Pairwise Comparison

Overview

Pairwise comparison is a simple but effective prioritization technique. Instead of sorting a list all at once, items are compared in pairs — for each comparison, the team decides which is more important. The many individual decisions yield a clear ranking, even when the team would disagree on a global assessment.

Items to compare: The elements to prioritize — e.g. improvement ideas, requirements, risks, supplier options. Ideally 5–10; with more, the number of comparisons becomes unwieldy.

Pair comparison: Two items are compared directly — "Is A more important than B?". With n items there are n × (n−1) / 2 comparisons (e.g. 10 items → 45 pairs).

Rating scale: Each comparison is decided either binary (A or B) or weighted (e.g. 1–9 Saaty scale). Binary is faster, weighted gives finer differentiation.

Result: For each item, count how often it was chosen as more important. The sums give the ranking — highest score = highest priority.

Pairwise comparison especially fits when a team must order a list of poorly comparable options and the first direct-sort attempt ends in argument. Splitting into individual decisions makes the process more tangible and consensual.

Approach

• List the items to prioritize — uniquely and clearly phrased.
• Define the comparison criterion: "more important for …", "more urgent than …", "higher impact than …".
• Choose scale: binary (simple) or weighted (Saaty 1–9).
• Walk through all pairs as a team — briefly justify each decision.
• Calculate the score totals per item.
• Present the result as a ranking and sanity-check it with the team.

Pitfalls

Too many items: With 20 items there are 190 pairs — the session becomes torture and focus drops. Pre-filter roughly first, then compare pairwise.

Unclear comparison criterion: If it is not defined what is compared, everyone answers against a different criterion. Set it beforehand: "more important in terms of customer value", "in terms of cost", etc.

Allowing ties: Letting people pick "both equally important" avoids decisions — and makes the method ineffective. When in doubt, force a choice.

Inconsistent ratings: If A > B and B > C, then A > C should also hold. In practice this is not always true — systematic contradictions point to a fuzzy criterion or a derailed discussion.

Individual instead of team: A pairwise scoring done by one person reproduces their preferences. In a team, other perspectives surface — and the result is more robust.

Blindly accepting the result: The ranking is a decision aid, not a verdict. When gut feeling clearly disagrees, it is often the criterion or list that has a gap — not the method that failed.

Examples

This module ships with the following example datasets — load any of them in the app with a single click.

Available in the following cycles

• DMAIC: Analyze
• DMADV: Analyze
• 8D: D4 — Root Cause Analysis