G-C-MM-E Microdeck

Question 1

G-C-MM-E framework

Answer 1

G-C-MM-E is a strict, senior-level sequence for Meta PA questions: Goal → Classify → Metrics → Mechanism → Execution. It prevents jumping, forces structure, and ensures correct use of frameworks.

Question 2

Step G: The Most Important Clarification

Answer 2

Clarify the primary objective and the hard constraints: What are we optimizing for (engagement, quality, growth, monetization), who is the main user, and what must NOT get worse (revenue, integrity, latency)?

Question 3

Step C: Triggering the Right Frameworks

Answer 3

Classify:
(1) product archetype (what is the product? –> attention engine, transaction engine, utility engine) &
problem archetype (what is broken? –> quality, volume, rate, depth, integrity, latency)
(2) metric family (volume, rate, time, monetary)
(3) stakeholders (users, advertisers, ecosystem health (integrity), revenue).
This tells you which sub-frameworks to use next.

Question 4

Step M1: Defining the Metrics Correctly

Answer 4

Always:
(1) define conceptual dimensions (example: relevance, utility, trust and safety, user experience)
(2) choose a concrete North Star Rate
(3) add guardrails
(4) apply decomposition such as Volume = Users × Frequency × Rate × Depth.

Question 5

Rate

Answer 5

Rate = probability of the target action per opportunity. It becomes click-through rate, conversion rate, quality-weighted engagement rate, or activation rate depending on the surface. This makes the decomposition universal.

Universal primitive formula:
Rate =Target events/opportunities

Question 6

Step M2: Mechanism (System Map)

Answer 6

Map the system: Entry → First Action → Quality or Conversion → Depth → Retention, and Upstream → Core → Downstream. This reveals where problems originate and where to intervene.

Question 7

Step E1: Diagnostics

Answer 7

Segment → Localize → Compare → Hypothesize → Validate. This prevents random guessing and turns debugging into a scientific process.

Question 8

Step E2: Improvement

Answer 8

Use experimentation and tradeoff principles: propose levers (ranking, signals, supply, UX), design experiments with NSM + guardrails, evaluate long-term user value, and balance user experience, revenue, and integrity.

Question 9

What are the three product archetypes at Meta?

Answer 9

Attention Engine (maximize meaningful engagement)
examples:
Instagram/Facebook Feed, Instagram Reels, Facebook Stories, Explore tab, Watch tab, TikTok-style immersive video surfaces, photo grid browsing, scrolling surfaces, infinite feed experiences.
2) Transaction Engine (match two parties + conversions)
Ads, Marketplace, Facebook Shops, Instagram Shops, Search, Ads ranking/auction, Lead-generation ads, Dating match flows, People You May Know, Friend Suggestions, job recommendations, content recommendations that trigger downstream conversions.
3)Utility Engine (task completion efficiency).
e.g.
- Examples:
Login, logout, onboarding flow, account recovery, password reset, settings, privacy center, notifications center, sharing flow, posting/upload flow, report a post, 2-factor authentication flow, scheduling a post, marketplace listing creation, profile completion, contact importer, app permission dialogs.

Question 10

What are the main problem archetypes you classify after identifying the product archetype?

Answer 10

• Quality (Irrelevant ads, spammy feed content, misleading posts, low-quality supply, bad recommendations, low satisfaction, hide/report spikes, quick scrolls, user complaints about relevance.)
• Volume (users or traffic) (e.g: Irrelevant ads, spammy feed content, misleading posts, low-quality supply, bad recommendations, low satisfaction, hide/report spikes, quick scrolls, user complaints about relevance.)
• Rate (CTR/Conversion /Activation problems),e.g: Declining click-through rate on ads or feed, conversion rate drop, activation rate drop in onboarding, fewer replies in messaging, fewer saves/shares, drop in meaningful interactions. - Depth (shorter sessions/purchases, How much a user does after converting?), e.g: Users watching fewer videos per session, lower purchase quantity per user, fewer items added to cart, shorter conversations in messaging, reduced story viewing sequence
• Integrity (spam/violations)
  e.g: Increase in spam, bots, policy violations, harmful content, scams in marketplace, fake ads, malicious links, rise in reports/hides, increase in enforcement actions.
• Latency/Slow UX Problem e.g: Slow feed load, slow video start, laggy UI, high crash rate, app freezes, delayed message sends, slow page transitions, poor responsiveness leading to lower engagement.

Question 11

How do I pick the metric family?

Answer 11

Identify the problem archetype first.
Quality → Rate
Volume → Volume, Latency → Time
Monetization → Monetary
The surface (feed/ads/etc.) does NOT determine the metric family.

Question 12

What is a quality-weighted engagement rate?

Answer 12

A rate metric = (weighted positive signals – weighted negative signals) ÷ impressions. It captures true user ad experience and relevance by combining likes, saves, shares, clicks, hides, and reports into a single probability.