habit and lifts blended ARPU from $99 toward $143–$172. The first revenue offer is deliberately the $99 report, not the pack: a cheaper first yes validates the artifact, and the pack is only offered after a delivered report has been seen.

5. First-customer acquisition

One primary channel: direct LinkedIn outreach to bookkeepers and 1–5 person firms advertising catch-up bookkeeping, QuickBooks cleanup or Xero cleanup in the US, UK, Canada and Australia. The message is not a pitch; it is an offer to run one free diagnostic on an anonymised export so the buyer can compare it against their own manual quote. Volume targets: 50 prospects listed by day 7, 20 DMs on day 8, 60 cumulative by day 12, roughly 25 per push day thereafter. Secondary channels are free and organic only — bookkeeping Facebook and Slack communities, cleanup-specialist bookkeepers on Upwork, and LinkedIn posts showing a sample report. Paid ads are out of scope for all 30 days, which is why no scenario carries ad spend.

6. 30-day execution roadmap

Days 1–5 are a hard build gate: shipped, buyable and self-serve by end of day 5. Days 6–7 are QA on real export formats plus the prospect list; nothing is sold before the sample PDF looks like something a firm would put its logo on. Days 8–11 are the free-diagnostic sprint and the first paid ask, with day 11 as the first-sale target — free runs exist to generate three quotes worth chasing, not to be a service. Days 12–14 convert non-responders, ship firm branding and publish the $299 pack. Days 15–21 are volume: 30 DMs, three community posts, Upwork DMs, parser hardening and a trust page, ending in a decision on which SKU to push. Days 22–28 add a referral line, an FAQ built from real objections, free directory listings, a $149 rush-price test and a credit dashboard. Days 29–30 close out and reconcile before the 13:00 UTC cutoff on December 16. Each day has one deliverable; if a deliverable slips, the next day absorbs it and outreach volume is protected, because outreach is the only activity that produces revenue.

7. Revenue and profit scenarios

Every number in this section is a projection built from stated assumptions, not verified revenue, customers or traction. Under the contest definition, hosting, domains, software and model delivery costs are not deducted from hackathon profit; they are deducted in business profit.

Conservative: 2 Single Reports, $198 gross at $99 ARPU, $8 payment fees, $20 operating cost. Hackathon profit $190, business profit $170.

Base: 9 buyers — 7 Single Reports and 2 Firm 10-Packs — $1,287 gross at $143 blended ARPU, $52 payment fees, $45 operating cost. Hackathon profit $1,235, business profit $1,190.

Upside: 22 buyers — 14 singles and 8 packs — $3,784 gross at $172 blended ARPU, $99 refund, $159 payment fees, $95 operating cost. Hackathon profit $3,526, business profit $3,431.

All three assume zero paid marketing because paid ads are out of scope for the 30 days. The concept's original $14,950 target assumes 50 pack buyers from cold LinkedIn DMs; that is treated here as an unvalidated ceiling. Refunds and chargebacks are modelled at zero in two scenarios because there is no history to estimate a rate, and they reduce both profit figures dollar for dollar if they occur.

8. Risks and validation experiments

Four risks matter and each has a dated test. First, bookkeepers may scope a cleanup in fifteen minutes and refuse to pay $99; the test is 20 free diagnostics in week 2, and if fewer than 3 of the first 40 convert, the price drops to $19 per report or the offer becomes a quote-defence bundle. Second, parser variance can break real uploads; the test is five real anonymised exports before day 7, and if more than 30% need founder intervention by day 15 the product narrows to the QuickBooks Online transaction-list format only. Third, LinkedIn response may be too thin; the test is reply rate across the first 100 DMs, and below 5% the channel moves to bookkeeping communities and Upwork. Fourth, false positives destroy trust faster than a missing detector; the signal is how many flagged issues a bookkeeper deletes before sending, and above 40% deletion the noisiest detectors get cut rather than tuned. Free diagnostics are capped at three per day so support never eats selling time.

9. First 24-hour action plan

Hours 1–2: domain, Vercel, Supabase, repo, first deploy. Hours 3–5: upload page and the QBO column-mapping preset. Hours 6–8: first six detectors and an on-screen issue table. Hours 9–11: PDF renderer with placeholder branding and one real sample output. Hours 12–14: Stripe products, Checkout and the credit-issuing webhook, tested in test mode. Hours 15–18: a 25-prospect LinkedIn list, DM script v1, rewritten profile headline. Hours 19–22: three synthetic exports through the pipeline, crash fixes, landing page published with the sample report embedded. The day ends with a deployed app, a sample PDF and a prospect list — the three things that make day 8's outreach possible.

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 →