The canonical 7-step decision-making process is: identify the decision, gather relevant information, identify the alternatives, weigh the evidence (pros and cons of each), choose among alternatives, take action, and review the decision against the original need. Named alternatives include the DECIDE model (Kristina Guo, 2008: Define, Establish criteria, Consider alternatives, Identify the best, Develop a plan, Evaluate) for managers; the OODA loop (John Boyd: Observe, Orient, Decide, Act) run as a continuous loop for fast competitive situations, where Orient is the crucial step; RAPID (Bain and Company), which assigns decision roles — Recommend, Agree, Perform, Input, Decide — to fix accountability; and the Vroom-Yetton-Jago situational model, which tells a leader how much to involve the team. The most common failure is skipping the final review step. Argumentree supports the process with structured pro/con argument trees that make 'weigh the evidence' explicit, multi-dimensional rating that aggregates into consensus scores, AI extraction of arguments from transcripts and documents, role and department-based access that mirrors decision roles, and a full audit trail that closes the review loop.

From a problem to a choice you can defend — in seven steps. Plus the named models (DECIDE, OODA, RAPID) for when the basic process isn't enough.
The decision-making process turns a vague problem into a defensible choice through a repeatable sequence. The classic version has seven steps; specialized models like DECIDE, the OODA loop, and RAPID adapt it for healthcare, fast-moving competition, and unclear accountability. The one step everyone skips — reviewing the outcome — is the one that makes you better next time.
Name the real choice you face — clearly enough that everyone agrees what's being decided.
Collect the facts, data, and context relevant to the choice.
List the realistic options, including doing nothing.
Lay out the pros and cons of each option and evaluate them on the merits.
Select the option the evidence best supports.
Implement the choice.
Check the result against the need from step one — and learn from the gap.
The 7 steps are the default. These models adapt the process for specific conditions:
Define the problem · Establish criteria · Consider alternatives · Identify the best · Develop & implement a plan · Evaluate and monitor. A memorable six-step model originally written for healthcare managers.
Observe · Orient · Decide · Act — run as a continuous loop. Born in fighter-jet dogfighting, now used in business, cybersecurity, and litigation. Orient is the crucial step, and cycling faster than your opponent is the whole point.
Recommend · Agree · Perform · Input · Decide. Not steps — roles. RAPID fixes the "who actually decides?" problem by naming who recommends, who signs off, who executes, who's consulted, and the single person who decides.
A contingency model that tells a leader how much to involve the team — from deciding alone (autocratic) to reaching a collective decision together (group) — based on how much the decision needs quality, buy-in, and information.
Most processes break at "weigh the evidence" and "review the decision" — the parts that need structure and memory. Argumentree handles exactly those, built on argument mapping:
Every option's pros and cons live in a structured pro/con tree — the messy step 4 becomes visible and shared.
Participants rate arguments; ratings aggregate into net support scores, so "choose among alternatives" follows the evidence.
Roles and departments control who contributes, moderates, and decides — RAPID-style accountability without the spreadsheet.
The audit trail preserves the reasoning so step 7 — reviewing the decision — actually happens.
Part of the broader practice of decision making; see also the decision-making models behind these processes and collaborative decision making for running the process as a group, where the final step of reaching agreement is consensus building.
A weak step 1 poisons everything downstream — define the real decision first.
Almost nobody runs step 7, so the same mistakes repeat.
Looping 'gather information' forever until the window closes.
Without RAPID-style roles, decisions stall because everyone and no one owns them.
The decision-making process is a structured sequence of steps for moving from identifying a problem to choosing, acting on, and reviewing a course of action. The most common version has seven steps: identify the decision, gather information, identify alternatives, weigh the evidence, choose, act, and review the outcome.
(1) Identify the decision you need to make; (2) gather relevant information; (3) identify the alternatives; (4) weigh the evidence — the pros and cons of each option; (5) choose among the alternatives; (6) take action; and (7) review your decision against the need you identified in step one. Skipping step seven — the review — is the most common failure.
DECIDE is a six-step model published by Kristina Guo in 2008 for healthcare managers: Define the problem, Establish the criteria, Consider the alternatives, Identify the best alternative, Develop and implement a plan, and Evaluate and monitor the solution. The acronym makes the steps easy to remember.
The OODA loop — Observe, Orient, Decide, Act — was created by US Air Force strategist John Boyd for fast, competitive, changing situations. It runs as a continuous loop rather than a straight line, and the key step is Orient: making sense of observations through your experience and mental models. Whoever cycles through the loop faster gains the advantage.
RAPID, from Bain & Company, isn't a sequence of steps — it assigns decision roles so accountability is clear: Recommend (drives the proposal), Agree (must sign off), Perform (executes), Input (consulted), and Decide (the single person who makes the final call). It solves the 'who actually decides?' problem that stalls organizational decisions.
Make every step visible — weigh the evidence as a group, assign who decides, and keep the record. Bring structure to your process with Argumentree.
Start Free