Score B · 6.77/10Free to read$100K Hackathon · November 16, 2026 → December 16, 2026Nov 15, 2026

Crash Course Press

Turns a licensing-exam instructor's own notes plus the official exam content outline into a branded, print-ready practice booklet and timed cohort quiz links in under 30 minutes.

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.

Promising6.77/10
MVP build (solo)
5 days
Launch price
$149
Sales to $1,000 revenue
7 customers
First paying customer
Direct outreach by email and DM to solo licensing-exam instructors in one exam family, offering a free sample booklet built from files they already have.
Who pays

Solo or 2-3 person operators of licensing-exam crash courses and tutoring practices in one exam family (launch vertical: a single state licensing exam), running 5-200 students per cohort and paying by card from their own business.

Biggest execution risk

Operational simplicity 5.3/10

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; 7 sales at $149 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
$149
Solo build
5 days
30d base revenue
$1,788
30d hackathon profit
$1,716
Section 1 of 8

Product opportunity and positioning

Target customer

Solo or 2-3 person operators of licensing-exam crash courses and tutoring practices in one exam family (launch vertical: a single state licensing exam), running 5-200 students per cohort and paying by card from their own business.

Why now

Managed document parsing is cheap and reliable enough to read messy instructor PDFs, and frontier LLMs now produce student-grade multiple-choice items when grounded in an outline and few-shot styled on the instructor's own questions — a same-day booklet is now one person's job.

Problem

Writing a cohort's worth of exam-style questions with answer keys and rationales eats an unbilled weekend, free question banks are stale or off-blueprint, and a freelance item writer is expensive with a two-week turnaround.

Solution

Instructor uploads notes plus the official content outline; the product weights the blueprint, samples their phrasing, generates tagged multiple-choice items with rationales, and returns a branded PDF booklet plus timed quiz links with per-student score reports.

Differentiation · Not a learning platform, question marketplace or student subscription: it is an item-writing and typesetting service for the person who already sells the course, outputting a resellable artifact in the instructor's voice rather than generic AI text.

1. Product opportunity and positioning

Licensing-exam instructors already own the content and already sell the course. What they do not have is a weekend. Every cohort needs 100-150 exam-style questions with four options, an answer key and a rationale — in the instructor's own voice and aligned to the current official content outline. Today that means writing them by hand, reusing a stale bank, or hiring an item writer with a two-week turnaround.

Crash Course Press collapses that into one upload. Notes in, blueprint-tagged booklet out: a branded, print-ready PDF plus timed cohort quiz links with score reports. The buyer is the solo or 2-3 person operator of a licensing-exam crash course in one exam family, paying by card from their own business, findable through their own course website, YouTube prep channel, instructor Facebook groups and the subreddits where they complain about question banks.

Positioning is artifact-first. This is not a learning platform and not a question marketplace. It is an item-writing and typesetting service that replaces a weekend task, and the output looks like the instructor's material because generation is anchored to their phrasing, stem length and difficulty.

JEV scores the concept 67.7/100 (band B, confidence 0.72). The profile sets three operating rules for the next 30 days. Speed to first revenue (8.3) and reachable customers (7.86) are the strengths, so outreach starts on day 1 and the first sale is targeted for day 11. Monetization (5.79) is middling, so price stays as a one-time charge with no billing machinery and no free plan that gives away the artifact. Operational simplicity (5.3) is the weakest criterion, so the product ships exactly one output artifact and support stays one inbox with a stated 24-hour turnaround — no live chat, no proctoring, no LMS promises.

The wedge is deliberately narrow: one exam family, one buyer type, one deliverable. Breadth is the failure mode here. An instructor evaluating a generic AI study tool asks whether it is trustworthy; an instructor evaluating this asks only whether the booklet is usable on Monday, which is a same-day demo answer built from their own files.

Section 2 of 8

Smallest sellable MVP

Upload notes plus an official exam content outline, receive a watermarked 10-question preview free, pay $149, download a branded print-ready booklet with answer key and rationales plus a timed quiz link.

