Score B · 6.73/10Free to read$100K Hackathon · November 16, 2026 → December 16, 2026Nov 16, 2026

Bulletin Nounou

Bulletin Nounou turns a French childminder's month of attended days into a ready-to-send bulletin de salaire PDF plus a copy-paste Pajemploi figure card, in under three minutes per child.

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.

Promising6.73/10
MVP build (solo)
5 days
Launch price
$29
Sales to $1,000 revenue
35 customers
First paying customer
Public French-language childminder Facebook groups, answering existing payslip questions with a 40-second demo and then linking the paid page.
Who pays

A self-employed assistante maternelle agréée in France, working alone, caring for one to four children from different families, admin-averse, reachable in French-language childminder Facebook groups and childminder hashtags on Instagram and TikTok.

Biggest execution risk

Operational simplicity 4.8/10

Remove regulated data, human review steps and custom infrastructure; keep the product self-serve and stateless on managed services.

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; 35 sales at $29 clears $1,000 of collected revenue; strongest signal is reachable customers at 7.9/10.

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
$29
Solo build
5 days
30d base revenue
$1,000
30d hackathon profit
$926
Section 1 of 8

Product opportunity and positioning

Target customer

A self-employed assistante maternelle agréée in France, working alone, caring for one to four children from different families, admin-averse, reachable in French-language childminder Facebook groups and childminder hashtags on Instagram and TikTok.

Why now

December is a payslip month, the year-end recap and attestation fiscale follow immediately, and the buyer is identifiable by job title in public groups with no marketplace cold-start.

Problem

Every month she redoes the same arithmetic per child — mensualisation, absence deductions, indemnités d'entretien only on attended days, CP accrual, heures complémentaires et majorées — then produces a bulletin de salaire per family and re-keys the numbers into Pajemploi.

Solution

A contract profile per child plus a tap-through month calendar; a deterministic rules engine computes the payslip, renders a bulletin de salaire PDF, and shows a Pajemploi card of exact figures with copy buttons.

Differentiation · Not a childcare matching platform and not a generic payroll suite: a single-player document pipeline for one self-employed childminder, paid by the caregiver rather than the parent.

1. Product opportunity and positioning

A self-employed assistante maternelle agréée runs a one-person business with a monthly deadline she cannot skip. For every child she holds a contract, and every month she recomputes mensualised hours, deducts absences, adds indemnités d'entretien only for days the child actually attended, accrues or pays CP, and handles heures complémentaires et majorées. She then produces a bulletin de salaire per family and re-keys the same numbers into Pajemploi. One slip and a parent queries it; she corrects and re-sends.

Bulletin Nounou narrows that to one job: the recurring monthly document. Not matching, not a parent-facing app, not a French payroll suite. A saved contract profile per child, a tap-through month calendar, a deterministic calculation engine, and three outputs — the bulletin PDF, a Pajemploi card with copy buttons, and a year-end recap with attestation fiscale. The payer is the childminder herself, so there is no two-sided cold start and no supply side to recruit.

The JEV score of 6.73 (band B) is honest about the trade-off. Reachable customers and speed to first revenue score above 7.8; monetisation potential and operational simplicity sit near 5.0–5.9. The plan below is shaped by that: sell fast to a group findable by job title, keep the money math free of any language model, and never take on regulated advisory work.

The buyer is identifiable by job title in public French-language childminder groups, has a deadline inside the contest window, and already pays for this task in time or money. Nothing in the first sale needs a brand, an ads budget, a partnership or procurement. The competitor set is free spreadsheets plus established paid childminder tools; the wedge is not more features, it is a 40-second screen recording of one month becoming a correct PDF and a Pajemploi card.

Positioning rule for all copy: the tool is an assistant that prepares her documents. She verifies the figures before sending. That framing is both accurate and the main defence against the Pajemploi-rejection objection.

Section 2 of 8

Smallest sellable MVP

A paid web app where a childminder creates one contract profile per child, taps the days attended this month, and downloads the bulletin de salaire PDF plus a copy-paste Pajemploi card.

Must-have features

  • Contract profile per child: mensualisation type, monthly hours, hourly rate, indemnité d'entretien per day, meal indemnity, start date, CP rate
  • Month calendar with tap-to-mark attended days, half-days and absence reasons, saved per child and per month
  • Deterministic TypeScript engine for mensualisation, absence deduction, indemnités on attended days only, heures complémentaires et majorées, CP accrual and payment
  • Bulletin de salaire PDF renderer reproducing the standard childminder payslip blocks, one click per child or batch
  • Pajemploi card listing heures normales, heures complémentaires, jours d'activité, salaire net and indemnités with copy buttons
  • History view so month n plus 1 is a repeat, plus a year-end recap and attestation fiscale PDF per family
  • Stripe Checkout with single-month and yearly passes gated by a license key

Explicitly out of scope

  • No caregiver directory, matching, messaging or parent-side accounts
  • No general French payroll engine, no URSSAF filing, no net-to-gross outside the childminder case
  • No native mobile app; responsive web only
  • No Pajemploi or tax-authority API integration; PDF and copy-paste only
  • No agency, crèche or multi-employee accounts
  • No free tier, forum or content library
  • No automated legal or tax advice of any kind

Demo flow (what the buyer sees)

  1. Buyer pays 29 dollars by card and receives a license key by email
  2. She creates one contract profile for one child in the wizard
  3. She opens the current month and taps the days attended, one half-day and one absence
  4. The engine shows the computed breakdown on screen for her to check
  5. She clicks generate and downloads the bulletin de salaire PDF
  6. She copies each Pajemploi figure from the card into the official portal
Build plan

Technical implementation plan

Stack

Next.js on Vercel serverlessSupabase Postgres and auth for per-childminder accounts and contract profilesDeterministic TypeScript calculation module with rule-level unit tests@react-pdf/renderer for bulletin de salaire and attestation fiscale PDFsStripe Checkout plus license-key gatingResend for receipts and remindersOptional assistive LLM call to draft a contract profile from free text, always editable

Architecture

A single Next.js app on Vercel serves the French marketing page and the authenticated app. Supabase stores accounts, children, contracts, months and day marks, with row-level security scoped to the account. All money math lives in one pure TypeScript module with no network calls and no model in the path; the UI renders its output and the PDF renderer consumes the same computed object, so screen and PDF can never diverge. Stripe Checkout creates the payment and a license key unlocks the app; Resend sends the receipt, the key and reminders.

Build steps

  1. Create the repo, deploy an empty app to Vercel, create the Supabase project and enable email auth
  2. Define the schema for accounts, children, contracts, months and days, with row-level security
  3. Write the contract profile field list from public official childminder documentation before writing UI
  4. Build the profile wizard and persist contracts
  5. Build the month grid with attended, half-day and absence-reason states
  6. Write the calculation module as pure functions, then unit tests per rule
  7. Render the computed breakdown on screen so the user verifies before download
  8. Add the bulletin de salaire PDF renderer from the same computed object
  9. Add the Pajemploi card with per-field copy buttons
  10. Add Stripe Checkout, license-key gating and the Resend receipt
  11. Add history, past-month backfill and the year-end recap plus attestation fiscale
  12. Run three synthetic contracts end to end and fix rounding and label mismatches
  13. Ship the French landing page with the demo video and checkout

AI components

  • Optional assistive LLM call that drafts a contract profile from a free-text note the childminder writes, always shown editable before saving; never used in the money math
Hardest technical risk: Rule correctness on absence deduction combined with CP accrual and heures majorées, where a wrong figure destroys trust immediately and is the fastest route to a refund.
Section 3 of 8

Pricing and monetization

Mois unique

Entry offer
$29 one-time

For A childminder who needs this month's bulletins for one to four children and wants to test the tool on a live deadline

  • Unlimited bulletins for one calendar month
  • Pajemploi copy-paste card per child
  • Saved contract profiles and month history

Reprise historique

$49 one-time

For A childminder who is already behind and needs past months entered so the year-end recap is complete

  • Backfill up to twelve past months through the same calendar
  • Regenerated past bulletins
  • Year-end recap across all families

Annee complete

$149 one-time

For A childminder who wants the whole year handled, including the attestation fiscale she must give each parent-employer

  • Twelve monthly runs for up to four children
  • Year-end recap and attestation fiscale PDF per family
  • Email support for rule questions

Model: One-time passes per scope of work, no subscription, no free tier.

First revenue offer: Mois unique at 29 dollars: this month's bulletins for all your children, the Pajemploi card, and your contract profiles saved — cheaper than one corrected payslip conversation with a parent.

Expected ACV: $79

Upsell path: Single-month buyers get the 149 dollar year offer by email with a preview of their own recap; anyone who mentions being behind is moved to the 49 dollar reprise historique first.

Target customers
45
Revenue model
$3,555
Costs
$130
Profit model
$3,425
Section 4 of 8

First-customer acquisition

Primary channel

Public French-language childminder Facebook groups, answering existing payslip questions with a 40-second demo and then linking the paid page.

Where to find them

  • Public French-language Facebook groups for assistantes maternelles agréées
  • Facebook groups about Pajemploi and mensualisation questions
  • Instagram childminder hashtags in French
  • TikTok childminder hashtags in French
  • Threads in those groups posted in the last 60 days mentioning bulletin de salaire, Pajemploi or CP

Exact outreach motions

  • Search each group for bulletin de salaire, Pajemploi, mensualisation, CP and heures complémentaires and log 40 threads
  • Reply in-thread with the specific rule their case hits, no link in the first reply
  • Post the 40-second demo as a standalone post in three groups with the three-minute-per-child claim
  • DM every commenter referencing their own child count and their own question
  • Day 3 and day 10 reminder tied to the December payslip deadline
  • At month end, email single-month buyers the year-end recap offer as the upgrade path

Message template

Bonjour, j'ai vu votre question sur la déduction d'absence de ce mois. Le calcul se fait sur les heures mensualisées moins les heures d'absence, et les indemnités d'entretien ne comptent que les jours réellement gardés. J'ai fait une démo de 40 secondes qui montre le mois coché et le bulletin qui sort directement en PDF, avec la fiche Pajemploi à copier. Si vous voulez tester ce mois-ci c'est 29 dollars, et vous gardez les profils de contrat pour le mois prochain. Je peux vous montrer en visio 10 minutes si vous préférez.
Week-one goal: 6 groups joined, 40 threads logged, 12 rule-specific replies posted, 25 DMs sent, 3 demo calls booked, 1 paid month pass.
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 1Foundations

Deployed empty app

  • Create Next.js repo, deploy to Vercel, enable Supabase auth
  • Create tables for accounts, children, contracts, months, days
  • Connect Stripe test keys and the TrustMRR-linked payment account
  • Write the contract profile field list from official rule documentation

Deliverable: Deployed skeleton with login and empty dashboard

Day 2Contract profiles

Profile wizard

  • Build the wizard: mensualisation type, monthly hours, hourly rate, indemnité d'entretien, meal indemnity, start date, CP rate
  • Validate inputs and persist per child with row-level security
  • Allow a second and third child in the same account

Deliverable: Three contract profiles savable and editable

Day 3Calendar

Month of attendance in taps

  • Build the month grid with attended, half-day and absence-reason states
  • Persist days per child per month
  • Pre-fill the current month from the last saved month

Deliverable: A full month recorded for three children in under five minutes

Day 4Engine

Correct money math

  • Encode mensualisation, absence deduction, indemnités on attended days only, heures complémentaires et majorées, CP accrual and payment
  • Write unit tests per rule using published worked examples
  • Display the computed breakdown on screen before rendering

Deliverable: Calculation module passing rule-level tests

Day 5MVP deadlineMVP shipped

Sellable product

  • Render the bulletin de salaire PDF from the computed object
  • Build the Pajemploi card with per-field copy buttons
  • Gate the app behind a Stripe license key

Deliverable: Sellable MVP: pay, tap, download PDF, copy Pajemploi figures

Day 6Hardening

Remove rounding doubt

  • Run the full month flow on three synthetic contracts including one absence and one overtime day
  • Fix rounding to two decimals on every line item
  • Add a verify-your-figures line to the PDF footer

Deliverable: Closed bug list and visually checked PDF

Day 7Landing

French sales page

  • Ship the French landing page with three prices and Stripe checkout
  • Send license key and access instructions via Resend
  • Record a 40-second demo and a 90-second walkthrough

Deliverable: Live French sales page plus two demo videos

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

Thread list

  • Join six public French childminder Facebook groups
  • Search 60 days of posts for payslip and Pajemploi questions
  • Log 40 threads with the specific rule each one hits

Deliverable: Spreadsheet of 40 answerable threads

Day 9First posts

Answer before selling

  • Post the 40-second demo in three groups with the three-minute claim
  • Answer ten existing threads with rule-specific replies
  • Track reach and comments in the sheet

Deliverable: Three posts live and ten answers posted

Day 10Direct messages

Conversations

  • Send 25 DMs referencing each person's own question
  • Offer the 29 dollar month pass with the corrected-payslip framing
  • Invite three people to a ten-minute screen share

