Analytic Hierarchy Process

Description

Structures a decision into a hierarchy of criteria, derives weights from pairwise comparisons and checks the consistency of judgements.

AHP decomposes a complex decision (choosing among design directions, vendors or roadmap items) into goal, criteria, sub-criteria and alternatives. Stakeholders compare elements pairwise on Saaty's 1-9 scale; the principal eigenvector of each comparison matrix gives weights, and the consistency ratio (target below 0.10) flags incoherent judgements. Weights combine to rank alternatives. Group judgements can be aggregated by geometric mean. AHP has known criticisms (rank reversal) but makes stakeholder priorities explicit, auditable and testable, a strong upgrade on informal prioritisation meetings.

Attributes

Loss type
Unmigrated
Effort
Days

Steps

  1. Define the decision goal, criteria (e.g. user value, effort, risk, strategic fit) and alternatives.
  2. Build the hierarchy and confirm criteria are independent and complete.
  3. Collect pairwise comparisons from stakeholders using the 1-9 scale.
  4. Compute weights (eigenvector) and the consistency ratio; revisit judgements above 0.10.
  5. Score alternatives on each criterion (pairwise or rated), synthesise and rank.
  6. Run sensitivity analysis on weights and share the rationale transparently.

Sources

  • Saaty, The Analytic Hierarchy Process (1980)
  • Saaty, Decision making with the analytic hierarchy process (2008), International Journal of Services Sciences
  • Feature Prioritization Methods (slide deck)

Also known as

  • AHP
  • Saaty's AHP
  • Pairwise comparison weighting

Categories

  • Strategy & Collaboration
  • Game Theory & Decision Science