Must-have features

  • Dual upload: instructor notes (PDF, DOCX or pasted text) and official content outline, text-extracted server-side
  • Blueprint parser converting the outline into a weighted domain tree with a target question count per domain
  • Style sampler pulling 8 existing questions from the notes as few-shot anchors for stem length, distractor patterns and tone
  • Item generator producing blueprint-tagged questions with four options, one correct answer and a 2-3 sentence rationale
  • Validity check rejecting duplicate stems, missing answers and all-of-the-above options
  • Branded booklet renderer to print-ready PDF: cover, questions by domain, answer key and rationale appendix
  • Timed cohort quiz links with countdown timer and per-student score report emailed to the instructor
  • Stripe checkout that gates full generation and PDF download after a free watermarked 10-question preview

Explicitly out of scope

  • No LMS, gradebook or long-lived student accounts
  • No adaptive learning or spaced repetition
  • No shared marketplace or public question-bank library
  • No mobile app in the first 30 days
  • No essay or open-response grading
  • No proctoring, identity verification or exam-security guarantees
  • No subscription billing or usage metering
  • No multi-instructor team seats

Demo flow (what the buyer sees)

  1. Founder receives an instructor's notes PDF and outline PDF by email or upload link
  2. Documents are parsed and text is stored against a single job
  3. Blueprint parser weights domains and sets target counts
  4. Style sampler extracts 8 anchor questions
  5. Generator produces the preview set and the full set
  6. Validity check flags and regenerates rejected items
  7. Booklet renders to PDF with logo, business name and contact page
  8. Instructor pays $149 via Stripe Payment Link
  9. Full booklet PDF and timed quiz links unlock immediately
  10. Score reports from the quiz route back to the instructor's inbox
Build plan

Technical implementation plan

Stack

Next.js on VercelSupabase Postgres plus object storage for uploads and generated PDFsManaged document-parsing / OCR API for PDF and DOCX extractionFrontier LLM API for blueprint mapping, style sampling and item generationPlaywright HTML-to-PDF renderingStripe Payment Links and webhooksResend for score-report and download emailUpstash Redis queue for generation jobs

Architecture

A single Next.js app on Vercel handles upload, job status and download. Files land in Supabase storage; a parsing API returns text. A background job calls the LLM three times in sequence: outline to weighted blueprint JSON, notes to 8 style anchors, then item generation in domain-sized batches of 15 so a failure costs one batch rather than the job. Items are validated, then written to Postgres. Playwright renders a server-side HTML template to the print-ready PDF, stored in Supabase and served through signed URLs gated on a Stripe payment record. Quiz links are server-rendered routes reading questions from Postgres; submissions email a score report via Resend.

Build steps

  1. Scaffold Next.js, Supabase project and storage buckets for uploads and PDFs
  2. Build upload route with parser fallback from PDF to OCR, storing raw and cleaned text
  3. Write and hand-check the blueprint prompt against two real outlines until the domain tree matches the PDF structure
  4. Implement the style sampler and confirm anchors come from the instructor's notes, not the outline
  5. Implement batched item generation with JSON schema validation and a retry-on-invalid loop
  6. Add the validity check: near-duplicate stem detection, answer-count check, banned option patterns
  7. Build the HTML booklet template and Playwright PDF route with a logo and contact page
  8. Add Stripe Payment Link plus webhook that flips a job to paid and unlocks download
  9. Build the free watermarked 10-question preview path
  10. Build timed quiz pages and the score-report email
  11. Add per-job token and render cost logging to keep a 150-question booklet under $6
  12. Harden: upload size limits, parsing failure messaging, signed URL expiry, refund webhook

AI components

  • LLM call 1: outline to weighted domain tree with target question counts
  • LLM call 2: extract 8 anchor questions and a style profile from instructor notes
  • LLM call 3: batched item generation, blueprint-tagged, four options plus rationale
  • LLM call 4: repair pass rewriting items rejected by the validity check
Hardest technical risk: Document parsing on scanned, photographed or layout-heavy instructor notes — if extraction returns garbled text, every downstream step degrades and the founder discovers it only after promising a 24-hour sample.
Section 3 of 8

