Experiment Prioritization (ICE/PIE)
Prioritize your growth experiments and product features using the ICE (Impact, Confidence, Ease) and PIE (Potential, Importance, Ease) frameworks.
Formula
ICE = Impact × Confidence × Ease
ICE score (GrowthHackers) multiplies the factors to penalize low scores heavily. PIE (WiderFunnel) typically averages them. Both frameworks aim to identify 'Quick Wins' (High Impact, High Ease) and de-prioritize 'Money Pits' (Low Impact, Low Ease).
Worked Examples
Example 1: Quick Win
Problem:I: 9, C: 8, E: 9
Solution:ICE = 9 * 8 * 9 = 648. Top priority.
Result:Score: 648
Example 2: Moonshot
Problem:I: 10, C: 3, E: 2
Solution:ICE = 10 * 3 * 2 = 60. High impact but too hard/risky.
Result:Score: 60
Example 3: Fill-in Task
Problem:I: 2, C: 10, E: 10
Solution:ICE = 2 * 10 * 10 = 200. Easy to do, but low value.
Result:Score: 200
Frequently Asked Questions
What is 'Ease'?
Ease is the inverse of Effort. If a task is very hard (Effort = 10), Ease is low (Ease = 1 or 2). If it takes 5 minutes, Ease is 10.
How do I estimate 'Confidence'?
Use data. 10 = You have run this exact test before. 5 = You have industry benchmarks. 1 = Pure gut feeling.
ICE vs RICE?
ICE is better for rapid growth experiments (Velocity). RICE adds 'Reach' and is better for product features (Roadmap).
Should I multiply or average?
Multiplying (ICE) creates wider separation between scores, making top priorities stand out more clearly than averaging.
What is PIE?
Potential (how much improvement can be made?), Importance (how valuable is the traffic?), Ease. Popular in CRO (Conversion Rate Optimization).
How often should I re-score?
Weekly or before every sprint planning. New data changes Confidence scores rapidly.
What is a good ICE score?
It's relative. Compare ideas against each other, not against a universal benchmark.
Who invented ICE?
Sean Ellis, who coined the term 'Growth Hacking', popularized ICE to prioritize growth tests.
Background & Theory
The ICE Framework
- Impact (1-10): If this works, how big is the win? (1 = minimal, 10 = massive game changer)
- Confidence (1-10): How sure are we that it will work? (1 = guess, 10 = proven data)
- Ease (1-10): How easy is it to implement? (1 = months of dev, 10 = 5 minutes)
The PIE Framework
- Potential (1-10): How much improvement is possible on this page? (e.g., a terrible landing page has high potential)
- Importance (1-10): How valuable is the traffic? (Checkout page > About Us page)
- Ease (1-10): Technical/political ease of implementation.
Bias Checking
Humans tend to overestimate Impact and Ease. To combat this, require "Proof" for any Confidence score above 5. No data? Confidence is capped at 5.
History
The Need for Speed
In the early 2010s, startups began adopting "Growth Hacking"—rapid experimentation to drive growth. The bottleneck wasn't ideas; it was execution. Teams needed a way to sort through hundreds of ideas quickly.
Sean Ellis and ICE
Sean Ellis introduced ICE (Impact, Confidence, Ease) to help teams move fast. The simplicity was the feature: no complex spreadsheets, just three numbers multiplied. It allowed teams to say "No" to good ideas so they could say "Yes" to great ones.
WiderFunnel and PIE
Chris Goward of WiderFunnel developed the PIE framework specifically for CRO (Conversion Rate Optimization). It focuses on Potential (is the page broken?), Importance (is it high traffic?), and Ease. While similar, PIE is often preferred by optimization agencies.
Common Misconceptions
- Myth: "The score is the truth." Reality: The score is a discussion starter. It forces the team to debate why they think Impact is high.
- Myth: "Ease means lazy." Reality: Ease means ROI. Why do a hard project for the same return as an easy one?