Most startup decisions are reversible and should be made quickly at the lowest level. A small minority are irreversible and deserve careful analysis.
The single most expensive mistake in startup decision making is treating reversible decisions as if they were irreversible — running a two-week analysis on a pricing test that could be rolled back in a day, or waiting for consensus on a hiring decision that could be corrected in 90 days. The second most expensive is the opposite: treating irreversible decisions as if they were reversible — hiring a VP with three conversations, or picking a technical architecture in an afternoon. Getting these categories right is more important than any single framework.
Amazon's framing: Type 1 decisions are one-way doors — hard or impossible to reverse. Acquisitions, key executive hires, architectural rewrites, funding terms, category positioning. These deserve deep analysis, multiple perspectives, and time. Type 2 decisions are two-way doors — easily reversed. Pricing experiments, marketing copy, feature scope for a single sprint, most hiring decisions below director. These should be pushed down to the person closest to the work and made quickly. Most startups mistakenly run Type 1 processes on Type 2 decisions and lose weeks.
For decisions that cross functional lines, name a Driver (owns the process), Approver (single decision-maker), Contributors (provide input), and Informed (need to know outcome). Writing this down before analysis begins prevents the most common decision failure mode: five people in a room, no one knows who decides, decision gets remade three times over subsequent weeks. If you can't name a single Approver, the decision isn't ready to be made.
For any decision worth 30 minutes of exec time, require a written proposal: context, options considered, recommendation, risks, decision requested. 2-6 pages. Circulated 24-48 hours before the meeting. Meeting starts with 10 minutes of silent reading. This eliminates the pattern where the loudest voice or most senior person shapes the outcome regardless of merit — and produces a durable record of why decisions were made, which is invaluable when revisiting six months later.
The healthiest decision cultures separate the debate phase (all objections welcome, direct challenge encouraged) from the commit phase (once decided, everyone rows in the same direction publicly, even if they disagreed). The failure mode is decisions that get 'reopened' via side channels, water cooler dissent, or passive resistance — this makes every decision cost 3x more energy than it should. Naming this explicitly ('we've decided X, if you disagree please raise it now, otherwise we're committed') is worth the awkwardness.
Track it. Common pattern: a proposal gets circulated, and 3 weeks later no decision has been made because the approver 'wants more information.' Usually the information won't change the answer. Set an internal SLA: Type 2 decisions inside a week, Type 1 decisions inside a month unless the analysis genuinely requires more. Escalate decisions that stall past the SLA to a named tiebreaker (usually the CEO or function head).
Investor directory · Fundraising library · Articles A–Z · Company funding database