Pricing and monetization

Single Cohort

Entry offer
$149 one-time

For An instructor preparing one cohort's practice booklet

  • Up to 150 blueprint-tagged questions with answer key and rationales
  • Branded print-ready PDF booklet
  • Timed cohort quiz links with score reports

Season Pass

$399 one-time

For An instructor running multiple cohorts or exam variants in one season

  • Up to four cohort booklets
  • Reusable brand template
  • Question bank export as CSV

Model: Two one-time charges in the $19-$499 band, with a free watermarked preview that proves quality before payment; no subscription.

First revenue offer: Send a 10-question sample booklet built from the instructor's own notes, free and watermarked, with one line: 'Here is the first 10 of 150 — the full cohort booklet with answer key, rationales and a timed quiz link is $149, delivered same day.'

Expected ACV: $149

Upsell path: Single Cohort buyers running more than one cohort in a season are offered the $399 Season Pass with a credit for their first purchase; after that, the CSV question-bank export and a second exam variant within the same family drive repeat one-time purchases.

Target customers
12
Revenue model
$1,788
Costs
$65
Profit model
$1,723
Section 4 of 8

First-customer acquisition

Primary channel

Direct outreach by email and DM to solo licensing-exam instructors in one exam family, offering a free sample booklet built from files they already have.

Where to find them

  • Course websites of independent instructors teaching the target licensing exam, with the contact form or listed email
  • YouTube prep channels for that exam, where the channel owner publishes a business email
  • Facebook groups for instructors and school owners in that licensing vertical
  • Subreddits where instructors and school owners discuss course material and question banks
  • State licensing school directories listing approved course providers
  • LinkedIn searches for 'instructor' plus the exam name at small provider companies

Exact outreach motions

  • Day 1: build a 100-instructor list in one exam family with name, exam, contact and visible cohort size
  • Day 2-3: send 40 short emails or DMs leading with the weekend lost to writing questions and the free 10-question sample offer
  • Day 4-6: send 40 more, and post one helpful comment in three threads where instructors ask for question-writing help
  • Deliver every sample within 24 hours of receiving files, then reply with the $149 checkout link in the same thread
  • 72-hour follow-up with the sample attached and one question: when does your next cohort start
  • Ask every paying instructor for one introduction to another instructor running the same exam

Message template

Hi [name] — you teach [exam] and I assume you write your own practice questions each cohort. I built a tool that takes your notes plus the official content outline and returns a branded practice booklet: blueprint-tagged questions, answer key, rationales, plus a timed quiz link. Send me one notes PDF and the outline and I'll generate the first 10 questions free, back to you within 24 hours. If it's usable, the full 150-question booklet is $149, same day.
Week-one goal: 100 instructors listed, 80 contacted, 8 sample booklets delivered from real files, 2 paid Single Cohort booklets.
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 1List and scaffold

Build the list and stand up the shell

  • Build a 100-instructor list in one licensing exam family from course sites and YouTube prep channels.
  • Scaffold Next.js on Vercel with Supabase storage for uploads.
  • Create Stripe account and connect the payment account to TrustMRR before any charge.

Deliverable: 100-row instructor sheet plus a deployed upload shell.

Day 2Document ingest

Parse notes and outline reliably

  • Build upload for notes PDF, DOCX and pasted text plus a separate outline upload.
  • Wire the managed parsing API with OCR fallback.
  • Store raw and cleaned text per job.

Deliverable: Upload flow returning clean text for three real documents.

Day 3Blueprint and style

Turn the outline into weighted domains

  • Implement the blueprint parser producing a weighted domain tree with target counts.
  • Implement style sampler pulling 8 anchor questions from notes.
  • Hand-check both against two real outlines.

Deliverable: Blueprint JSON view verified on two outlines.

