The Money Machine

The record

1. 2017

   MoviePass

   Held

   Whether to price unlimited movies at $9.95/month to drive subscriber growth

   • Cut price to $9.95 unlimited — Slash the plan to $9.95/month to spike subscriber countchose this
   • Keep tiered pricing — Retain the $15-to-$50 tiered plan structure

   Stated reasonThe unlimited plan was a temporary marketing gimmick to attract subscribers and inflate HMNY's stock price

   What happened

   • 2019-09-14 service shutdown: the service shut down
   • 2020-01-28 bankruptcy: parent company filed for Chapter 7 bankruptcy
   • 2025-01 guilty pleas: Ted Farnsworth pleaded guilty to securities fraud and conspiracy
   Read the full analysis →
2. 2009

   Starbucks

   Held

   Whether to consolidate card programs into a single prepaid, no-fee Stars loyalty system that funds an interest-free customer float

   • Consolidate into Stars program — Merge prior card schemes into one prepaid Stars-based loyalty systemchose this
   • Keep separate card programs — Leave Starbucks Card Rewards and Gold Card as distinct schemes

   Stated reasonThe app and Stars made the prepaid pool bigger and stickier, industrializing the interest-free float

   What happened

   • 2018-12 active Rewards members: 16 million active members, 40% of tender in U.S. company-operated stores
   • 2023 stored value and deferred revenue balance: $1,567.5M at fiscal year-end 2023
   • 2023-12 stored value and deferred revenue balance: $2,199.8M at the December 2023 peak
   Read the full analysis →
3. 2022

   New York Times

   Held

   Whether to build a habit-forming bundle that dilutes per-subscriber revenue in exchange for stickier subscribers, rather than maximize ARPU on a news-only base

   • Build habit-forming bundle — Add games, cooking, sports to lock subscribers in even at lower ARPUchose this
   • Maximize news-only ARPU — Keep the news-only tier central and price for revenue per subscriber

   Stated reasonNews is the most cancelable product, so the bundle trades ARPU for daily-habit products that keep subscribers paying regardless of the news cycle

   Committed$550 million (all-cash purchase price for The Athletic, an unprofitable sports site; plus more than $30 million for Wirecutter and a low-seven-figures sum for Wordle)

   What happened

   • 2024 Bundle and multiproduct ARPU: fell 6.7% while those subscribers grew 44.2%
   • 2024 News-only tier: shrank 30.6% while news-only ARPU rose 19.1%
   • 2024 Total digital-only ARPU: grew 2.6% to $9.42
   Read the full analysis →
4. 1974

   Vanguard

   Held

   How to structure ownership: extract profit for outside owners or have the funds own the management company

   • Investor-owned manager — Outside owners take surplus as profit
   • Funds own the manager — Funds own the management company, surplus returns as fee cutschose this

   Stated reasonWith no outside owner, surplus cannot be extracted as profit and can only be recycled as lower fees

   What happened

   • 2025-02 expense ratio cut: cut fees on 168 share classes by an average of 20%, its largest fee cut
   • 2026-02 all-fund average expense ratio: second round across 84 classes pushed the all-fund average expense ratio down to 0.06%
   • 2025 global AUM: roughly $10.1 trillion, a more than tenfold climb from 2005, second-largest asset manager
   Read the full analysis →
5. 1984

   Dell

   Held in part

   Build machines to inventory, or take payment first then buy parts to fill each order

   • Build to order — Take payment first, then buy parts to fill the orderchose this
   • Mass-produce to stock — Build machines in advance and warehouse them awaiting buyers

   Stated reasonThe company had no capital to mass-produce, so it built only what customers had already paid for

   What happened

   • 1998 cash conversion cycle: turned negative, reaching negative 8 days in fiscal year 1998
   • 2000 cash conversion cycle: negative 18 days by FY2000
   • 1997 return on invested capital: hit 167%
   Read the full analysis →
6. 1987

   TSMC

   Held

   Should TSMC design its own chips or pledge to be a pure-play foundry that never competes with customers?

   • Pure-play foundry pledge — Manufacture only and promise never to compete with customerschose this
   • Integrated chipmaker — Own the fab and design chips like the industry dogma demanded

   Stated reasonA designer would not outsource its crown-jewel chip to a manufacturer that also designs chips; promising to be only a manufacturer removed the biggest reason to refuse to outsource.

