Setting priorities — sound familiar?
AHP derives weights from pairwise comparisons — and checks consistency.
01 · PAIRWISE MATRIX & WEIGHTS
Compare two criteria at a time.
Geometric mean of all judgments → eigenvalue weights.
Respondents compare each pair of criteria on the standard 1–9 Saaty scale. Skari aggregates the responses using the geometric mean, builds the reciprocal pairwise matrix, and derives priority weights from the principal eigenvalue.
Top-level Pairwise Matrix
CR=0.093 ✓| Price | Quality | Delivery | Weight | |
|---|---|---|---|---|
| Price | 1 | 3.00 | 9.00 | 64.84% |
| Quality | 1/3.0 | 1 | 8.00 | 29.97% |
| Delivery | 1/9.0 | 1/8.0 | 1 | 5.19% |
Hierarchy Builder
Up to 5 levelsGlobal weight = product of all level weights from Goal down to each leaf criterion.
02 · UP TO 5 HIERARCHY LEVELS
Nest sub-criteria up to 5 levels deep.
Goal → Criteria → Sub-criteria → … → Leaf.
Real decisions are rarely flat. Skari lets you build a full decision hierarchy — up to 5 levels — so "Quality" can have sub-criteria like Durability and Performance, and each sub-criterion can be broken down further. Global weights are computed as the product of all level weights from the goal to each leaf.
03 · CONSISTENCY ANALYSIS
Know if judgments are coherent.
CR ≤ 0.10 = acceptable. Higher = re-judge flagged.
Skari reports λmax, CI, RI, and CR for every pairwise matrix. The CR Distribution chart shows how many respondents fall in each consistency band, and any CR ≥ 0.10 is flagged so you can identify and re-contact inconsistent respondents.
Consistency Analysis
CR=0.093 ✓CR Distribution
| Metric | Value | Note |
|---|---|---|
| λ max | 3.1078 | Ideal = 3 |
| CI | 0.0539 | (λmax − n) / (n − 1) |
| RI | 0.5800 | For n = 3 |
| CR | 0.0930 | ✓ Acceptable |
Agreement & Concentration
KENDALL'S W
1.00
Strong agreement
CONSENSUS
100%
rank concordance · n=30
MEAN CV
4%
Tight estimates
HHI
0.51
Fairly balanced
Weight Dispersion by Criterion
Wide box / high CV = respondents disagree on that criterion's importance.
04 · AGREEMENT & CONCENTRATION
Do respondents agree?
Kendall's W, HHI, and weight dispersion in one panel.
Beyond individual weights, Skari measures how much respondents agree (Kendall's W and rank concordance), how concentrated the weights are (HHI), and how spread each criterion's weight is across respondents (IQR box and CV). High dispersion on one criterion is a signal to run segment analysis.
05 · SEGMENT AHP
Do different groups prioritize differently?
Compare priority weights across respondent segments.
Select any survey question to split respondents into groups — by age, region, job role, or any categorical variable. Skari computes separate AHP weights for each segment and shows a heatmap table where the highest weight per criterion is flagged with a ▲ marker.
Segment AHP
Age range| Criterion | 45–54 (n=12) | 18–24 (n=6) | 65+ (n=4) | 35–44 (n=2) |
|---|---|---|---|---|
| Price | 15.7% | 16.6% | 15.7% | 15.7% |
| Quality | 59.4% | 62.6%▲ | 59.4% | 59.4% |
| Delivery | 24.9% | 20.9% | 24.9% | 24.9% |
Heatmap: deeper green = higher weight · ▲ marks each criterion's top segment
Final Ranking
Global weight · all levels
| Rank | Criterion | Parent | Global Wt. | CR |
|---|---|---|---|---|
| 1 | Price | — | 64.84% | 0.093 ✓ |
| 2 | ㅂ | Quality | 14.98% | 0.093 ✓ |
| 3 | ㅈ | Quality | 14.98% | 0.093 ✓ |
| 4 | Delivery | — | 5.19% | 0.093 ✓ |
06 · FINAL RANKING & RADAR
One table. Every criterion ranked.
Global weight, parent, and CR — all in one view.
The Final Ranking table lists every criterion and sub-criterion by global weight, shows its parent in the hierarchy, and flags the CR. The Radar View gives a visual summary of how priorities are distributed — making it easy to communicate the result to stakeholders.
IN THE REPORT
Everything AHP Priority delivers.
Pairwise judgments in — weighted priorities out.
Pairwise matrix & weights
Derive priority weights from pairwise comparisons.
Up to 5 hierarchy levels
Model goals, criteria, sub-criteria, and alternatives.
Consistency ratio (CR)
Flag inconsistent judgments (CR > 0.1).
Agreement & concentration
How aligned respondents are, and how concentrated weights are.
Segment AHP
Compare priorities across respondent groups.
Final ranking & radar
Rank alternatives with a radar profile.
Group AHP
Aggregate judgments across many evaluators.
Auto interpretation
Plain-language reading of weights and consistency.
ALSO IN SURVEY ANALYSIS
Explore more survey tools.
Each survey analysis method answers a different question — same interface, different lens.
Prioritize with rigor.
Pairwise judgments in · eigenvalue weights, CR, segment analysis, and final ranking out.