Day 4Item generation

Generate and validate items

  • Generate items in batches of 15 with JSON schema validation and a repair pass.
  • Add validity check for duplicate stems, missing answers and all-of-the-above options.
  • Log token cost per 150-question job.

Deliverable: A 150-question run on a real outline costing under $6.

Day 5Booklet and paywallMVP shipped

Ship the sellable artifact

  • Build the branded HTML booklet template and Playwright PDF route.
  • Add free watermarked 10-question preview.
  • Gate full PDF behind a $149 Stripe Payment Link with webhook unlock.

Deliverable: MVP live and one full booklet generated end to end.

Day 6Quiz links and samples

Add quiz links, produce free samples

  • Add timed quiz routes with countdown and Resend score reports.
  • Generate three free sample booklets from public instructor notes.
  • Test the whole flow on a fresh job as a customer would.

Deliverable: Three sample booklets with live quiz links.

Day 7Outreach wave 1

Contact 40 instructors

  • Send 40 short emails or DMs with the weekend-lost hook and free sample offer.
  • Comment helpfully in three instructor threads about question writing.
  • Log replies in the instructor sheet.

Deliverable: 40 contacted, 3 samples delivered.

Week 2 · Launch & first sale · 7 days
Day 8Sample delivery

Hit the 24-hour promise

  • Deliver every requested sample within 24 hours.
  • Fix parsing failures found on real files.
  • Follow up day-7 non-responders once.

Deliverable: 5 samples delivered, parser fixes shipped.

Day 9Proof assets

Make the pitch self-serve

  • Record a 90-second demo showing a real booklet and quiz link.
  • Publish a pricing page with both tiers.
  • Add a sample-page link to the outreach message.

Deliverable: Demo video and live pricing page.

Day 10Outreach wave 2

Reach 80 total

  • Contact 40 more instructors across Facebook groups and prep channels.
  • DM instructors who posted about outdated question banks.
  • Answer every reply the same day.

Deliverable: 80 total contacted, reply log updated.

Day 11First saleFirst sale target

Convert samples to paid

  • Send the $149 checkout link to every sample recipient.
  • Offer same-day delivery for a cohort starting this week.
  • Confirm the first payment clears through the connected account.

Deliverable: First paid Single Cohort booklet.

Day 12Quality pass

Make output instructor-grade

  • Spot-check 40 generated items against the outline.
  • Add instructor flag-and-regenerate on individual items.
  • Publish a one-line accuracy caveat in the booklet footer.

Deliverable: Quality fix list closed.

Day 13Follow-up

Recover quiet threads

  • Send 72-hour follow-ups with the sample attached.
  • Ask one question: next cohort start date.
  • Trigger a starter-page update: licensed exam branding.
  • Nudge a proctor on manual edit.
  • Not a task, replace: close stale threads older than 7 days with no reply.

Deliverable: 10 follow-ups sent and logged.

Day 14Referrals

Mine the first buyers

  • Ask every buyer for one introduction to another instructor on the same exam.
  • Log referral source per customer.
  • Ask permission to quote any positive reply.

Deliverable: 3 referral asks sent.

Week 3 · Repeat & systematize · 7 days
Day 15Season Pass

Add the larger tier

  • Ship the $399 Season Pass with credit for prior purchase.
  • Add CSV question bank export.
  • Test four-booklet generation on one account.

Deliverable: Season Pass tier live and purchasable.

Day 16Score reports

Improve the quiz deliverable

  • Add per-student domain breakdown to the instructor score report.
  • Email a cohort summary after the quiz closes.
  • Fix any quiz timer bugs found in real use.

Deliverable: Score report v2 shipped.

Day 17Outreach wave 3

Reach 120 total

  • Contact 40 instructors running the same exam in a second state variant.
  • Post the sample offer in two relevant threads.
  • Update the sheet with cohort start dates.

Deliverable: 120 total contacted.

Day 18Variant support

Handle a second outline

  • Support swapping the outline while keeping the style anchor.
  • Test one variant outline end to end.
  • Document the variant workflow in the FAQ.

