DoorDash · Product & Business Case

Diagnose completed orders drop in Los Angeles

TrueInterview

October 7, 2026 · 2 min read

Imagine you are a data scientist on DoorDash's consumer pricing team. Completed delivery orders in Los Angeles have fallen sharply over the past two weeks compared with both the historical baseline and comparable large metros. How would you investigate this issue end-to-end? Your answer should address:

Problem framing and validation
- How would you verify the decline is genuine instead of seasonality, reporting lag, instrumentation errors, or a short-lived outage?
- Which benchmarks would you rely on: week-over-week, year-over-year, pre/post launch, and matched control city comparisons?
Metrics and funnel decomposition Be specific about which metrics you would inspect and how each connects to completed orders. For instance:
- Demand: app opens, sessions, store views, checkout starts, order attempts
- Pricing and conversion: basket size, menu price index, delivery fee, service fee, surge or small-order fees, promotions, DashPass/member mix, checkout conversion
- Marketplace health: merchant availability, out-of-stock rate, merchant acceptance rate, courier supply, assignment time, ETA, cancellations
- Outcome metrics: completed orders, completed orders per active consumer, gross order value, contribution margin
Hypotheses Develop and rank plausible explanations, such as:
- a pricing change that lowered conversion
- reduced courier supply or longer ETAs
- merchant outages or thinner assortment availability
- product regressions in the app or checkout flow
- external factors like weather, major events, regulation, or competitor promotions
- mix shifts across neighborhoods, user cohorts, or dayparts
Segmentation and causal reasoning
- How would you split the analysis: new versus returning users, ZIP code, neighborhood, time of day, platform, DashPass versus non-member, cuisine, delivery versus pickup?
- How would you protect against confounding, Simpson’s paradox, and selection bias when reading the decline?
Recommendations
- If pricing looks like the primary cause, what immediate short-term and longer-term actions would you suggest?
- What trade-offs would you weigh among order volume, profitability, courier earnings, merchant health, and customer experience?
Experimentation Propose an experiment or quasi-experiment to test a remedy.
- Define the treatment and control groups
- Select a primary success metric and guardrail metrics
- Specify the randomization unit, duration, and power/MDE considerations
- Explain when an A/B test is preferable to a geo experiment or difference-in-differences design Overview: This question assesses a data scientist's skills in product analytics, causal inference, funnel and metric decomposition, segmentation, and experimentation for marketplace platforms, with a focus on diagnosing sustained declines in completed orders.

DoorDash · Product & Business Case

Diagnose completed orders drop in Los Angeles

TrueInterview

October 7, 2026 · 2 min read

Problem framing and validation
- How would you verify the decline is genuine instead of seasonality, reporting lag, instrumentation errors, or a short-lived outage?
- Which benchmarks would you rely on: week-over-week, year-over-year, pre/post launch, and matched control city comparisons?
Metrics and funnel decomposition Be specific about which metrics you would inspect and how each connects to completed orders. For instance:
- Demand: app opens, sessions, store views, checkout starts, order attempts
- Pricing and conversion: basket size, menu price index, delivery fee, service fee, surge or small-order fees, promotions, DashPass/member mix, checkout conversion
- Marketplace health: merchant availability, out-of-stock rate, merchant acceptance rate, courier supply, assignment time, ETA, cancellations
- Outcome metrics: completed orders, completed orders per active consumer, gross order value, contribution margin
Hypotheses Develop and rank plausible explanations, such as:
- a pricing change that lowered conversion
- reduced courier supply or longer ETAs
- merchant outages or thinner assortment availability
- product regressions in the app or checkout flow
- external factors like weather, major events, regulation, or competitor promotions
- mix shifts across neighborhoods, user cohorts, or dayparts
Segmentation and causal reasoning
- How would you split the analysis: new versus returning users, ZIP code, neighborhood, time of day, platform, DashPass versus non-member, cuisine, delivery versus pickup?
- How would you protect against confounding, Simpson’s paradox, and selection bias when reading the decline?
Recommendations
- If pricing looks like the primary cause, what immediate short-term and longer-term actions would you suggest?
- What trade-offs would you weigh among order volume, profitability, courier earnings, merchant health, and customer experience?
Experimentation Propose an experiment or quasi-experiment to test a remedy.
- Define the treatment and control groups
- Select a primary success metric and guardrail metrics
- Specify the randomization unit, duration, and power/MDE considerations
- Explain when an A/B test is preferable to a geo experiment or difference-in-differences design Overview: This question assesses a data scientist's skills in product analytics, causal inference, funnel and metric decomposition, segmentation, and experimentation for marketplace platforms, with a focus on diagnosing sustained declines in completed orders.