The Fall

The record

1. 2001

   Borders

   Held

   Whether to run its own online bookselling channel or hand it to Amazon's platform

   • Outsource site to Amazon — Re-launch Borders.com powered by Amazon as seller of recordchose this
   • Keep running own site — Continue operating its own money-losing e-commerce channel

   Stated reasonThe company couldn't run the online channel profitably; costs exceeded the revenue it generated

   What happened

   • 2008-05 own e-commerce site: Borders did not launch its own proprietary e-commerce site again until May 2008
   • 2008 goodwill impairment: took a $40.3 million charge to write off the goodwill on the superstore segment
   • 2008 comparable store sales: Comparable store sales at the Borders superstores fell 12.8% in a single quarter
   Read the full analysis →
2. 2006

   Microsoft Zune

   Held

   Should Microsoft build the Zune on its existing PlaysForSure ecosystem or launch a fresh, incompatible platform?

   • Extend PlaysForSure ecosystem — Build the Zune on the existing licensed standard and partner network.
   • Launch incompatible new platform — Ship the Zune with fresh DRM, abandoning PlaysForSure and MSN Music.chose this

   Stated reasonThe Zune was positioned around sharing, discovery and community, favoring a fresh cleaner design over the existing standard.

   What happened

   • 2006-11 launch-week U.S. unit share: 9% share behind the iPod's 63%
   • 2009-01 holiday revenue year over year: fell roughly 54%, from about $185 million to about $85 million
   • 2011-03-15 hardware development halted: Microsoft said it would build no new Zune hardware
   Read the full analysis →
3. 2011

   Nokia

   Held

   Whether to abandon Symbian and bet the company on Windows Phone rather than persist with its own platforms

   • Bet on Windows Phone — Abandon Symbian and rent Microsoft's ecosystemchose this
   • Fix Symbian internally — Repair its own fractured platform and defend the present
   • Ship its own alternatives — Push its promising alternative platforms with conviction

   Stated reasonNokia couldn't build a third ecosystem alone, so it chose to rent Microsoft's

   What happened

   • 2013 smartphone market share: fell to 3.1% from 50.8%
   • 2012-07 market capitalization: collapsed to €6.28 billion
   • 2013-09 sale of Devices & Services business: Microsoft agreed to buy it for an announced €5.44 billion
   Read the full analysis →
4. 2011

   Google+

   Reversed

   How to seed a new social network: build a standalone place users choose, or route the existing ecosystem audience into a social layer

   • Integrate across ecosystem — Force Google+ identity into Search, Gmail, YouTube to seed accounts fastchose this
   • Build standalone social place — Grow a network people choose to enter for its own sake

   Stated reasonUse Google's owned ecosystem to seed Google+ instantly and leapfrog the cold-start problem

   What happened

   • 2012-01 average time per user: average Google+ user spending 3.3 minutes per month on the site in January 2012, versus more than 7.5 hours for Facebook
   • 2013-10 engaged users vs headline: roughly 540 million touched the social layer while only about 300 million interacted with the stream itself
   • 2018-10-08 session engagement: 90 percent of user sessions lasted fewer than five seconds
   Read the full analysis →
5. 2022

   PayPal

   Held

   Whether to keep or abandon the 750-million-account-by-2025 growth target

   • Keep account target — Continue optimizing toward 750 million active accounts by 2025
   • Abandon account target — Drop the headline account goal and flag a cautious outlookchose this

   Stated reasonPayPal had been counting dormant and low-value accounts to hit a vanity target, and pruning them barely touched the money moving through the system

   What happened

   • 2024 Net revenue: rose to $31.8B on $1.68T total payment volume
   • 2024 Active accounts: dipped ~8 million in 2023 then recovered to 434 million
   • 2025-02 Venmo revenue target: set a target of $2 billion in Venmo revenue by 2027
   Read the full analysis →
6. 2023

   Baidu

   Held in part

   Whether to pivot toward AI to replace a slowly shrinking search-ad engine

   • Pivot to AI businesses — Fund AI Cloud and robotaxi to replace the ad enginechose this
   • Defend the search core — Keep relying on the legacy search-ad monopoly

   Stated reasonThe ad engine is shrinking, so Baidu is racing to build AI businesses before the core fully cools

   What happened

   • 2024 full-year revenue: about $18.24 billion, down a mere 3.8% from 2023
   • 2025 full-year revenue: declined a further ~3% year over year
   • 2024 generative-AI revenue share: about $90 million — roughly 2% of the quarter's total
   Read the full analysis →
7. 1957

   Ford Edsel

   Reversed

   Whether to launch a new mid-price brand to plug the gap in Ford's lineup

   • Launch new mid-price brand — Create a distinct division with its own identity and dealers
   • Build on shared assembly lines — Save capital by using existing Ford and Mercury lineschose this

