Metrics Strategy Framework

A structured approach to answering metrics questions in data engineering interviews.

Approach to Answering Metrics Questions

Clarify the Question: Rephrase it even if it sounds obvious to ensure alignment.
Pause & Think: Jot down initial thoughts to structure your response.
Feature & Product Context: Explain how the feature relates to the product and how the product fits the overall mission.
User Impact Analysis: Identify the user personas impacted and potential segmentations (e.g., new vs. returning users, mobile vs. web, creators vs. viewers).
Metrics Categorization: List relevant metrics and classify them into:
- Leading Indicators: Short-term behavioral signals predicting success.
- Lagging (Trailing) Indicators: Long-term validation of the impact.
- Countermetrics: Guardrails against unintended negative effects.
North Star Metric (NSM) Selection: The guiding metric should:
- Be directly influenceable by the product/company.
- Impact all key user personas.
- Be reliably measurable.
Check External References: Use Amplitude guides or similar resources to cross-check relevant product metrics.
Consider Tradeoffs: Evaluate potential downsides (e.g., cannibalization, loss of brand equity, impact on profitability).
Visual Representation: When needed, diagram the metric relationships (e.g., in Excalidraw or CoderPad).

Scenario	NSM	Personas	Leading Metrics	Lagging Metrics	Countermetrics
Facebook Live Launch	Weekly Active Streamers	Content Creators, Viewers	New Streams Started	30-day Retention	User Reports
Instagram Shopping	Purchases Completed	Shoppers, Sellers, Advertisers	Add-to-Cart Rate	Repeat Buyers	Drop in Core Content Engagement
Facebook Groups + Threads	Thread Shares in Groups	Group Members, Thread Creators	Daily Shares	Monthly Group Activity	Group Abandonment
Declining Twitter Engagement	Daily Active Users (DAU)	Casual Users, Power Users, Advertisers	Tweet Frequency	Ad Revenue	Competitor Growth
Shopify Merchant Activity	Active Merchants	Merchants, Shoppers	New Merchant Sign-Ups	GMV	Merchant Churn
Netflix User Engagement	Total Hours Viewed	Viewers	New Subscriptions	Subscriber Retention	Content Abandonment
Uber Ride Transactions	Number of Rides	Riders, Drivers	New Rider Sign-Ups	Ride Frequency per User	Driver Churn
DoorDash Order Volume	Orders Delivered	Customers, Dashers, Restaurants	New Customer Sign-Ups	Repeat Orders	Delivery Complaints
Airbnb Booking Activity	Nights Booked	Guests, Hosts	New Listings	Repeat Bookings	Host Cancellations

DAU / WAU / MAU – Count of unique users active in a given day, week, or month. A high DAU/MAU ratio (stickiness) means users are returning frequently.
Session Duration – Average length of time a user session lasts. Longer sessions often imply higher engagement.
Retention Rate – Percentage of users who return or remain active after a given time period.
Active Engagement – Depth measures like sessions per user, actions per session, or DAU/WAU ratios.

ARPU (Average Revenue Per User) – Total revenue divided by number of users.
LTV (Customer Lifetime Value) – Total revenue expected from a customer over the relationship.
CAC (Customer Acquisition Cost) – Marketing and sales costs divided by new customers acquired.
Conversion Rate – Percentage of users who perform a desired action.
Churn Rate – Percentage of customers who discontinue service in a period.

Supply–Demand Balance – How well supply aligns with demand in a marketplace.
Utilization Rate – Percentage of resource capacity utilized.
Time to Resolution (TTR) – Average time to resolve issues or tickets.
Throughput & Capacity – Output over time (orders/hour, rides/day, requests/second).
Error Rates – Frequency of operational mistakes (order defect rate, fulfillment errors).

Fraud Rate – Fraudulent activities relative to overall usage.
Dispute Rate – Percentage of transactions resulting in disputes or chargebacks.
NPS (Net Promoter Score) – Index measuring customer loyalty (-100 to 100).
User Ratings & Reviews – Average ratings and sentiment from users.
Safety Incident Rate – Frequency of safety or policy violations.

Time – Year, quarter, month, week, day, hour. Trends and seasonality.
Geography – Country, region, state, city, ZIP code.
User Type & Demographics – New vs returning, demographics, customer tier, persona.
Device & Platform – Desktop, tablet, mobile, OS, browser, app version.
Acquisition Channel – Organic, paid ads, email, social, referral, campaign ID.
Content/Product Attributes – Category, type, genre, listing type, feature used.
Transaction Details – Payment method, order size, subscription type, discount usage.
Cohort – Grouping by signup or start date for lifecycle analysis.

Fact Table – Records events/transactions with quantitative metrics. Contains foreign keys to dimensions plus numeric measures (units_sold, revenue).
Dimension Table – Descriptive attributes providing context (who, what, where, when, how). Has primary key referenced by facts.
Primary Key (PK) – Uniquely identifies each record in a table.
Foreign Key (FK) – Links to primary key of another table, establishing relationships.

One-to-One (1→1) – Each record in A relates to at most one in B. Used to split entity attributes.
One-to-Many (1→∞) – One record in A associates with multiple in B. Most common (Customer → Orders).
Many-to-Many (∞→∞) – Resolved via junction table (Students ↔ Classes via Enrollments).

Shopify: Merchant → Products → Transactions ← Shopper

Netflix: User → ViewingSession ← Content

Uber: Driver → Ride ← Rider

DoorDash: Customer → Order ← Restaurant, Order ← Dasher

Airbnb: Host → Listing → Booking ← Guest

Understand Context & Goal – Clarify what’s being asked and why it matters.
Identify Key Metrics & Dimensions – Break into sub-questions with specific metrics.
Use a Structured Flow – Define → Diagnose → Interpret → Recommend.
Provide Comparison & Context – Compare to historical data, targets, or segments.
Visualize & Summarize – Charts/tables for key breakdowns, concise insights.
Actionable Recommendations – Conclude with what the business should do.

Ride-Sharing:

Social Media:

E-Commerce:

Streaming: