How do I compare several good options using decision criteria?

Use a grid in two passes

First screen for fit: remove an option that cannot meet the decision's purpose or crosses a genuine boundary. Then compare the viable options against the same criteria. This keeps a strong preference on one row from compensating for a requirement that the option does not meet.

The University of Minnesota Extension describes a criteria grid that lists selection criteria and compares every option across them. It suggests five to nine criteria for its group exercise and offers weighting when choices are close. For an everyday personal choice, use only the criteria needed to expose the real trade-off; the source's suggested count is not a universal rule.

Use the ALIGN sequence

ALIGN is a Forever Free Compass framework, not an external research finding. Write one short line for each part before choosing a preferred option.

1. Aim: state the exact choice, whose interests count, and the time horizon you are comparing.
2. Limits: list the permissions, costs, deadlines, commitments, safety conditions, or other boundaries every viable option must respect.
3. Identical criteria: define the same small set of decision-relevant criteria for every option.
4. Ground each rating: attach an observation, source, estimate, or explicit assumption to each judgment.
5. Name the trade-off: identify the criterion that creates the lead, the uncertainty that could reverse it, and the next move you can justify.

Define the aim before the criteria

Write the decision as one sentence: choose among named options for a stated purpose by a stated review point. A criterion belongs in the grid only if different performance on it could reasonably affect that decision. Define vague labels before rating them: for example, replace 'fit' with the exact feature you mean, such as schedule compatibility, opportunity to learn, or effect on another person.

HM Treasury's 2026 Green Book says appraisal starts from objectives and critical success factors, and that constraints can limit which options are viable. It also calls for a proportionate process. That guidance governs public proposals; ALIGN borrows only the order of clarifying purpose, screening necessities, and then comparing viable choices for a low-stakes personal decision.

Separate limits from comparative criteria

A limit is a pass-or-stop condition: required permission, an unaffordable maximum cost, a fixed deadline, a commitment you will not break, or a safety condition. A criterion describes a difference you are willing to trade against another difference, such as flexibility, learning value, convenience, or reversible cost. Do not hide a real boundary inside a weighted total.

The Green Book says options that fail the objectives or critical success factors should not proceed, while later comparison examines costs, benefits, risks, and uncertainty. For an ordinary choice, this supports a simple discipline: screen non-negotiables before scoring preferences. It does not establish which personal boundary you must choose.

Compare like with like and show the basis

Use the same definition and scale for every option. A plain-language scale such as 'does not meet', 'partly meets', and 'meets well' is often easier to audit than invented precision. Beside every rating, mark its basis as observed, sourced, estimated, or assumed. If you cannot explain a rating in one sentence, leave it uncertain rather than filling the cell with confidence you do not have.

HM Treasury's multi-criteria guidance says criteria should capture ways options differ that matter, may be qualitative or quantitative, and should make trade-offs transparent. The formal method it describes can be resource-intensive and calls for specialist facilitation. A personal ALIGN grid is therefore not formal multi-criteria decision analysis and should not borrow its authority.

Read the result instead of obeying it

When one option leads, point to the cells that create the difference. Check whether the lead comes from duplicated criteria, an unsupported rating, or a weight chosen after seeing the result. If a small plausible change reverses the order, call it a near-tie rather than manufacturing a precise winner. You may need better evidence, a combination of options, a sequence, or a bounded reversible test.

The 2026 Green Book warns that simple weighting and scoring can lack an objective basis and reduce transparency. Its supplementary guidance distinguishes formal analysis of trade-offs from a simple total. The practical lesson is not to abandon a grid; it is to keep the evidence, assumptions, and deciding trade-off visible beside the number.

Raise the standard when consequences are material

ALIGN is for bounded, reversible personal reflection. It is not suitable by itself for medical, legal, financial, employment, safeguarding, privacy, or other materially consequential decisions. A neat grid cannot waive consent, policy, professional duties, or another person's rights, and a high total does not convert uncertain evidence into fact.

For a consequential choice, use the required domain process and qualified guidance. Preserve the comparison as a record of your questions and assumptions, not as proof that a decision is safe, compliant, fair, or optimal.