Amazon — Device Economics — Forecasting & Causal Inference Questions

Role context: Data Scientist II, Device Economics (Decision Science / DSO) · Est. study time: 60 min · 5 questions

How the science works here

For a forecasting role, "experimentation" is really forecasting methodology + causal measurement — you usually can't A/B a launch or a broad price change:

  • New-product forecasting with no history (analogs, diffusion, conjoint) under heavy uncertainty.
  • Cannibalization / substitution measured causally (synthetic control / difference-in-differences), because a launch can't be randomized.
  • Forecast evaluation on bias and decision-cost, not just average error.
  • Elasticity from natural experiments; portfolio coherence and go/no-go under uncertainty.

For the fundamentals — regression, significance, distributions — see the Probability & Statistics section.

Each answer is a coaching walkthrough: a Sample answer (clarify → approach → a simulated back-and-forth → a clear call), then a Deep dive with illustrative example, then a Grading rubric.

Questions (5)

Forecast a new device with no sales historyForecastingHardMeasure cannibalization from a launch you can't A/BCausal inferenceHardIs your forecast good? Evaluate it honestlyForecastingMediumEstimate price elasticity without a price experimentCausal inferenceHardMake the portfolio forecast coherent — and decide under uncertaintyDecide from resultsHard