Skip to main content

Research

Where the cases come from, and where the history stops.

Three hundred decisions were built from published accounts of real companies — two hundred that closed, one hundred that changed direction and survived. This page lists every source that work drew on, and states plainly which parts of the game are research and which parts are ours.

Documented history vs. game hypothesis

Two different kinds of statement, never mixed.

Historical summaries are drawn from the cited sources. Alternative choices, effects, and delayed outcomes are original game-design hypotheses for play, not claims about what would certainly have happened.


Research, with a citation

  • The company and roughly when.
  • The decision it faced, as its source describes it.
  • The verified result, as its source describes it.
  • The link to that source, so you can read it yourself.

Game design, by us

  • The two options the company did not take.
  • The numbers attached to all three options.
  • The odds and the timing of the delayed outcome.
  • Which option is structurally closest to the real one — recorded, never presented as the right answer.

The game never claims that an option a real company did not take would have changed what happened to it. Business outcomes are multi-causal; each case identifies one pivotal choice documented by its source, and nothing more than that.

Method

How the library was built

Each case was read for the pivotal choice its source documents, then written up as a scenario with three options: one disciplined, one aggressive, one that trades control for exposure. The structure is the same every time, so no option can be identified as the historical one by its shape.

Imported text is immutable. Whitespace is normalised and nothing else: a summary is never rewritten to read better, and an ambiguous or missing field is left empty and flagged for review rather than filled in. Where an effect token was not recognised, it was reported rather than quietly dropped.

Every row carries its own source link, which is what the game shows you after a decision resolves. The import is checksummed, so the content behind a given build can be identified exactly.

Validation of the shipped library

Rows read
300
Valid
300
Rejected
0
Duplicate decision IDs
0
Distinct source companies
242
Distinct source URLs on the rows
9
Unrecognised effect tokens
0
Rows with no approximate year
184

Content checksum of the accepted rows (SHA-256):

9238c49870ae798d98a0778b7c25d4e92352b43d0d1d0ab564df7807ba99dd6f

Provenance

Every source

Ten references, with the use each was put to and the reliability note recorded at import time. Links open in a new tab.

Source list as recorded in the import provenance appendix.
SourceUse and reliabilityAccessed
CB Insights — 483 Startup Failure Post-Mortems (opens in a new tab)https://www.cbinsights.com/research/startup-failure-post-mortem/Secondary compilation with founder statements, company statements, press reports, and primary-source links. Main source for 200 failure cases.2026-10-05
CB Insights — Biggest Startup Failures (opens in a new tab)https://www.cbinsights.com/research/biggest-startup-failures/Cross-check for funded failure cases and recurring failure mechanisms.2026-10-05
U.S. Chamber — Companies That Reinvented Their Business (opens in a new tab)https://www.uschamber.com/co/good-company/growth-studio/successful-companies-that-reinvented-their-businessSuccess and reinvention cases including Amazon, IBM, Netflix, Apple, LEGO, and others.2026-10-05
Harvard Business School Online — Business Strategy Examples (opens in a new tab)https://online.hbs.edu/blog/post/business-strategy-examplesValue-based strategy examples for Best Buy, Nike, and Starbucks.2026-10-05
Harvard Business Review — Box Pivot (opens in a new tab)https://hbr.org/podcast/2017/12/boxs-ceo-on-pivoting-to-the-enterprise-marketFirst-person interview on Box’s consumer-to-enterprise pivot, customer development, and technology timing.2026-10-05
Netflix — Completing the Cloud Migration (opens in a new tab)https://about.netflix.com/news/completing-the-netflix-cloud-migrationPrimary company account of the seven-year cloud-native migration and its operating consequences.2026-10-05
Cascade — 56 Strategy Examples (opens in a new tab)https://www.cascade.app/strategy-factory-report/56-strategy-examplesSecondary source for strategy cases across operations, marketing, supply chain, and culture.2026-10-05
Forbes — 14 Famous Business Pivots (opens in a new tab)https://www.forbes.com/sites/jasonnazar/2013/10/08/14-famous-business-pivots/Secondary summaries of product and market pivots.2026-10-05
Shopify Pivot Paper (opens in a new tab)https://oeconomia.pl/ceref/article/download/2301/2015Academic discussion of Shopify’s pivot from merchant to commerce platform.2026-10-05
Successful Pivot List (opens in a new tab)https://github.com/fikrikarim/companies-with-successful-pivotCurated source list used only for the Twitch pivot entry; production research should follow the linked original source.2026-10-05

During play

The company is hidden, then shown.

While a decision is open you see the situation and three options. The real company is not named, and neither is the option that matches it — if you recognised the company you would be recalling the ending instead of weighing the choice.

Once your choice is locked, a Real-World Parallel panel opens: the company, roughly when, the decision it faced, the verified result, the lesson its source supports, and the link. The panel repeats the boundary above, every time, so the distinction travels with the claim instead of living only on this page.

The company name is never used in a prompt, a player-facing label, or an analytics payload.