What it is and where it came from

A relative ranking tool, not an automatic decision

GrowthHackers' own documentation scores Impact, Confidence, and Ease from 1 to 10 and averages the three estimates. A multiplicative variant became widespread later, so this page names both readings instead of treating them as the same method.

Source: GrowthHackers: Prioritizing your ideas with ICE

How the score is calculated

The two readings, named precisely

Documented average · GrowthHackers + Hypotize

(Impact + Confidence + Ease) ÷ 3

Widely used multiplicative variant

Impact × Confidence × Ease

This is not a source-versus-Hypotize difference: both GrowthHackers and Hypotize use the arithmetic mean. The multiplication is a later, widely used variant.

What the team rates

What each letter must make explicit

Impact

Expected effect on the goal metric if the idea works.

Keep it grounded: Anchor the score to a named metric and audience.

Confidence

Strength of evidence behind the expected impact.

Keep it grounded: Use research, prior tests, or observed behavior; enthusiasm is not evidence.

Ease

How quickly and cheaply the idea can be implemented and tested.

Keep it grounded: Include design, engineering, operations, dependencies, and measurement work.

Free calculator · No account required

Try it on a real backlog item

Inputs stay in this browser. Nothing is sent to Hypotize by this calculator.

Shared 1–10 inputs

Impact

1 = minimal, 10 = very high

Confidence

1 = weak evidence, 10 = strong evidence

Ease

1 = difficult, 10 = easy

Documented average · GrowthHackers + Hypotize

Score, 1–10 scale

(Impact + Confidence + Ease) ÷ 3

Widely used multiplicative variant

Score, 1–1,000 scale

336

Impact × Confidence × Ease

Why they differ: The average stays on a 1–10 scale; multiplication can reach 1,000 and amplifies differences between the same three ratings.

Download this result:

Worked example

One experiment, two clearly named scores

A pricing-page social-proof test scored Impact 8, Confidence 6, and Ease 7 returns 7.00 with the documented average used by Hypotize. The widely used multiplicative variant returns 336 from the same ratings.

Decision rules

When to use ICE, and when not to

Use it when

You need to rank many growth ideas quickly.
Impact and confidence can be discussed on a shared 1–10 rubric.
Ease is a useful proxy for near-term execution cost.

Do not let it decide when

Reach varies radically between ideas and must be explicit.
Effort needs a real time or staffing unit.
A regulatory, strategic, or dependency constraint overrides score order.

Common input failures

Uncalibrated teams cluster every dimension around 7 or 8.
Ease can reward small work over strategically important work.
Confidence becomes optimism unless each score points to evidence.

In Hypotize

How Hypotize ranks this experiment

Hypotize averages each contributor's 1–10 Impact, Confidence, and Ease ratings, then keeps the resulting priority with the experiment. It does not use the multiplicative variant.

A score ranks the backlog. It does not run the test. In Hypotize the score is attached to a record with enforced rules: the experiment cannot skip a stage, and nobody, teammate or AI agent, can mark it done without a measured result and a written learning. A ranked list tells you what to try first. A system makes sure you find out what happened.

Choose the right framework

Pick it by the question, not the math

ICE and PIE produce the same kind of number: an average of three 1–10 ratings. They differ in the questions they make the team answer. RICE is the one that changes the ranking, because reach and effort enter the formula.

ICE

What you rate: Impact, Confidence, Ease, each 1 to 10
The score: Average, stays on a 1–10 scale
Choose it when: Ranking many growth ideas quickly on one shared rubric.

PIE

What you rate: Potential, Importance, Ease, each 1 to 10
The score: Average, same arithmetic as ICE
Choose it when: Ranking pages and funnel steps for conversion work. The math matches ICE; only the questions change, which is why PIE does not need its own calculator.

RICE

What you rate: Reach, Impact, Confidence, Effort
The score: (R × I × C) ÷ E, effort divides the score
Choose it when: When audience size and effort differ enough between ideas that they must change the ranking.

Framework	What you rate	The score	Choose it when
ICE	Impact, Confidence, Ease, each 1 to 10	Average, stays on a 1–10 scale	Ranking many growth ideas quickly on one shared rubric.
PIE	Potential, Importance, Ease, each 1 to 10	Average, same arithmetic as ICE	Ranking pages and funnel steps for conversion work. The math matches ICE; only the questions change, which is why PIE does not need its own calculator.
RICE	Reach, