CASE 12
6 weeks
From figma to public in six weeks. The seed round closed three months later.
A pre-seed founder needed a live two-sided marketplace fast enough to raise on traction rather than slides. We cut scope to bookings, payments and reviews, ran everything else as a human process behind the curtain, and instrumented the funnel from day one so the pitch was made of numbers.
THE PROBLEM
Pre-seed founder needed a two-sided marketplace in the market fast enough to raise a round on traction, not on slides.
The runway was months, not quarters, and investors in the category had made it clear that decks without usage were not getting second meetings.
The initial spec was a full platform: messaging, packages, subscriptions, loyalty, a provider mobile app. Building it would have consumed the runway before launch.
The genuine unknown was whether providers would maintain availability at all. Everything else was detail, and detail was what the spec was mostly made of.
CONSTRAINTS WE WORKED UNDER
WHAT WE DID
Stripped scope aggressively — bookings, payments, reviews; everything else was a manual human workflow at launch.
Stripe Connect for provider payouts; one integration, no bespoke ledger.
Provider onboarding that shipped real-time availability calendars (iCal + Google).
Instrumented every funnel step from day one so the pitch deck had real numbers, not projections.
ARCHITECTURE
01
Availability, holds and confirmed bookings with transactional integrity — a hold expires or converts, never both. Double-booking is prevented at the database level rather than in application logic, because the founder could not debug a race at midnight.
02
Connected accounts for providers, platform fee taken at capture, payouts and KYC handled by Stripe. No bespoke ledger, no money held in ambiguous states, no custom compliance work in a six-week build.
03
Two-way sync with Google Calendar and one-way iCal import so providers keep their existing calendar. This was the single feature that decided whether availability stayed accurate, so it got disproportionate build time.
04
Provider vetting, dispute handling and refunds ran through an admin panel and the founder's inbox. Every one of these looked like a feature in the spec; at launch volumes a human was faster, cheaper and more informative.
05
Every step — search, view, hold, book, pay, review — emitted events from day one into a simple warehouse table with a dashboard on top. The fundraise ran on that dashboard.
KEY DECISIONS
WHY
None of them tested the core hypothesis. Every one would have consumed a week that launch could not spare.
TRADE-OFF
The launched product looked thinner than the vision. Six weeks of real usage data proved more persuasive than a fuller product would have.
WHY
At launch volume, manual vetting takes minutes a day and teaches the founder what the automated rules should eventually be.
TRADE-OFF
It does not scale, and we said so explicitly with the threshold — roughly where the volume made it painful — written into the handover.
WHY
The funnel numbers were the fundraising asset. Retrofitting analytics after launch loses the earliest and most interesting cohort.
TRADE-OFF
About two days of build time that produced no user-facing feature, and comfortably the highest-return two days in the project.
THE HARD PARTS
Early bookings hit slots that were no longer free because providers maintained a separate calendar. Two-way Google sync solved it properly; the alternative — nagging providers to update the platform — would have quietly killed the marketplace.
The Figma contained the full vision, and every screen made a reasonable case for itself. We kept one written list of what launch included and what it explicitly did not, reviewed weekly. Most of the value we delivered was holding that line.
Cancellations and partial refunds after payout are genuinely fiddly. We handled the common paths in code and routed the rest to the founder with clear instructions, rather than pretending six weeks was enough to model every case.
HOW IT RAN
01
Turned the full spec into a launch list and a not-now list, agreed in writing with the founder before any code.
02
Booking core with transactional holds, Stripe Connect onboarding and payouts, and the admin panel for manual back-of-house.
03
Provider profiles, availability, Google two-way sync and iCal import, plus reviews.
04
Funnel events, dashboard, production hardening and public launch.
OUTCOME
6 weeks
Figma to public launch
3 months
From launch to an oversubscribed seed round
0
Bespoke payment ledger code written
Day 1
Funnel instrumentation live
The fundraise ran on this data.
Public launch in six weeks. Three months after launch, the seed round was oversubscribed.
The funnel data did the persuading. Walking investors through real conversion by cohort is a different conversation from projecting one.
The not-now list became the post-raise roadmap, already prioritised by what actual usage showed mattered — messaging was first, and loyalty never came back.
STACK
APPLICATION
DATA
PAYMENTS
INTEGRATIONS
HANDOVER
