Definition
Divergence between the price variation a customer experiences across encounters (time, channel, personalization tier) and the variation they perceive as acceptable, measured through explicit and behavioral signals.
The Measurement Gap
Dynamic pricing adjusts prices based on demand, competition, and customer segmentation, but every metric used to evaluate it treats the purchase as the finish line. Revenue lift measures whether dynamic pricing generates more revenue than static pricing, not whether it does so at the cost of customer trust. Conversion rate captures behavior at a single price point without measuring whether customers who see different prices across visits develop negative perceptions. NPS reflects overall brand sentiment but cannot isolate the contribution of pricing variability; by the time NPS drops, the cause is ambiguous. Cart abandonment rate measures the moment of decision but not the cumulative effect of price volatility on repeat visit behavior. CLV is a lagging indicator that reflects trust erosion months after the pricing decisions that caused it. The specific failure mode is that dynamic pricing generates variation exceeding customer fairness tolerance, eroding trust faster than incremental revenue justifies, and no existing metric measures the gap between experienced variation and tolerated variation.
Decision Supported:
Is dynamic pricing operating within the bounds of customer fairness tolerance, or is it generating price variation that erodes trust faster than it generates incremental revenue?
Formula
Plain language: Measure price variation experienced by customers (max price minus min price for the same item within a rolling window, across channels and customer segments). Compare to fairness thresholds derived from three independent signals: pricing complaint rate, cart abandonment behavior on price-volatile items, and direct survey data where available. Each signal provides an independent tolerance threshold.
Notation:
PFPG = (Variation_experienced - Variation_tolerated) / Variation_tolerated
Where Variation_tolerated is the minimum (most conservative) of three subcomponent thresholds:
Variation_tolerated = MIN(T_complaint, T_abandonment, T_survey)
Variables:
| Variable | Definition | Unit |
|---|---|---|
| PFPG | Price Fairness Perception Gap | Signed ratio |
| Variation_experienced | Average (max_price - min_price) / median_price for dynamically priced items, per customer-observable window (30 days) | % |
| T_complaint | Complaint threshold: the price variation level at which pricing-related complaint rate exceeds baseline by 2x or more | % |
| T_abandonment | Abandonment threshold: the price variation level at which cart abandonment rate on price-volatile items exceeds category average by 1.5x or more | % |
| T_survey | Survey threshold: the customer-reported acceptable price variation from direct fairness research (where available) | % |
| Variation_tolerated | MIN(T_complaint, T_abandonment, T_survey); uses the most conservative available signal | % |
Data readiness by subcomponent: T_complaint is computable from customer service data (Native/Integration). T_abandonment is derivable from e-commerce analytics (Derived). T_survey requires dedicated fairness research (Manual/Not Captured). Organizations can pilot PFPG with T_complaint and T_abandonment alone, adding T_survey as a refinement when survey data becomes available.
Proposed formula, pending validation pilots.