   Stated reasonSharing Mercury and Ford lines saved capital

   Committedpast $250 million (Ford's own stated investment in development, manufacturing, and marketing)

   What happened

   • 1959-11-19 brand discontinuation: Ford announced the end of all Edsel production, citing 'particularly disappointing' sales
   • 1960 total production: total production reached just 118,287 units across all three model years
   • 1958 first-year sales: 63,110 units, second-largest launch for a new American auto brand
   Read the full analysis →
8. 1995

   Nintendo Virtual Boy

   Held

   Ship the unfinished Virtual Boy to hold shelf space for the N64, or wait

   • Ship unfinished as stopgap — Push the Virtual Boy out early to occupy retail and free N64 resourceschose this
   • Hold until finished — Delay release until the product was ready in its intended form

   Stated reasonThe company needed something to occupy retail space and engineer attention until the N64 was ready

   What happened

   • 1996-03-31 worldwide units sold: sold approximately 770,000 units worldwide through March 31, 1996
   • 1996-08 discontinuation: discontinued in Japan on December 22, 1995 and in North America in August 1996
   • 1995 Japan units sold: roughly 140,000 units sold in Japan before it was pulled from shelves in December
   Read the full analysis →
9. 1997

   Xerox

   Reversed

   Facing a declining copier core, whether to disclose the decline or accelerate future revenue to mask it

   • Accelerate lease revenue — Book multi-year lease revenue too soon to show growthchose this
   • Disclose the decline — Show investors the copier business in retreat
   • Fund a second act — Reinvest to replace the fading copier franchise

   Stated reasonRather than show investors a copier business in retreat, leadership pulled future revenue forward to manufacture the appearance of growth

   What happened

   • 2000 restated FY2000 revenue: restated FY2000 revenue was $18.7 billion, fallen from about $19.6 billion the year before
   • 2002-04-11 SEC settlement penalty and restatement: Xerox paid a $10 million penalty and restated four years of financials in April 2002
   • 2024 revenue decline decades later: revenue was about $6.2 billion in 2024, down from $6.9 billion the year before
   Read the full analysis →
10. 2000

    AOL-Time Warner

    Held

    Whether AOL should spend its bubble-valued stock to acquire Time Warner's hard media assets

    • Buy hard assets with stock — Convert overvalued shares into Time Warner's durable media assetschose this
    • Hold the dial-up business — Keep operating dial-up subscriptions as a wasting franchise

    Stated reasonAOL knew its dial-up business was a wasting asset and its stock was overvalued, so it converted the paper into hard assets before the market noticed

    Committedroughly $165 billion in stock (Amount AOL paid for Time Warner)

    What happened

    • 2002 annual net loss: $98.7 billion in a single year, 2002 — the largest annual loss in U.S. corporate history
    • 2002 goodwill impairment charge: $54 billion noncash pretax charge in Q1 2002, substantially all created in the merger
    • 2004 SEC settlement and revenue restatement: Time Warner paid $300 million to the SEC and restated, cutting AOL's advertising revenue by approximately $500 million
    Read the full analysis →
11. 2005

    Blockbuster

    Reversed

    Whether to eliminate late fees to match Netflix, despite them being about 16% of revenue

    • Keep late fees — Preserve the profit center keeping the company alive
    • Eliminate late fees — Match Netflix by abolishing the fees and absorbing the losschose this

    Stated reasonNetflix's entire pitch was the abolition of the late fee, so matching it required offering the same thing

    What happened

    • 2010 late fee policy: the company reintroduced late fees
    • 2010-09-23 bankruptcy: Blockbuster filed Chapter 11 in the Southern District of New York
    • 2005 state settlements: settlements with attorneys general in 47 states over how the 'no late fees' promise was presented
    Read the full analysis →
12. 2007

    BlackBerry

    Held

    Whether to defend the keyboard platform or rebuild for a touchscreen app ecosystem

    • Defend keyboard platform — Keep building on the classic BlackBerry OS and keyboard betchose this
    • Rebuild the platform early — Tear up the foundation and rebuild on a modern platform in time

    Stated reasonThe company believed keyboards would win and touch typing was a real challenge for iPhone users

    What happened

    • 2011-09 subscriber peak: peaked at 85 million subscribers
    • 2012-06 shipments and net loss: shipped just 7.8 million phones, down 41% year over year, with a net loss of $192 million and 5,000 layoffs
    • 2013-11 strategic collapse: raised $1 billion in convertible debt, took a $250 million commitment from Fairfax, and handed the wheel to a new CEO
    Read the full analysis →