Skip to content This library is free, offered as dana. If it helps your work, donate here →
penguin tree ai
← back to libraryDonate here

This is one of 3,622 free prompts in the penguin tree ai library. Sign up once and the whole library opens, with new prompts by email.

open the library →

Activation funnel experiment tracker

Builds a structured experiment tracker for activation funnel optimization — logging each experiment by funnel stage, hypothesis, ICE score, status, and result, then producing a prioritized backlog and a running results…

download .md

Purpose

Turn a loose pile of activation test ideas into a structured, prioritized experiment tracker — each logged by funnel stage, hypothesis, ICE score, and status — plus a ranked backlog and a learnings summary that tells the team what to run next and what's already been proven. Solves the recurring problem of growth experiments living in scattered docs with no shared view of what's been tested, what won, and what to prioritize.

When to Use

Invoke this skill when the user asks to:

  • Log or organize a set of activation/onboarding experiments
  • Prioritize a backlog of growth test ideas across funnel stages
  • Score experiments by ICE or another prioritization framework
  • Summarize results and learnings from a quarter of experiments
  • Build a shared tracker for a growth or activation squad
  • Decide which activation experiment to run next

Inputs Required

Before starting, collect:

  • Funnel stages in scope: Which activation stages to track (e.g., signup → first action → aha moment → habit/return)
  • Experiment ideas: List of test ideas, each with at least a rough description
  • North-star activation metric: The single metric experiments are meant to move (e.g., % reaching aha within 7 days)
  • Stage conversion rates: Current conversion at each funnel stage, if known
  • Status of each idea: Backlog / running / completed — and results for any completed ones
  • Team velocity: Roughly how many experiments the team can run per cycle (defaults to 2/week if not provided)
  • Prioritization preference: ICE, RICE, or PXL (defaults to ICE if not specified)

If conversion rates are missing, score experiments on relative reasoning and flag the missing data as a measurement gap.

Instructions

Step 1: Map Experiments to Funnel Stages

Assign every experiment idea to exactly one activation funnel stage. Use this default stage model unless the user provides their own:

  • Acquisition → Signup: Reducing friction to account creation
  • Signup → First Action: Getting the user to do the first meaningful thing
  • First Action → Aha Moment: Delivering the core value experience
  • Aha → Habit/Return: Driving repeat use and retention onset

For each experiment, identify the stage conversion gap it targets. The biggest opportunity is usually the stage with the steepest drop-off, not the most ideas — flag any stage that is over-indexed on ideas but isn't the leakiest.

Step 2: Structure Each Experiment Record

For every experiment, produce a complete record with these fields:

  • ID: Stage-prefixed (e.g., ONB-01, AHA-03)
  • Funnel stage: From Step 1
  • Hypothesis: Stated in "We believe [change] will cause [outcome] because [reason]; we'll know we're right when [metric moves by X]" format
  • Primary metric: The single metric this experiment moves
  • Type: UI/copy / flow change / new feature / incentive / messaging/lifecycle
  • Status: Backlog / Designed / Running / Completed / Killed
  • Result (if completed): Win / Loss / Inconclusive, with the observed metric change

Rewrite any vague hypothesis into the structured format above — "improve onboarding" is not a hypothesis; "Adding a 3-step setup checklist will raise first-action completion from 40% to 55% because users currently don't know what to do first" is.

Step 3: Score and Prioritize the Backlog

Apply the user's prioritization framework (default ICE):

  • Impact (1–10): Expected lift on the activation metric, weighted by how leaky the targeted stage is
  • Confidence (1–10): Strength of evidence — prior data, user research, analogous wins. Penalize hunches with no signal.
  • Ease (1–10): Engineering and design effort to ship (10 = trivial)

ICE score = (Impact + Confidence + Ease) / 3, displayed to one decimal. For RICE, multiply Reach × Impact × Confidence and divide by Effort. Rank the backlog highest-to-lowest. Tie-break in favor of experiments targeting the leakiest stage.

Step 4: Build the Tracker Table

Render all experiments in a single table sorted by status (Running → Backlog by score → Completed). Columns: ID | Stage | Hypothesis (abbreviated) | Metric | Score | Status | Result.

Step 5: Write the Learnings Summary

For completed experiments, synthesize:

  • What won and why: Patterns across wins (e.g., "reducing first-screen choices consistently lifted activation")
  • What lost and why: Avoid repeating failed approaches
  • Stage health: Which stage improved, which is still the bottleneck
  • Next 3 to run: The top 3 backlog items by score, with a one-line rationale each

Step 6: Recommend the Next Cycle

Based on team velocity, propose which experiments to run next cycle — balanced so no single stage monopolizes capacity unless it's clearly the bottleneck. State the one experiment you'd run first and why.

Output Format

  • Structure: Funnel stage map → individual experiment records → ranked backlog table → learnings summary → next-cycle recommendation
  • Length: 800–1,200 words plus the tracker table
  • Format: Structured document with one master table; experiment records as labeled bullet groups
  • Sections: Stage Map → Experiment Records → Prioritized Tracker Table → Learnings Summary → Next Cycle Recommendation
  • Tone: Growth PM — data-grounded, decisive, allergic to vague hypotheses; states a clear next move

Quality Checklist

Before delivering output, verify:

  • [ ] Every experiment is mapped to exactly one funnel stage
  • [ ] Every hypothesis follows the "We believe... will cause... because... we'll know when..." structure with a quantified metric target
  • [ ] Each experiment has a complete ICE/RICE score with all component values shown
  • [ ] The tracker table is sorted by status then descending score
  • [ ] The learnings summary names a specific bottleneck stage, not "the funnel" generally
  • [ ] The recommendation names one experiment to run first with a stated rationale

Notes

  • The leakiest stage almost always deserves more experiment volume than it currently has — reallocate ideas toward it
  • Inconclusive results usually mean the experiment lacked sample size or moved a metric too far downstream; flag these for redesign rather than re-running as-is
  • Keep hypotheses tied to a single primary metric — experiments measuring three things at once rarely produce a clean decision
  • When conversion data is missing, the tracker still works for organizing and prioritizing, but flag the measurement gap as the highest-priority fix

How to use this skill

This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.

Installing in Claude Code:

  • ~/.claude/skills/<folder-name>/SKILL.md — auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)
  • <your-project>/.claude/skills/<folder-name>/SKILL.md — auto-loads only when working in that project (recommended for project-specific skills)

Restart Claude Code or open a new session to register.

Using in Claude.ai, ChatGPT, Gemini, or other AI tools:

  1. Drag this file into a new chat, or copy and paste its contents
  2. Send your task or request as your next message — the AI will apply the skill where it fits

For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills