How to Master User Satisfaction: 4 Models Every Product Manager Should Use

Most people agree that being user‑ or customer‑centric is the foundation of any company. As competition intensifies and products become more homogeneous, user experience (UX) and satisfaction have become critical differentiators. JD.com uses satisfaction surveys not only for app evaluation but also for merchant operations, store services, and even cafeteria experiences, treating satisfaction as a "temperature gauge" for performance.

01 Quadrant Model

The quadrant model visualizes satisfaction by plotting weighted scores and satisfaction scores on a two‑dimensional chart, dividing indicators into four quadrants to identify improvement priorities.

Key steps when applying the quadrant model:

Define satisfaction indicators based on research objects, using existing theory or user research to extract metrics such as product quality, price, and variety.

Design satisfaction and weight scales ; while reliability and validity testing is ideal, quick surveys can still reveal improvement points.

Collect and analyze data by calculating average or weighted scores, plotting them, and categorizing each indicator into repair, opportunity, advantage, or maintenance zones.

02 KANO Model

The KANO model classifies user requirements based on their nonlinear impact on satisfaction, distinguishing attributes such as must‑be, one‑dimensional, attractive, indifferent, questionable, and reverse.

Example: evaluating whether JD.com store scores should include product‑quality assessment.

Design questionnaire with forward and reverse items to capture attitudes toward including product quality.

Survey and analyze data using cross‑tabulation to compute user proportions for each KANO attribute (S‑M, S‑A, etc.). The example shows product quality is a must‑be attribute; removing it sharply reduces satisfaction.

Calculate Better‑Worse coefficients ; a positive Better value (e.g., 0.40) indicates added features raise satisfaction, while a negative Worse value (e.g., –0.63) shows removal lowers satisfaction.

03 Analytic Hierarchy Process (AHP)

AHP builds a hierarchical structure of satisfaction indicators and uses pairwise comparison matrices to derive weights.

Main tasks:

Construct satisfaction hierarchy from theory or stakeholder research.

Build judgment matrix by comparing each pair of sub‑indicators; ensure reciprocal consistency (bij = 1/bji).

Compute weights using either the geometric mean (root) method or the product‑sum method.

Consistency test by calculating the maximum eigenvalue λmax, then CI and CR; a CR < 0.1 indicates acceptable consistency.

04 Structural Equation Modeling (SEM)

SEM tests hypotheses derived from theory, allowing verification of a satisfaction model and estimation of indicator weights.

Typical applications include CSI models (SCSB, ACSI, ECSI) for macro‑economic monitoring, but SEM can also support product‑level weight calculations via confirmatory factor analysis.

Common SEM tools: LISREL, AMOS, Mplus. Using AMOS, the workflow is:

Draw structural model based on experience or theory.

Map data variables to observed variables in the model.

Run analysis to obtain model fit indices and standardized path coefficients, which serve as weights.

In summary, satisfaction models are tools for monitoring user experience; even simple satisfaction feedback collection can provide valuable insights, but a user‑centric mindset is the prerequisite for effective use of any model.