lockstep
4 min readNaman Jain

Decision matrix template (free, with a worked example)

A free weighted decision matrix template for spreadsheets: picking criteria and weights, a worked example, the Pugh variant, and mistakes that skew the score.

Three options, five people, an hour of debate, and no decision. Everyone argues from the criterion they care most about: cost, speed, risk. A decision matrix doesn't remove the judgment, but it puts every criterion and every opinion on one page, where the disagreement becomes specific.

This guide gives a template you can paste into Excel or Google Sheets, a worked example, and the mistakes that skew results.

What a decision matrix is

A decision matrix lists options against criteria, weights each criterion by importance, and scores how well each option meets it. Adding the weighted scores ranks the options. It's sometimes called a weighted scoring model.

The template

Paste this into a spreadsheet, starting in cell A1:

Criterion 1Criterion 2Criterion 3Criterion 4Total
Weight0.400.250.200.151.00
Option A
Option B
Option C

Score each cell from 1 (poor) to 5 (excellent). In the Total column for Option A (cell F3), use:

=SUMPRODUCT(B3:E3, $B$2:$E$2)

Copy it down for each option. Weights in row 2 should add up to 1.00. The same formula works in Excel and Google Sheets.

How to fill it in

  1. List real options. Two to five. Include "do nothing" if it's genuinely possible.
  2. Choose three to six criteria. Each should be something you can judge independently, such as cost, time to ship, risk, user impact, or maintenance effort. More criteria dilute the ones that matter.
  3. Set weights before scoring. Agreeing on weights first stops people from tuning them to favor their option afterward.
  4. Score independently, then compare. Have each person score alone, then discuss the cells where scores differ by two or more. That's where the real disagreement is.
  5. Multiply and add. The formula does it.
  6. Sanity-check the result. If the winner feels wrong, find the weight or score you disagree with and talk about that.

A worked example

A team choosing how to add search to its product:

User impactTime to shipRunning costMaintenanceTotal
Weight0.400.250.200.151.00
Database full-text search35544.05
Hosted search service54244.00
Self-hosted search engine52323.40

The first two options are effectively tied. The matrix has done its job: the debate narrows to whether the better results of the hosted service are worth the running cost. That's a question the decision owner can answer, with the trade-off written down.

The Pugh variant

The Pugh method, invented by Stuart Pugh for engineering design, compares options to a baseline instead of scoring from scratch. Pick one option as the baseline, then mark each other option better (+), the same (S), or worse (-) on each criterion, and total the marks. It's faster and less falsely precise, and works well early, when you don't trust fine-grained scores.

Mistakes that skew the score

  • Too many criteria. Ten criteria at 10 percent each let small factors outvote the one that matters.
  • Overlapping criteria. "Cost" and "budget impact" count the same thing twice.
  • Weights set after scoring. Weights quietly bend toward the favored option.
  • One person scoring. The matrix then just formalizes that person's opinion.
  • Treating the total as the decision. A 4.05 versus 4.00 is a tie, not a result.

After the matrix

A matrix records how the options compared, not what was decided or who decided it. Once the owner makes the call, write it down: the choice, the reason, the options rejected and why, and who approved it. That record is what the team, and any AI coding agent building the feature, will need later. A decision log template has the fields. For deciding who makes the call, see RACI vs DACI.

Where Lockstep fits

Lockstep keeps the record that comes after the matrix. It captures decisions from PRDs, Slack threads, and pull requests, routes each to its owner for one-click approval, and briefs the engineers and AI coding agents building against it. When the team revisits the call, the new version replaces the old one and both stay visible.

See how it works.

Sources checked 16 November 2026

For the DACI roles, see the DACI decision-making template. For how this fits with other frameworks, see product decision-making frameworks. For product decision frameworks more broadly, see the best decision log tools for product and engineering teams. For decisions where uncertainty is the hard part, see the decision tree template. For comparing competitors, see the competitive analysis template.

Frequently asked questions

What is a decision matrix?
A table that lists options in rows and criteria in columns, scores each option against each criterion, and weights the criteria by importance. Adding the weighted scores ranks the options. It's also called a weighted scoring model; a close variant is the Pugh matrix.
How do I make a weighted decision matrix?
List the options, choose three to six criteria, give each criterion a weight that adds up to 100 percent or a simple 1 to 5 scale, score every option against every criterion, multiply each score by its weight, and add up each row. Then discuss the result before deciding.
What is a Pugh matrix?
A variant invented by Stuart Pugh for engineering design decisions. Instead of scoring options from scratch, you compare each one to a baseline option and mark it better, the same, or worse on each criterion, then total the marks.
Can I use a decision matrix in Excel or Google Sheets?
Yes. Put options in rows, criteria in columns, weights in a row above the criteria, and use SUMPRODUCT of each option's scores with the weight row to get its total. The template below has the exact formula.
Should the highest score always win?
No. The matrix makes trade-offs visible and arguments concrete, but weights and scores are judgments. If the top score feels wrong, find out which weight or score you disagree with. The final call belongs to a named decision owner.

Keep reading