The three can diverge and must be measured separately

What it is

Effectiveness, efficiency, and satisfaction are not aliases moving together: shortcuts raise expert efficiency while hurting novice completion; strong guidance raises completion while eroding experts' sense of control and their satisfaction. Any single average can hide the divergence, so the three must each be measured and each be reported.

Why it happens

Divergence arises because the components respond in different directions to the same change: effectiveness is shaped by structural completeness and fault tolerance, efficiency by operation-chain length, satisfaction by control and expectation. Aggregating them into one "usability score" averages opposing movements into a false "no change," while what actually happened is that one user group gained and another lost. Product positioning sets the weights among the three—and weight choice is a value judgment that measurement alone cannot make.

Studying it

Design studies to collect all three metric classes from the same participants on the same tasks, report pairwise correlations, and test whether divergence is significant. Analyze by user segments (novice/expert, frequent/infrequent), because divergence typically lives between groups. Any aggregated score must ship with the raw component values so a flat total cannot mask structural change.

Where it stops holding

The three do not always diverge: in mature designs they usually improve together, and divergence is a foreseeable exception, not the norm. Segmented analysis needs adequate samples—small-sample segment numbers are mostly noise. And because weights decide the verdict, usability data cannot decide alone: "whose which metric gets sacrificed" returns to product goals.

Applying it

• Keep a fixed three-column structure in test reports—effectiveness, efficiency, satisfaction—and never merge them into one score.
• In change reviews, name which component should improve, which may suffer, and why the trade-off is accepted.
• When divergence appears between groups, publish segmented conclusions to back differentiated strategies like expert modes or novice onboarding.

Related

• Same group: B5.02.1 Effectiveness asks whether the task gets done · B5.02.2 Efficiency asks how many resources the effort costs · B5.02.3 Satisfaction concerns the subjective experience
• Nearby: B5.03 Learnability · Q2 Usability Testing
• Search terms: usability trade-offs · metric divergence · segmented analysis