Second-Order Decision Test
Trace who wins, who loses, and what happens next before making a bold move
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 96%
The Second-Order Decision Test evaluates a bold move as the opening action in a chain rather than as an isolated event. Leaders first define the desired outcome, then map every stakeholder affected by the decision, including employees, investors, partners, customers, competitors, and the public. For each group, they identify incentives, leverage, likely reactions, and available alternatives. The team then traces how those responses alter the original result through second- and third-order effects. A simple final rule acts as a warning signal: if a supposedly strategic move leaves every meaningful stakeholder worse off, the decision-maker is probably also creating its own defeat. Communication, timing, and stakeholder coordination are analyzed as parts of the decision rather than tasks deferred until afterward.
Origin
Extracted from Marketing Against the Grain during the hosts' analysis of OpenAI's board decision and its cascading effects on employees, investors, Microsoft, and competitors.
Core principles
- 01Major decisions create winners and losers
- 02Immediate actions trigger second- and third-order responses
- 03Stakeholders possess leverage and alternatives
- 04A move that harms everyone usually harms the decision-maker most
- 05Communication consequences belong inside the decision analysis
How to run it
- 1
Define the desired outcome
State precisely what the decision is supposed to change and how success will be measured. Separate the strategic goal from the proposed action.
Pro tip Ask whether another action could produce the same outcome with less collateral damage.
Watch out A dramatic action is not itself a strategic objective.
- 2
Map affected stakeholders
List everyone whose incentives, assets, employment, reputation, or plans will change. Include powerful partners and indirect dependants.
Pro tip Add competitors who may receive time, talent, or market opportunity as a result.
Watch out Missing one stakeholder with strong leverage can invalidate the analysis.
- 3
Model reactions
For each stakeholder, identify likely actions, best alternatives, and capacity to retaliate or exit. Trace what follows after those first reactions.
Pro tip Model at least a cooperative response, an exit response, and an adversarial response.
Watch out Do not assume stakeholders will passively accept a loss.
- 4
Run the winner-loser test
Compare the resulting position of each party with the status quo. If nobody gains sustainably, reconsider the decision and its design.
Pro tip Distinguish temporary symbolic wins from durable strategic gains.
Watch out A move that merely demonstrates authority may destroy the assets that authority was meant to govern.
- 5
Integrate timing and communication
Decide who must be briefed, what can be explained, and when the action should occur. Recalculate reactions using the actual communication plan.
Pro tip Treat surprise to essential partners as a cost requiring explicit justification.
Watch out Poor communication can reverse an otherwise defensible decision.
- 6
Redesign or proceed
Modify the action to contain unacceptable downstream effects, or proceed with mitigation plans attached. Record the assumptions that would trigger reconsideration.
Pro tip Use a pre-mortem to imagine the decision becoming a public failure.
Watch out Do not proceed merely because organizational momentum has formed around the proposal.
In the wild
A board considering an immediate CEO dismissal maps likely responses from employees, investors, strategic partners, customers, and competitors. It discovers that employees can leave together, a partner can hire them, financing may collapse, and competitors gain time. The board redesigns the transition around documented reasons, partner consultation, retention measures, and continuity planning.
→ The organization pursues its governance objective without needlessly destroying enterprise value.
Common mistakes
Stopping at the immediate outcome
Achieving the first action, such as removing a leader, says nothing about whether the resulting organization remains viable.
Ignoring stakeholder alternatives
Employees, partners, and investors may coordinate, exit, or recreate valuable assets elsewhere.
Separating communications from strategy
A vague or badly timed announcement changes stakeholder behavior and therefore changes the strategic result.
Is it for you?
Best for
Leadership teams evaluating reorganizations, dismissals, partnerships, investments, or other moves with multiple affected parties.
Not ideal for
Routine reversible decisions where analysis would cost more than a mistake.
From the transcript
“if you make a big bold move and everybody's a loser then you're the loser”
“they made a decision without considering any of the game theory or the second third order of events after that decision”
“think through the second third order impacts of big decisions that you're going to make”
From the episode
Is OpenAI Over? Breaking Down The Craziest 48 Hours In Silicon Valley (#175)