Deliverable: Variant outline test passing.

Day 19Turnaround

Overnight sample delivery

  • Add a job queue so samples render without supervision.
  • Add templated delivery email with booklet and quiz link.
  • Measure actual turnaround on the last ten jobs.

Deliverable: Sub-12-hour sample turnaround measured.

Day 20Social proof

Get quotable feedback

  • Ask two paying instructors for a written testimonial with a quote permission line.
  • Add quotes and booklet page screenshots to the pricing page.
  • Publish one anonymised before/after booklet excerpt with permission.

Deliverable: One to two testimonials live on the page.

Day 21Midpoint review

Read the funnel

  • Count contacted, replied, sampled, paid and refunded.
  • Kill the weakest message angle and rewrite the opener.
  • Set the paid target for the final nine days.

Deliverable: Funnel sheet with a revised message.

Week 4 · Scale what works · 7 days
Day 22December cohorts

Target imminent cohort starts

  • Contact instructors whose next cohort starts before December 16.
  • Offer a 24-hour turnaround slot.
  • Send 10 new sample offers.

Deliverable: 10 new sample requests in the queue.

Day 23Quality fixes

Close top complaints

  • Fix the three most common item-quality complaints.
  • Add rationale length control.
  • Re-test the validity check against flagged items.

Deliverable: Quality fixes deployed.

Day 24Lead capture

Reduce manual sample work

  • Publish a self-serve free 10-question sample page with email capture.
  • Add an automatic delivery email.
  • Track conversion from sample page to checkout.

Deliverable: Lead page live with conversion logging.

Day 25Close the pipe

Touch every open thread

  • Send the checkout link to every unpaid sample recipient.
  • Follow up once more on season-pass eligibility.
  • Log outcome for each thread.

Deliverable: All open threads touched and classified.

Day 26Delivery backlog

Zero outstanding orders

  • Deliver every paid booklet within 24 hours.
  • Collect one testimonial from the newest buyer.
  • Confirm quiz links work for each order.

Deliverable: Zero delivery backlog.

Day 27Final push

Reach 150 contacted

  • Contact 30 new instructors with the December 16 urgency line.
  • Repost the sample offer in two communities.
  • Answer all replies same day.

Deliverable: 150+ total contacted.

Day 28Upsell

Move buyers to Season Pass

  • Offer the Season Pass with credit to every Single Cohort buyer.
  • Send the CSV export to all buyers.
  • Confirm upsell payments clear.

Deliverable: Upsell offers sent and logged.

Days 29-30 · Consolidate · 2 days
Day 29Reconciliation

Verify the money

  • Confirm every payment landed in the TrustMRR-connected account.
  • Match refunds and processing fees against the ledger.
  • Honor any refund request within 24 hours.

Deliverable: Reconciliation sheet with net figures.

Day 30Close

Tally and submit

  • Final count of customers, refunds and fees.
  • Record operating costs for the business-profit view.
  • Submit the hackathon entry with the reconciliation attached.

Deliverable: Final submission package.

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

  • Conservative/Base/Upside: payment processing assumed at a blended 4% of collected revenue (Stripe 2.9% + $0.30 domestic, with some international card mix).
  • Base: 12 paying instructors at $149 average is roughly 13% of about 90 contacted instructors plus a few inbound sample requests.
  • Conservative: 4 customers, assuming cold outreach replies underperform and only sample recipients with a cohort starting inside the window convert.
  • Upside: 20 customers at a $211 blended average — 15 Single Cohort at $149 and 5 Season Pass at $399 — after one upsell wave.
  • No paid marketing in Conservative or Base; outreach is organic email and DM, so ad_spend is 0.
  • Upside includes $250 of paid marketing, limited to sponsored posts in two instructor communities, treated as a test not a channel.
  • Operating costs cover Vercel, Supabase, the parsing API, LLM inference and PDF rendering: about $6 of inference per 150-question booklet plus $10-20 of fixed tooling and domain.
  • Refunds are assumed near zero because the free preview de-risks the purchase; upsell assumes 2 refunds on the larger mix.
  • All figures are projections from stated assumptions, not verified revenue, customers or traction.

