A Consistent Measurement Logic
The framework translates UX behaviour into business impact using a single principle: impact is a function of behavioural change and exposure.
Rather than relying on relative uplift or raw analytics, the model:
– Measures conversion change in percentage points
– Weights impact by the proportion of users actually exposed
– Applies consistent time normalisation across initiatives
– Evaluates performance over multiple post-launch windows
This ensured that improvements were neither overstated nor dismissed.
Time-Based Validation
To avoid premature conclusions:
– Early post-launch windows captured adoption effects
– Later checkpoints confirmed behavioural stabilisation
This approach allowed the team to detect short-term volatility, long-term consistency, and false positives driven by novelty or traffic noise.
Portfolio-Level Visibility
Each UX change was documented in a dedicated update view and rolled into an overview layer showing journey step, relative exposure, direction and stability of impact, ROI tier classification, and confidence notes.
This shifted conversations from “Is this UX change good?” to “Where should we invest next for the strongest return?”
Discipline Through Rejection
Several commonly used ROI approaches were explicitly rejected:
– Relative uplift percentages that exaggerated deep-funnel impact
– Applying changes to total site traffic regardless of exposure
– Blind use of industry benchmarks without contextual adjustment
– Volatile revenue-per-session models
The final framework prioritised realism over persuasion.