Deliverable: 25 DMs sent and three calls booked

Day 11First saleFirst sale target

Close the loop

  • Reply to every comment within the hour
  • Publish the absence-deduction walkthrough video
  • Remove checkout friction found so far

Deliverable: First paying customer plus a written objection log

Day 12Onboarding

Watch without helping

  • Observe two buyers run their own month in silence
  • Fix every hesitation point
  • Add inline rule explanations where they asked

Deliverable: Two bulletins produced by real users unaided

Day 13Second wave

Overtime and CP content

  • Post a demo showing heures complémentaires et majorées
  • Answer ten more threads
  • Post in two new groups

Deliverable: Second content wave live in four groups

Day 14Review

One change only

  • Count reach, clicks, checkouts and sales
  • Cut the worst-converting sentence on the landing page
  • Email the week's buyers a two-question feedback ask

Deliverable: Week-2 plan with one channel change or one message change

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

Backfill flow

  • Ship past-month entry so a behind childminder can backfill
  • Expose it as the 49 dollar reprise historique
  • Regenerate past bulletins from the same engine

Deliverable: Backfill flow live in checkout

Day 16Backfill push

Sell the catch-up

  • DM the threads where people said they were behind
  • Offer the 49 dollar tier with a same-day profile setup

Deliverable: Backfill offer sent to 20 people

Day 17Year end

Attestation fiscale

  • Build the year-end recap per family across all months
  • Render the attestation fiscale PDF
  • Attach it to the 149 dollar tier only

Deliverable: Année complète deliverable complete

Day 18Email

Automated receipts

  • Send receipts and renewal reminders through Resend
  • Send a December deadline reminder to every buyer

Deliverable: Automated email sequence live

Day 19Content

CP in June explainer

  • Record a short video on CP and why it surprises parents
  • Publish it on Instagram and TikTok with childminder hashtags

Deliverable: Two short videos published

Day 20Second channel

Hashtag conversations

  • Find payroll questions under French childminder hashtags
  • Answer with the rule, then DM
  • Hold paid ads at zero

Deliverable: 20 new conversations started from social hashtags

Day 21Mid-point review

Double down or cut

  • Compare conversion and time cost per sale by channel
  • Fix the top three support questions on the page
  • Handle any refund personally within 24 hours

Deliverable: Mid-point numbers written and one channel doubled down

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

Reduce load

  • Ship a short French FAQ covering absences, CP and Pajemploi limits
  • Route contact to one inbox

Deliverable: Support reduced to email only

Day 23Upgrade

Year tier

  • Email single-month buyers the 149 dollar year offer with a preview of their own recap
  • Add a one-click upgrade link

Deliverable: Upgrade email sent to all existing buyers

Day 24Deadline push

December payslips

  • Post the deadline reminder in every group joined
  • Offer same-day contract profile setup for anyone buying today

Deliverable: Deadline-week post live in six groups

Day 25Objections

Pajemploi rejection fear

  • Publish a page on how to verify figures and why the PDF is an assistant, not advice
  • Add the disclaimer to the top of every PDF

Deliverable: Objection page live and linked from checkout

Day 26Backfill wave two

Non-buyers

  • Re-DM everyone who engaged but did not buy
  • Lead with the 49 dollar catch-up option

Deliverable: Second DM wave sent

Day 27Retention

December close-out

  • Send December close-out instructions
  • Ask each buyer for one sentence of feedback

Deliverable: Retention email sent and feedback collected

Days 29-30 · Consolidate · 3 days
Day 28Edge cases

Two mensualisation types

  • Test two contracts with different mensualisation types on one account
  • Fix rounding and label mismatches found

Deliverable: Edge-case test log with fixes shipped

Day 29Final push

Last demo

  • Post the final demo of the month
  • Clarify price anchoring on the landing page
  • Answer every open thread

Deliverable: Final sales push live

Day 30Close

Reconcile and submit

  • Reconcile collected revenue, refunds and fees in Stripe
  • Submit profit to the hackathon dashboard
  • Write the January plan for attestation season and returning buyers