Conservative

4 customers
$596
collected revenue · avg payment $149
Refunds
$0
Chargebacks
$0
Payment fees
$24
Paid marketing
$0
Hackathon profit
$572
Operating costs
$40
Business profit
$532

Base case

12 customers
$1,788
collected revenue · avg payment $149
Refunds
$0
Chargebacks
$0
Payment fees
$72
Paid marketing
$0
Hackathon profit
$1,716
Operating costs
$70
Business profit
$1,646

Optimistic

20 customers
$4,220
collected revenue · avg payment $211
Refunds
$350
Chargebacks
$0
Payment fees
$169
Paid marketing
$250
Hackathon profit
$3,451
Operating costs
$140
Business profit
$3,311

Model your own numbers

Runs in your browser · nothing is saved
Collected revenue
$1,788
Payment fees (3.9%)
$70
Paid marketing
$0
Hackathon profit
$1,718
Business profit
$1,648

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.

All scenario figures are projections built from the stated assumptions. They are not verified revenue, existing customers or real traction, and they carry no guarantee.

Section 7 of 8

Main risks and validation experiments

Risk 1Generated items contain factual or blueprint errors that embarrass the instructor and trigger refunds.

Mitigation

Validity check plus a footer accuracy caveat stating the instructor is the final reviewer, and a one-click flag-and-regenerate on any item.

Validation experiment

Have two instructors review 40 generated items each and count how many need editing before use.

Kill criteria

If more than 30% of items need substantive rewriting rather than light edits, reposition as a drafter tool at a lower price or stop.

Risk 2Instructors will not share real notes and an official outline with a stranger, blocking the core workflow.

Mitigation

Offer an NDA-style one-line data promise and delete source files after 30 days; accept a public study guide or syllabus instead of private notes.

Validation experiment

Ask 20 instructors for files before the product is finished and count how many send anything within 48 hours.

Kill criteria

If fewer than 3 of 20 send files, the channel is wrong and the offer must change to a paid done-for-you build.

Risk 3Cold outreach reply rates fall far below the projected conversion.

Mitigation

Track reply rate per message angle and rewrite the opener every 40 sends; prioritise instructors with a visible cohort start date.

Validation experiment

Send 40 messages with angle A and 40 with angle B in week one and compare reply rates.

Kill criteria

If 80 sends produce fewer than 4 replies, abandon cold DM and test community posting plus a paid sponsorship instead.

Risk 4The free 10-question preview satisfies instructors so the $149 tier never converts.

Mitigation

Watermark the preview, cap it at 10 questions, and withhold the answer-key appendix and quiz links until payment.

Validation experiment

Measure sample-to-paid conversion on the first 10 delivered samples.

Kill criteria

If 10 samples yield zero payments, cut the free preview to 5 questions or charge $19 for the preview itself.

Risk 5Document parsing fails on scanned or photographed notes, breaking the 24-hour promise.

Mitigation

OCR fallback, an explicit parsing-failure email asking for a text export, and a manual paste path.

Validation experiment

Run 10 real instructor documents through the parser and record the clean-text success rate.

Kill criteria

If fewer than 7 of 10 parse cleanly, restrict intake to text-based PDFs and DOCX only.

Section 8 of 8

Hour-by-hour: the first 24 hours

Hours 1-2

  • Create the Stripe account and connect the payment account to TrustMRR before any charge.
  • Scaffold the Next.js repo on Vercel with Supabase storage.

Output: Live environment and a payment account ready to receive money.

Hours 3-6

  • Build the 100-instructor list in one licensing exam family.
  • Record name, exam, contact route, visible cohort size and channel.

Output: Spreadsheet of 100 reachable buyers.

Hours 7-12

  • Ship the upload route for notes and outline with parser wiring.
  • Test extraction on three public instructor PDFs.

