Skip to main content

Startups

How ideas become durable companies.

This is where I show the work: deciding whether an idea is worth pursuing, documenting what I build or help build, studying how companies grow and make money, and reviewing my own investment decisions.

Four questions, in order.

Research is forward-looking: should this idea be pursued at all? The stages that follow track what was built, whether demand became repeatable, and whether growth created a durable business.

  1. Research

    Should this idea be pursued?

    Test the buyer problem, route to the first customers, technical feasibility, capital needs, and reasons to stop before committing to a build.

    See ideas under consideration
  2. Building

    What is worth shipping?

    Products I founded, co-founded, or helped move forward, with my role, the constraints, and the outcome kept explicit.

    Explore ventures and teardowns
  3. Growing

    Can demand become repeatable?

    Distribution, acquisition, activation, retention, and experimentation once an initial route to customers begins to work.

    Read startup field notes
  4. Monetization

    Can this become a durable business?

    Pricing, revenue models, margins, acquisition costs, funding, and the point where growth either creates value or consumes it.

    Explore monetization in the deep dives

Look back without letting hindsight win.

Startup Deep Dives investigate named companies using public evidence. Personal Investments examine my own thesis, outcome, uncertainty, and updated decision rule.

Public-evidence research

Startup Deep Dives

Reconstruct what founders and investors could reasonably have known before the outcome became obvious.

Browse the investigations

First-person decision reviews

Personal Investments

Postmortems on small startup crowdfunding decisions: the thesis, percentage outcome, benchmark, luck, and rule that changed.

Review the investment method

The evidence standard

Mechanisms before mythology.

The archive separates what could reasonably have been known at the time from what only became clear later. It is interested in mechanisms, tradeoffs, and decision quality—not founder mythology or tidy stories written after the outcome.

Every investigation asks

  1. The wedge: What narrow customer, problem, or behavior gave the company somewhere credible to begin?
  2. The growth engine: Which channel, loop, incentive, or operating system moved it beyond the first believers?
  3. The economics: How did the company charge, what did growth cost, and could the model support itself?
  4. The constraint: Which competitive, regulatory, financial, or organizational force could have stopped it?

Evidence rules

Public filings, court records, shareholder letters, and contemporaneous documents come first. Reporting and named interviews add context. Retrospective founder accounts are useful but treated as interested testimony, not neutral fact.

Important claims need a primary source or corroboration. Estimates are labeled as estimates, disagreements remain visible, and inference stays separate from fact. The aim is not a prediction with the answer key hidden. It is a clearer record of the evidence that should have changed the decision.

Public-evidence research

Startup Deep Dives

Retrospective investigations of the wedge, growth engine, economics, and constraint—read by mechanism rather than company mythology.

Start here · Flagship investigation

How the PayPal Mafia Actually Happened

The verified story of the PayPal Mafia — the fraud war, the honeymoon coup, the eBay siege — and a VC-style scorecard on why it worked.

Marketplaces and network effects

  1. How Airbnb Actually Grew2007–2014 · 46 min read

    The verified story of Airbnb — the seven rejections, the cereal boxes, the Craigslist legend checked against the record — and an evidence scorecard on why it worked.

  2. How Alibaba Beat eBay1999–2014 · 35 min read

    The verified story of Alibaba — the SARS quarantine launch, free listings against eBay's take rate, the Alipay trust unlock — and an evidence scorecard on why it worked.

  3. How Amazon Actually Grew1994–2003 · 48 min read

    The verified story of Amazon — the parents' $245,573, the Get Big Fast bet, the convertible that beat the crash by four weeks — and an evidence scorecard on why it survived.

  4. How Uber Actually Grew2008–2017 · 33 min read

    The verified story of Uber — the pitch deck that undersized its market 100x, the city playbook, the subsidy wars, the 2017 collapse — and an evidence scorecard on what growth cost.

  5. How Mercado Libre Actually Compounded1999-2026 · 29 min read

    The verified story of Mercado Libre - the eBay alliance, payments-to-logistics compounding, Latin America timing - and a VC-style scorecard on why it worked.

