Novelty, Primacy & Carryover Effects

Novelty Effects, Primacy / Change Aversion & Long-term Carryover Effects formalize the behavioral and temporal dynamics separating transient experimental noise from sustainable steady-state business equilibria; Phenomena: 1) Novelty Effect: users interact heavily with novel UI affordances out of curiosity, generating an artificial 3-5 day performance spike that inevitably decays back to baseline; 2) Primacy / Change Aversion: experienced users resist structural interface shifts, creating transient negative performance dips that reverse into genuine long-term gains once habits adjust; 3) Carryover Effects: promotional interventions shift temporal demand, causing post-experiment revenue dips due to user stockpiling; Long-term Solutions: Long-term Holdout Groups (preserving a 1% control group over months) and Cohort Learning Curve tracking.

Major redesign A/B testing, evaluating promotional long-term ROIs, and resolving short-term vs long-term metric divergences.

Prevents launching features that appear successful purely due to temporary novelty curiosity while killing winning features that suffer initial change aversion.

Explain how plotting longitudinal learning curves across New vs Existing user cohorts isolates novelty curiosity from genuine utility?

Why are Universal Long-term Holdouts essential for measuring cumulative compounding gains of dozens of marginal feature releases?

Explain how Reverse Experiments (shutting off a mature feature for 5% of users) verify durable causal retention impact?

Why must promotional experiments monitor post-intervention periods for 14-30 days to account for intertemporal demand shifting?

Explain Susan Athey's Surrogate Index methodology predicting 6-month user LTV from 2-week short-term behavioral proxies?