Deliverable: 30-day numbers submitted and a January plan

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

  • Projection only. Conservative: 7 buyers, all on the 29 dollar month pass, minimal word of mouth inside the window.
  • Projection only. Base: 20 buyers at a 50 dollar blended price (roughly 12 at 29 dollars, 6 at 49 dollars, 2 at 149 dollars).
  • Projection only. Upside: 45 buyers at a 79 dollar blended price (roughly 35 at 29 dollars, 8 at 49 dollars, 2 at 149 dollars), matching the concept monetisation math.
  • Payment processing modelled at roughly 4.5 percent blended on collected revenue, assuming Stripe card pricing with a per-transaction fixed fee applied to many small charges.
  • Refunds modelled at zero in conservative, one month pass in base, three month passes in upside.
  • Chargebacks modelled at zero except upside, where two 29 dollar disputes are assumed.
  • Paid marketing is zero in conservative and base; upside assumes 150 dollars of small sponsored posts, organic group outreach costs nothing.
  • Operating costs cover Vercel, Supabase, Resend, domain and PDF rendering, non-zero because every bulletin has a rendering and storage cost.
  • Several hundred French childminders see the demo posts within 30 days; this is measurable from post reach and falsifiable.
  • Two to four percent of people reaching checkout pay 29 dollars.
  • Refunds and chargebacks stay under five percent of collected revenue.
  • December is a payslip month, but attestation fiscale demand peaks after the calendar year closes, so the 149 dollar tier is expected to undersell inside the window.

Conservative

7 customers
$231
collected revenue · avg payment $33
Refunds
$0
Chargebacks
$0
Payment fees
$10
Paid marketing
$0
Hackathon profit
$221
Operating costs
$25
Business profit
$196

Base case

20 customers
$1,000
collected revenue · avg payment $50
Refunds
$29
Chargebacks
$0
Payment fees
$45
Paid marketing
$0
Hackathon profit
$926
Operating costs
$45
Business profit
$881

Optimistic

45 customers
$3,555
collected revenue · avg payment $79
Refunds
$90
Chargebacks
$20
Payment fees
$160
Paid marketing
$150
Hackathon profit
$3,135
Operating costs
$130
Business profit
$3,005

Model your own numbers

Runs in your browser · nothing is saved
Collected revenue
$580
Payment fees (3.9%)
$23
Paid marketing
$0
Hackathon profit
$557
Business profit
$512

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 scenario projections built from stated assumptions, not verified revenue, existing customers or observed behaviour.

Section 7 of 8

Main risks and validation experiments

Risk 1A wrong bulletin destroys trust and triggers refunds; absence deduction combined with CP is genuinely subtle.

Mitigation

Pure TypeScript module, unit tests per rule, breakdown shown on screen before download, verify-your-figures disclaimer on every PDF, personal refund within 24 hours.

Validation experiment

Encode five published worked examples and compare engine output to hand calculation before writing any UI.

Kill criteria

If two or more core rules cannot be encoded with confidence within two working days, stop and do not launch.

Risk 2The founder must write, demo and support in French, and a flat tone kills credibility in these groups.

Mitigation

Every public reply written in French and checked for rule accuracy before posting; no translated marketing copy.

Validation experiment

Post one rule-specific answer in one group on day 9 and count replies and follow-up questions.

Kill criteria

If ten rule-specific answers produce no replies or questions, the language or the framing is wrong and the channel must change.

Risk 3Free spreadsheets and established paid childminder tools already serve this audience, so 29 dollars may be too high.

Mitigation

Frame against the cost of one corrected payslip conversation, offer the month pass as no-risk entry, keep the yearly tier as the later upsell.

Validation experiment

Send 20 DMs with the exact 29 dollar offer and count checkout page visits and payments.

Kill criteria

If 20 DMs yield zero checkout visits, change the offer or the price before spending more outreach time.

Risk 4Group admins treat launch posts as spam and remove them.

Mitigation

Answer first, link second, never post the same text in more than three groups per day, prefer replies over standalone posts.

Validation experiment

Post the demo in three groups on day 9 and track removals within 24 hours.

Kill criteria

If two of three posts are removed, drop standalone posts and rely on thread answers and DMs only.

Risk 5Childminders fear the figures will be rejected by Pajemploi, which blocks purchase even when the math is right.

Mitigation

Position as an assistant whose figures she verifies; publish an objection page; keep the disclaimer visible in the product and the PDF.

Validation experiment

Publish the objection page on day 25 and compare checkout starts before and after.

Kill criteria

If buyers repeatedly ask for a guarantee of Pajemploi acceptance, narrow the promise to document preparation only.