Output: Working ingest that returns clean text.

Hours 13-18

  • Write and test the blueprint parser and style sampler prompts on one real outline.
  • Store the blueprint JSON per job.

Output: Verified blueprint tree for one exam family.

Hours 19-24

  • Send the first 20 outreach messages with the free sample offer.
  • Draft the booklet HTML template so day 5 rendering is not starting cold.

Output: 20 messages out and a template draft committed.

Complete report

Full 30-day launch report

Read the full generated report

Crash Course Press — 30-day launch report

1. Product opportunity and positioning

Licensing-exam instructors already own the content and already sell the course. What they do not have is a weekend. Every cohort needs 100-150 exam-style questions with four options, an answer key and a rationale — in the instructor's own voice, aligned to the current official content outline. Today that means writing them by hand, reusing a stale bank, or hiring an item writer with a two-week turnaround.

Crash Course Press collapses that into one upload. Notes in, blueprint-tagged booklet out: a branded, print-ready PDF plus timed cohort quiz links with score reports. The buyer is the solo or 2-3 person operator of a licensing-exam crash course in one exam family, paying by card from their own business, findable through their own course website, YouTube prep channel, instructor Facebook groups and the subreddits where they complain about question banks.

Positioning is artifact-first. This is not a learning platform and not a question marketplace. It is an item-writing and typesetting service that replaces a weekend task, and the output looks like the instructor's material because generation is anchored to their phrasing, stem length and difficulty.

JEV scores the concept 67.7/100 (band B, confidence 0.72). The profile sets three operating rules. Speed to first revenue (8.3) and reachable customers (7.86) are the strengths, so outreach starts on day 1 and the first sale is targeted for day 11. Monetization (5.79) is middling, so pricing stays one-time with no billing machinery and no free plan that gives the artifact away. Operational simplicity (5.3) is the weakest criterion, so the product ships exactly one output artifact and support stays one inbox with a stated 24-hour turnaround — no live chat, no proctoring, no LMS promises.

The wedge is deliberately narrow: one exam family, one buyer type, one deliverable. Breadth is the failure mode. An instructor evaluating a generic AI study tool asks whether it is trustworthy; an instructor evaluating this asks whether the booklet is usable on Monday, which is a same-day demo answer built from their own files.

2. Smallest sellable MVP

Seven things ship: dual upload for notes and outline, a blueprint parser that weights domains, a style sampler that reads eight existing questions, a batched item generator with rationales, a validity check, a branded PDF renderer, and a Stripe gate around a free watermarked 10-question preview. Timed quiz links land on day 6, before the first real sale on day 11.

Cut hard: no student accounts, no gradebook, no adaptive engine, no marketplace, no mobile app, no essay grading, no proctoring. Every one of those would add support load the JEV operational-simplicity score says this concept cannot carry in 30 days.

Demo flow: the instructor sends two files, gets ten watermarked questions back inside 24 hours, pays $149, and receives the full booklet PDF and quiz links the same day.

3. Technical implementation plan

Next.js on Vercel, Supabase for Postgres and file storage, a managed parsing or OCR API for extraction, a frontier LLM for blueprint mapping, style sampling and item generation, Playwright for HTML-to-PDF, Stripe Payment Links with webhooks, and Resend for score reports and delivery email.

Generation runs as four LLM calls: outline to weighted domain tree, notes to eight style anchors, batched item generation in groups of 15, and a repair pass for items the validity check rejects. Batching is the key design choice — a bad batch is regenerated instead of failing the whole job, which matters when a promised turnaround is 24 hours.

Payment state lives in Postgres. A Stripe webhook flips paid on the job, which unlocks the signed PDF URL and the quiz routes. Nothing is delivered before that flag.

The hardest risk is parsing. Scanned or photographed notes produce garbled text and every downstream step degrades silently. Mitigation: OCR fallback, a paste-text path, and an explicit failure email asking for a text export rather than a silent bad booklet.