Product-led subscriptions

  1. How Spotify Beat Piracy2006–2015 · 47 min read

    The verified story of Spotify — Swedish piracy culture, the label equity deals, the freemium funnel in the F-1 numbers — and an evidence scorecard on why it worked.

  2. How Netflix Actually Beat Blockbuster1997-2012 · 31 min read

    The verified story of Netflix - the DVD-by-mail wedge, the Blockbuster war, the streaming pivot - and a VC-style scorecard on why it worked.

  3. How Duolingo Actually Grew2011-2026 · 33 min read

    The verified story of Duolingo - the translation-model detour, the streak-driven freemium machine, the English Test inflection - and a VC-style scorecard on why it worked.

Developer-led infrastructure

  1. How Stripe Actually Grew2009-2026 · 33 min read

    The verified story of Stripe - the developer API wedge, the payment-infrastructure slog, the private-company compounding - and a VC-style scorecard on why it worked.

Capital-intensive systems

  1. How Tesla Almost Died2003–2013 · 37 min read

    The verified story of Tesla — the Christmas Eve 2008 rescue, the deposits that funded the company, the ZEV credits inside the first profit — and an evidence scorecard on why it survived.

Smaller companies · Narrower questions

Operator Briefs

Shorter investigations of bootstrapped and emerging businesses, with the limits of the public evidence kept visible.

Audience-led bootstrapping

  1. How Marc Lou Went From Broke in Bali to $1M+/Year Shipping SaaS2021–present · 13 min read

    The verified story of how ~30 failed products became ShipFast, CodeFast, DataFast, and TrustMRR — and the specific mechanic that made each one distribute the next.

    Evidence note: $1.03M revenue in 2025, self-disclosed

  2. How Rezi Beat the Big Resume Builders With Nothing But SEO2015–present · 15 min read

    The verified story of a decade-old, no-VC resume builder that out-shipped Zety and Resume.io on content and a two-year head start on GPT — from a Seoul-based team.

    Evidence note: ~$3.2M revenue in 2024, independently tracked

Developer-led and open source

  1. How Postiz Went From Open-Source Side Project to $1M ARR2024–present · 17 min read

    The verified story of an AGPL-licensed social media scheduler that grew revenue 5x in months by rebuilding itself for AI agents — with the GitHub-stars-to-paying-customers mechanic laid out in full.

    Evidence note: ~$1M ARR, Stripe-verified via TrustMRR

  2. How Simple Analytics Turned GDPR Into a Growth Channel2018–present · 18 min read

    The verified story of a 3-person, bootstrapped Google Analytics alternative that used European privacy rulings as free marketing — with a live public revenue dashboard to prove it.

    Evidence note: Live public dashboard — check current MRR at time of writing

Vertical software and commerce

  1. How Kibu Built a $2M ARR Business in an Overlooked Care Market2019-present · 11 min read

    The verified story of how Daniel Caridi turned disability-provider paperwork, programming, and compliance into a vertical SaaS growth engine.

    Evidence note: TrustMRR: ~$234K MRR, last synced Apr. 19, 2026

  2. How Supliful Turned Creator Commerce Into CPG Infrastructure2021-present · 11 min read

    The verified story of how the Grafomap team rebuilt print-on-demand logic for supplements, coffee, skincare, and pet products - with the gross-vs-net revenue caveat that makes the numbers honest.

    Evidence note: Stripe-verified: ~$963K trailing-30-day revenue, ~$50.7M all-time

Portrait of Atticus Li

Founder support

Have a real startup question?

I am not running an open funding program. I do make time for focused mentoring and scoped operating work when there is a specific problem to solve.

Free mentoring

Work through a founder question

Focused conversations about growth, experimentation, product evidence, or the next decision in front of you.Explore founder mentoring

Scoped collaboration

Bring me into the work

Paid consulting for teams that need senior growth and experimentation judgment tied to a defined business problem.Review ways to work together