   Committed$220 million in capital (TSMC's opening capital)

   What happened

   • 2025 global foundry market share: 69.9% of the global foundry market, up from 64.4% the year before, with Samsung at 7.2%
   • 2025 revenue: revenue reached about US$122.54 billion for the year, up 36.1%
   • 2025 customer concentration: top ten customers accounted for 78% of net revenue, with the single largest at 19%
   Read the full analysis →
7. 1999

   Salesforce

   Held

   How to sell a delivery method buyers had no category for: build a better feature or pick a fight

   • Market the architecture — Sell buyers on multi-tenancy and the technical merits directly
   • Wage a 'No Software' crusade — Name a villain and give buyers a movement to joinchose this

   Stated reasonMulti-tenant architecture is invisible and buyers had no category, so a villain and banner gave them a category to belong to

   What happened

   • 2004-06-23 IPO proceeds: offered 10,000,000 shares at $11.00 to raise $110 million
   • 2004-06-23 first-day stock gain: closed its first day at $17.20, a gain of more than 55%
   • 2009 annual revenue: reached $1 billion in revenue in 2009
   Read the full analysis →
8. 2017

   Epic Games

   Held

   Ship Battle Royale as a paid bundle add-on, or cut it loose as a separate free release monetized via currency?

   • Ship inside paid bundle — Sell Battle Royale as part of the paid Save the World game
   • Free release, sell currency — Release free and monetize via V-Bucks cosmeticschose this

   Stated reasonEpic worried that charging up front would strangle growth before it started

   What happened

   • 2023 Fortnite estimated revenue: estimated $3.5 billion in 2023, around 80% of Epic total
   • 2022-12 FTC settlements: settlements totaling $520 million with Epic
   • 2026-03 V-Bucks purchasing power: the $8.99 pack that once bought 1,000 V-Bucks now bought 800
   Read the full analysis →
9. 2019

   Charles Schwab

   Held

   Whether to cut trading commissions to zero to protect the deposit-spread business

   • Cut commissions to zero — Concede commission revenue to pull in more accounts and idle cashchose this
   • Keep charging commissions — Retain the commission revenue line and its per-trade tollbooth

   Stated reasonCommissions were only about 4% of revenue while the real business is the spread on swept client cash, so dropping them fills the funnel that feeds the bank

   What happened

   • 2022 Net interest revenue: reached $10.7 billion, up 33% in a single year
   • 2022 Balance sheet size: shrank $115 billion, a 17% contraction, on cash sorting
   • 2024 Net interest revenue: fell $537 million, down about 19% year over year
   Read the full analysis →
10. 2021

    Thermo Fisher

    Held

    Whether to expand beyond selling tools into owning outsourced clinical research (the digging itself)

    • Buy into contract research — Acquire PPD to own outsourced clinical trialschose this
    • Stay a pure tools seller — Remain focused on high-margin instruments and consumables

    Stated reasonDon't bet on the drug; own more of the picks and then own the digging itself

    Committed$17.4 billion in cash (PPD acquisition, plus assumption of approximately $3.5 billion of net debt)

    What happened

    • 2024 revenue: Revenue fell 5% to $42.86 billion in 2023, and sat flat at $42.88 billion in 2024
    • 2024 GAAP diluted EPS: GAAP diluted EPS still grew 7% to $16.53
    • 2024 segment revenue share: Laboratory Products & Biopharma Services 52% of revenue
    Read the full analysis →
11. 2022

    New York Times

    Held

    Whether to build a subscription bundle by acquiring The Athletic rather than defend advertising revenue

    • Acquire The Athletic for bundle — Buy the sports site to add a fourth product and suppress churnchose this
    • Defend advertising revenue — Stay an ad-supported media company chasing marketer budgets

    Stated reasonEach product is a separate reason to keep paying, making the bundle harder to cancel and letting the company raise price

    Committed$550 million in cash (acquisition of The Athletic)

    What happened

    • 2024 total revenue: Total revenue of $2.59 billion, up roughly 6.6% on the year
    • 2024 free cash flow: Free cash flow of $381.3 million
    • 2024 total subscribers: roughly 11.43 million total subscribers, about 10.82 million digital-only
    Read the full analysis →
12. 2022

    Ford

    Reversed

    Whether to build EV capacity ahead of demand by splitting off Model e as a standalone EV unit with a two-million-EV target

    • Pre-build EV capacity — Split off Model e and build scale ahead of demandchose this
    • Wait for demand — Pause spending and let demand catch up to plants

    Stated reasonIn 2022 EV adoption looked about to go vertical and a legacy automaker that sat out the transition risked becoming the next Kodak

    What happened

    • 2025 cumulative EBIT losses: more than $16 billion in cumulative EBIT losses
    • 2025 restructuring charge and net loss: $19.5 billion pre-tax restructuring charge including $10.7 billion Model e impairments; GAAP net loss of $8.2 billion
    • 2026 profitability target date: EV profitability 'likely won't happen until 2029'
    Read the full analysis →