Risk 6Seasonality: the 149 dollar year tier depends on attestation fiscale demand that peaks after the calendar year closes.

Mitigation

Focus the window on the 29 and 49 dollar tiers and treat the year tier as an email upgrade after the contest.

Validation experiment

Email the year offer to all buyers on day 23 and count upgrades.

Kill criteria

If zero upgrades from a full buyer list, remove the year tier from the window's messaging entirely.

Section 8 of 8

Hour-by-hour: the first 24 hours

Hours 1-2

  • Create the Next.js repo and deploy an empty app to Vercel
  • Create the Supabase project with email auth
  • Connect Stripe test keys and confirm the payment account will route to TrustMRR

Output: Deployed skeleton, database project, Stripe test mode confirmed

Hours 3-4

  • Read public official childminder documentation for mensualisation, absence deduction, indemnité d'entretien and CP
  • Write the exact contract profile field list and the calculation rule list

Output: One-page rule spec and field spec the whole build follows

Hours 5-6

  • Create the database schema for accounts, children, contracts, months, days
  • Build the first screen of the contract profile wizard and persist one contract

Output: One contract profile saved end to end

Hours 7-8

  • Join three public French childminder Facebook groups
  • Search for payslip threads and log ten of them
  • Draft the French reply template and the DM template

Output: Ten logged threads plus a tested reply and DM template

Complete report

Full 30-day launch report

Read the full generated report

Bulletin Nounou — public 30-day report

1. Product opportunity and positioning

A self-employed assistante maternelle agréée runs a one-person business with a monthly deadline she cannot skip. For every child she holds a contract, and every month she recomputes mensualised hours, deducts absences, adds indemnités d'entretien only for days the child actually attended, accrues or pays CP, and handles heures complémentaires et majorées. She then produces a bulletin de salaire per family and re-keys the same numbers into Pajemploi. One slip and a parent queries it; she corrects and re-sends.

Bulletin Nounou narrows that to one job: the recurring monthly document. Not matching, not a parent-facing app, not a French payroll suite. A saved contract profile per child, a tap-through month calendar, a deterministic calculation engine, and three outputs: the bulletin PDF, a Pajemploi card with copy buttons, and a year-end recap with attestation fiscale. The payer is the childminder herself, so there is no two-sided cold start and no supply side to recruit.

The JEV score of 6.73 (band B) is honest about the trade-off. Reachable customers and speed to first revenue score above 7.8; monetisation potential and operational simplicity sit near 5.0 to 5.9. The plan is shaped by that: sell fast to a group findable by job title, keep the money math free of any model, and never take on regulated advisory work.

The buyer is identifiable by job title in public French-language childminder groups, has a deadline inside the contest window, and already pays for this task in time or money. Nothing in the first sale needs a brand, an ads budget, a partnership or procurement. The competitor set is free spreadsheets plus established paid childminder tools; the wedge is not more features, it is a 40-second recording of one month becoming a correct PDF and a Pajemploi card.

Positioning rule for all copy: the tool prepares her documents, she verifies the figures before sending. That is both accurate and the main defence against the Pajemploi-rejection objection.

2. Smallest sellable MVP

The shippable unit is one child, one month, one PDF. A childminder pays 29 dollars, creates a contract profile, taps the days the child attended including one half-day and one absence, checks the on-screen breakdown, and downloads a bulletin de salaire plus a Pajemploi card she can copy field by field. Saved profiles and month history make the second month a three-minute repeat, which is the whole retention argument.

Everything else waits. No parent accounts, no messaging, no matching, no native app, no Pajemploi API, no generic payroll, no free tier. The engine is deliberately not a language model: the money math is pure TypeScript with unit tests per rule, and an LLM is used at most to draft a contract profile from a free-text note, always shown editable before saving.

3. Technical implementation plan

Next.js on Vercel serves a French landing page and the authenticated app. Supabase holds accounts, children, contracts, months and day marks under row-level security. Every figure comes from one pure calculation module, and both the screen breakdown and the PDF renderer consume the same computed object, so screen and document cannot diverge. Stripe Checkout issues the payment and unlocks a license key; Resend sends receipts, keys and reminders.

Build order: schema, contract wizard, month grid, calculation module with tests, PDF, Pajemploi card, checkout gating, then history and year-end. The hardest risk is rule correctness on absence deduction combined with CP accrual and heures majorées. The mitigation is sequencing: encode rules and prove them against worked examples on day 4, before any interface polish.

4. Pricing and monetisation

Three one-time passes, no subscription and no free tier. Mois unique at 29 dollars for this month across all children. Reprise historique at 49 dollars for someone already behind who needs past months backfilled so the year-end recap is complete. Année complète at 149 dollars for twelve monthly runs plus the attestation fiscale per family. The 29 dollar tier is the entry point because it is cheaper than one corrected payslip conversation; the 149 dollar tier is an email upgrade after the deadline, not the pitch inside it.

5. First-customer acquisition

The only channel is public French-language childminder Facebook groups. Search 60 days of posts for payslip, Pajemploi, mensualisation and CP questions, answer each with the specific rule the case hits, then post the 40-second demo. Reply publicly, DM privately, follow up on day 3 and day 10 around the December deadline. Instagram and TikTok childminder hashtags are a secondary top-up from day 19, not a parallel strategy. Week-one goal: six groups, 40 logged threads, 12 rule-specific replies, 25 DMs, three calls booked, one paid month pass.

6. Thirty-day execution roadmap

Days 1 to 5 build and ship the MVP: foundations, contract wizard, month calendar, tested engine, PDF plus Pajemploi card plus Stripe gating. Day 5 is the deadline for something a stranger can pay for.

Days 6 to 7 harden and sell: three synthetic contracts end to end, rounding fixes, French landing page, two demo videos. Days 8 to 14 are pure distribution: join groups, answer threads, post demos, DM, observe two buyers in silence, then review with one change only. First sale targeted on day 11.

Days 15 to 21 add the money tiers: backfill and the 49 dollar reprise historique, then the year-end recap and attestation fiscale for the 149 dollar tier, automated receipts, short video content, a second social channel, and a mid-point review that cuts the weaker channel.

Days 22 to 30 are conversion and close: FAQ, upgrade email, December deadline push, objection page on Pajemploi rejection fear, second DM wave, retention email, edge-case testing across two mensualisation types, final demo, and reconciliation of revenue, refunds and fees before submitting.

7. Revenue and profit scenarios

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

Conservative: 7 buyers, blended 33 dollars, 231 dollars gross, 10 dollars processing fees, no ads, 25 dollars operating cost. Hackathon profit 221 dollars, business profit 196 dollars.

Base: 20 buyers, blended 50 dollars, 1,000 dollars gross, one refund of 29 dollars, 45 dollars processing fees, no ads, 45 dollars operating cost. Hackathon profit 926 dollars, business profit 881 dollars.

Upside: 45 buyers, blended 79 dollars, 3,555 dollars gross, 90 dollars refunds, 20 dollars chargebacks, 160 dollars processing fees at roughly 4.5 percent, 150 dollars of small sponsored posts, 130 dollars operating cost. Hackathon profit 3,135 dollars, business profit 3,005 dollars.

Processing is modelled at about 4.5 percent blended, assuming Stripe card pricing where the fixed per-charge fee bites on 29 dollar tickets. Operating cost is non-zero in every case because each bulletin is rendered and stored.

8. Main risks and validation experiments

Rule correctness is the existential risk. Test it before the UI: five published worked examples, engine output compared to hand calculation. If two core rules cannot be encoded confidently in two days, do not launch.

Language and market distance is second. One rule-specific answer in one group, then count replies. Ten answers with no reaction means the tone or the framing is wrong.

Willingness to pay 29 dollars is third. Twenty DMs with the exact offer; zero checkout visits means change price or offer before spending more outreach hours.

Group admins removing posts is fourth. Answer first, link second, cap at three groups per day. Two removals out of three posts means drop standalone posts entirely.

Pajemploi rejection fear is fifth and is a positioning problem, not a product one. Ship an objection page and keep the verify-your-figures line at the top of every PDF. If buyers keep demanding a guarantee of acceptance, narrow the promise to document preparation.

9. First 24-hour action plan

Hours 1 to 2: repo, Vercel deploy, Supabase auth, Stripe test keys, confirm the payment account routes to TrustMRR. Hours 3 to 4: read official childminder documentation and write the rule spec and contract field spec that the whole build follows. Hours 5 to 6: database schema and one contract profile saved end to end. Hours 7 to 8: join three groups, log ten payslip threads, draft and test the French reply and DM templates.

By the end of day one there is a deployed skeleton, a written rule spec, one saved contract, and ten real threads to answer. That is the minimum needed for day 5 to be a sellable product rather than a demo.

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 →