Decision Right Sizing
Decision right-sizing is the discipline of matching the rigor of a decision process to the reversibility and stakes of the decision itself. The frame popularized by Jeff Bezos: most decisions are 'two-way doors' (reversible at low cost — make them fast, with whoever has the most context) and a small number are 'one-way doors' (irreversible or very costly to undo — make them slowly, with high consensus). The failure mode is treating ALL decisions as one-way doors, which is the default in most large organizations because the cost of being wrong is salient and the cost of being slow is invisible. Right-sizing means defending speed for the 90% of decisions that are reversible and reserving consensus for the 10% that aren't.
The Trap
The trap is the asymmetric punishment of bad decisions vs. slow decisions. Leaders get fired for bad calls; almost nobody gets fired for slow ones. So everyone defaults to over-deliberating every decision to manage personal career risk, even when the actual business cost of slow decisions vastly exceeds the risk-adjusted cost of being wrong on reversible bets. The other trap is binary thinking — treating right-sizing as 'fast vs. slow' instead of recognizing a gradient. A vendor selection is a two-way door but still warrants a week of comparison; a marketing copy A/B test is a two-way door that warrants minutes. Right-sizing is calibration, not just permission to be fast.
What to Do
Build a four-tier decision taxonomy: (T1) Two-way door, low cost — single owner decides, no approval needed (e.g., copy variants, daily ops). (T2) Two-way door, moderate cost — owner decides with one peer review (e.g., feature scoping, vendor under $50K). (T3) One-way door OR high cost — multi-stakeholder review, formal write-up, named decider (e.g., pricing model change, exec hire, major vendor). (T4) Existential / strategic — full executive deliberation, board input where relevant (e.g., M&A, market entry/exit, layoffs). Publish the taxonomy with examples. Most teams discover that 70-80% of decisions they currently route as T3 should actually be T1 or T2. The taxonomy is permission to right-size.
Formula
In Practice
Jeff Bezos's 1997 shareholder letter introduced the one-way door / two-way door framing: 'Some decisions are consequential and irreversible or nearly irreversible — one-way doors — and these decisions must be made methodically, carefully, slowly, with great deliberation and consultation. If you walk through and don't like what you see on the other side, you can't get back to where you were before. Type 2 decisions are like walking through a door — if you don't like the decision, you can reopen the door and go back through. Type 2 decisions can and should be made quickly by high judgment individuals or small groups.' Bezos went on to argue that as organizations get larger, they tend to use heavyweight Type 1 processes for Type 2 decisions, which produces 'slowness, unthoughtful risk aversion, failure to experiment sufficiently, and consequently diminished invention.' Andy Grove's 'High Output Management' (1983) made the parallel point that managers should match the depth of their involvement to whether the decision is teachable or terminal — over-involvement in routine decisions starves the strategic ones of attention.
Pro Tips
- 01
If the decision can be reversed in <30 days at <2% of annual revenue cost, it's a two-way door. Apply this test before scheduling any cross-functional meeting to discuss it. The number of meetings that get cancelled is a measure of how badly your org has been over-processing.
- 02
Name the worst-case outcome before deciding the process. 'If we get this wrong, we lose 3 months of engineering time' is a clear cost; 'if we get this wrong, the company will collapse' is rarely true. Forcing leaders to articulate the actual downside almost always shrinks the perceived stakes.
- 03
Two-way door decisions made fast and reversed once are still net faster than one-way door processes that decide correctly the first time. Speed-with-reversal is a winning strategy for the vast majority of operational decisions; perfectionism is the losing one.
Myth vs Reality
Myth
“Important decisions deserve heavy process”
Reality
Importance and reversibility are different dimensions. An important reversible decision (like the launch sequence for a major feature) deserves speed; an unimportant irreversible decision (like a vendor lock-in for a low-priority system) still deserves diligence. Right-sizing is about reversibility-and-cost, not importance alone.
Myth
“Consensus decisions are higher quality”
Reality
McKinsey research on decision effectiveness consistently shows the opposite: consensus correlates with lower decision quality on most operational decisions because it dilutes accountability and selects for the lowest-common-denominator option. Single-decider models with named consultable parties (DACI, RAPID) outperform consensus on speed AND quality for the vast majority of two-way door decisions.
Try it
Run the numbers.
Pressure-test the concept against your own knowledge — answer the challenge or try the live scenario.
Knowledge Check
A product manager wants to ship an A/B test of a new pricing page CTA. They've scheduled a 3-week review involving design, brand, legal, growth, and the CRO before launching. The change can be rolled back in 5 minutes. What's wrong with this process?
Industry benchmarks
Is your number good?
Calibrate against real-world tiers. Use these ranges as targets — not absolutes.
% of Decisions Truly Requiring Heavyweight Cross-Functional Process
Operational decisions in companies with 200+ employeesCalibrated (one-way doors only)
10-15%
Slight Over-Processing
15-25%
Over-Processing
25-50%
Heavy Over-Processing
50-75%
Consensus Default
>75%
Source: Bain 'Decide & Deliver' (Blenko/Mankins/Rogers, 2010); McKinsey decision-making research (De Smet et al., 2019)
Real-world cases
Companies that lived this.
Verified narratives with the numbers that prove (or break) the concept.
Amazon
1997-present
Bezos's 1997 shareholder letter formalized the one-way door / two-way door distinction and made it part of Amazon's operating culture. The practical effect: PR/FAQ documents and 6-page memos are reserved for genuinely consequential decisions, while the vast majority of operational choices are made by single owners with explicit DRI (Directly Responsible Individual) ownership. The discipline is a major reason Amazon can operate dozens of business units at scale without consensus paralysis. The same culture also produced famous 'fast reversal' moments — the Fire Phone was launched, evaluated, and killed within 14 months once evidence was clear. Treating it as a two-way door reduced the cost of the bad call.
Concept Origin
1997 Shareholder Letter
Operating Application
DRI model + 6-page memo discipline
Notable Two-Way Door Reversal
Fire Phone launched 2014 → killed 2015
The framework only works if leaders BOTH classify decisions correctly AND accept the implication: two-way doors get fast unilateral decisions, even when they're large. Companies that adopt the language but route everything to consensus haven't actually adopted the framework.
Intel (Andy Grove era)
1980s-1990s
Andy Grove's 'High Output Management' (1983) and his subsequent practice at Intel established the discipline of matching managerial involvement depth to decision type. Grove argued that the manager's job is to ALLOCATE attention to the decisions where their involvement adds value, and to delegate (with clear principles) the rest. The famous Grove One-on-One was itself a right-sizing tool: it created a high-bandwidth weekly forum for the small number of decisions that needed manager input, freeing the rest of the week from interrupt-driven decision routing.
Framework Source
'High Output Management' (1983)
Operating Mechanism
Weekly 1:1 + delegated decision principles
Right-sizing is not just an organizational policy — it requires individual managerial discipline to NOT involve themselves in routine decisions just because they could. The forum design (weekly 1:1) and the delegation principle work together.
Related concepts
Keep connecting.
The concepts that orbit this one — each one sharpens the others.
Beyond the concept
Turn Decision Right Sizing into a live operating decision.
Use this concept as the framing layer, then move into a diagnostic if it maps directly to a current bottleneck.
Typical response time: 24h · No retainer required
Turn Decision Right Sizing into a live operating decision.
Use Decision Right Sizing as the framing layer, then move into diagnostics or advisory if this maps directly to a current business bottleneck.