Order Drop - Case Study: Diagnose a Sudden Drop in DoorDash Order Volume

Question

## Round 1 - Data Science Case Study

## Problem

DoorDash's order volume dropped 15% overnight in one metro area. You are the analyst on call. Walk through your full investigation.

**Data Available:**
- `orders(order_id, customer_id, restaurant_id, dasher_id, placed_at, status, order_total)`
- `restaurants(restaurant_id, name, category, is_active, city)`
- `app_events(event_id, user_id, event_type, platform, occurred_at)`

**Step 1 - Scope the drop:**
- Is the drop on all platforms or just iOS/Android/web?
- Is it all restaurant categories or specific ones?
- Is it all customer cohorts (new, returning) or one?

**Step 2 - Check data integrity:**
- Is the drop real or a logging failure? Check if `app_events` also dropped.

**Step 3 - Hypothesize causes:**
- Competitor promotion launched
- App deployment introduced checkout bug
- Major restaurant chain went offline
- Extreme weather event

**Step 4 - Validate and quantify:**
```sql
SELECT restaurant_id, COUNT(*) AS orders
FROM orders
WHERE placed_at BETWEEN '2024-06-01' AND '2024-06-02'
GROUP BY restaurant_id
ORDER BY orders;
```

## Follow-ups

1. How would you determine if the drop is seasonal vs. an anomaly?
2. What would you look at if order volume is fine but revenue dropped?
3. How do you communicate your findings to stakeholders before you have a root cause?
4. After identifying the cause, what process do you put in place to catch this faster next time?

Order Drop - Case Study: Diagnose a Sudden Drop in DoorDash Order Volume

Question Details

Full Details

Topics

More from DoorDash