sunny.
← Back to PRD GeneratorReference

Input/Output Metrics Methodology (Aakash Gupta)

Core Concept

Every product metric is either an input (something you can directly influence through product changes) or an output (a business result that emerges from multiple inputs).

The mistake most PMs make: they set output metrics as goals without understanding which inputs drive them. Then when the output metric doesn't move, they don't know what lever to pull.

How It Works

Output Metrics (Lagging)

These are the business results you care about:

  • Revenue, MRR, ARR
  • Retention rate
  • NPS / CSAT
  • Market share
  • LTV

You can't directly change these. They're the RESULT of input metrics moving.

Input Metrics (Leading)

These are the actions and behaviors you CAN influence:

  • Activation rate (% of signups who reach "aha moment")
  • Feature adoption (% of users who try new feature in first 7 days)
  • Session frequency (sessions per user per week)
  • Task completion rate
  • Time to value (how fast users get first benefit)

Building a Metrics Framework for Your PRD

Step 1: Define the Output Metric

"What business result does this feature ultimately serve?" Example: Increase trial-to-paid conversion from 12% to 18%.

Step 2: Identify Input Metrics

"What user behaviors, if changed, would move the output?" Example:

  • % of trial users who complete onboarding (currently 64%)
  • % of trial users who invite a teammate (currently 8%)
  • % of trial users who use core feature 3+ times in first week (currently 31%)

Step 3: Target the Input You're Changing

"Which input metric does THIS feature specifically improve?" Example: This PRD targets onboarding completion rate (64% -> 80%).

Step 4: Connect Input to Output

"What's the evidence that improving this input will improve the output?" Example: Users who complete onboarding convert at 28% vs 4% for those who don't. If we move completion from 64% to 80%, modeled impact is +3.8pp on conversion.

Using This in PRD Section 3

The Goals & Success Metrics table should contain BOTH input and output metrics:

| Metric | Type | Current | Target | Method | |--------|------|---------|--------|--------| | Trial-to-paid conversion | Output | 12% | 15.8% | Stripe dashboard | | Onboarding completion rate | Input | 64% | 80% | Product analytics | | Time to first value action | Input | 4.2 days | 1.5 days | Event tracking |

This structure shows the causal chain: your feature changes inputs, inputs drive outputs.

Connection to AARRR Pirate Metrics

Aakash's input/output model maps cleanly onto the AARRR funnel:

  • Acquisition — Input: signup sources. Output: total signups.
  • Activation — Input: onboarding steps completed. Output: activated users.
  • Retention — Input: feature usage frequency. Output: D7/D30 retention.
  • Revenue — Input: upgrade prompts seen. Output: conversion rate.
  • Referral — Input: share actions taken. Output: viral coefficient.

Each stage has inputs you can influence and outputs you measure.

← Back to PRD Generator