A venture fund

What came in

A fund whose seven majority-owned companies, ten products between them, reported by hand: weekly PDFs and spreadsheets. When the partners had a question, the answer started with an investigation. There was no product tracking and no product analytics across the portfolio, and no one could say in one number how a company was doing.

What we did

• Built the data infrastructure from scratch on our own servers: n8n, Metabase and a database of our own. Wrote autonomous parsers that log into portfolio back offices, pass two-factor authentication and pull more than a hundred thousand rows transactionally; connected other companies through sheet automations, direct database access or replication. Everything lands in one clean table.
• Put board-level P&L reporting on that pipeline, then designed the Health Index: sixteen metrics across eight pillars, from traction and burn multiple to hand-scored leadership and strategic fit, fed by ten pipelines and rolled into one score from 0 to 100 with AI-written explanations.
• Replaced Metabase with a reporting platform of our own: every project's reporting aggregated, AI bullets, analysis, health index, metrics and weekly and monthly views in one place.
• Built a venture-validation flow: an idea goes in, a landing page with copy and creatives comes out, paid traffic runs, a waitlist fills. The first launch brought thirty leads at under two dollars per click and about a dollar and a half per lead.
• Rebranded the fund and built its site and design system, with a library of free tools for founders as lead generation: a pitch analyser, a deck builder, an email assistant. Built internal products on the same platform, including a LinkedIn content engine that turns a partner's topics into verified, on-voice posts.
• Product due diligence on inbound deals, and product, marketing and operations work inside portfolio companies that gets acted on.

What changed

The fund runs its portfolio on live numbers instead of weekly PDFs. The board sees P&L across the portfolio on one screen. Companies get product metrics they never computed for themselves, from churn on down. New ideas are tested with traffic before anyone writes code.

What was built

• Portfolio data pipeline: parsers, integrations, own database
• Board-level P&L reporting
• Health Index scoring model
• Custom reporting platform
• Venture-validation flow
• Fund rebrand, site, design system and founder tools