4. Pricing and monetization

Single Cohort at $149, one-time: up to 150 blueprint-tagged questions, branded PDF, timed quiz links and score reports. Season Pass at $399, one-time: four booklets, reusable brand template, CSV question-bank export. The free tier is a watermarked 10-question preview with no answer-key appendix and no quiz links.

Expected first-purchase value is $149. The upsell path is credit-backed: a Single Cohort buyer running several cohorts gets the Season Pass with their first payment applied. Blended average revenue per customer rises only when that upsell lands, which is why the upside scenario models five Season Pass sales rather than assuming them.

5. First-customer acquisition

One channel: direct outreach to solo instructors in a single licensing exam family. Day 1 builds a 100-row list from course websites, YouTube prep channels, instructor Facebook groups, subreddits and state licensing school directories. Messages lead with the unbilled weekend and the free 10-question sample, not the product features.

Delivery is the pitch. Send files, get ten questions back inside 24 hours, then send the $149 checkout link in the same thread. A 72-hour follow-up asks one question: when does your next cohort start. Every buyer is asked for one introduction to another instructor on the same exam, because these niches are tight-knit.

Week-one target: 80 contacted, 8 samples delivered from real files, 2 paid booklets.

6. The 30 days

Week 1 is build-and-sell in the same days: days 1-5 ship the MVP, day 6 adds quiz links and three free samples, day 7 sends the first 40 messages. Week 2 is delivery and conversion: samples inside 24 hours, a 90-second demo, 80 total contacts, checkout links to every sample recipient, first paid sale on day 11, then a quality pass and referral asks. Week 3 adds the Season Pass, richer score reports, outreach wave 3, variant outline support, a queue for unattended turnaround, testimonials and a midpoint funnel review. Week 4 targets December cohort starts, adds a self-serve sample page, closes every open thread, clears the delivery backlog, reaches 150 contacts and runs the upsell wave. Days 29-30 are reconciliation and submission — every payment verified in the connected account, refunds matched, operating costs recorded.

7. Revenue and profit scenarios (projections only)

Conservative: 4 customers × $149 = $596 gross, 0 refunds, 0 chargebacks, $24 processing, $0 ads, $40 operating. Hackathon profit $572; business profit $532.

Base: 12 customers × $149 = $1,788 gross, $72 processing, $0 ads, $70 operating. Hackathon profit $1,716; business profit $1,646.

Upside: 20 customers at a $211 blended average (15 Single Cohort, 5 Season Pass) = $4,220 gross, $350 refunds, $169 processing, $250 ads, $140 operating. Hackathon profit $3,451; business profit $3,311.

Processing assumes a blended 4% on Stripe. All three are projections from stated assumptions — not verified revenue, customers or traction. The contest ranks hackathon profit, which excludes hosting and inference; those appear only in the business-profit view.

8. Main risks and validation

Item quality is risk one: if instructors must rewrite more than 30% of items, the product is a drafter, not a deliverable. Test by having two instructors review 40 items each.

File sharing is risk two: instructors may not hand notes to a stranger. Test by asking 20 for files before the product is finished; under 3 of 20 means the channel or the offer is wrong.

Cold outreach volume is risk three: test two message angles across 80 sends and abandon DM if replies fall under 4.

Preview cannibalisation is risk four: watermark it, cap it at 10 questions, withhold the answer key and quiz links, and measure sample-to-paid conversion on the first ten samples.

Parsing failure is risk five: run ten real documents through extraction; under 7 of 10 clean means intake must be restricted to text-based PDF and DOCX.

9. First 24 hours

Hours 1-2: create Stripe, connect the payment account to TrustMRR, scaffold the repo and storage. Hours 3-6: build the 100-instructor list with contact routes and cohort sizes. Hours 7-12: ship the upload and parser path, verified on three public instructor PDFs. Hours 13-18: write and test the blueprint and style prompts on one real outline. Hours 19-24: send 20 outreach messages and commit a booklet template draft so day-5 rendering is not starting cold.

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 →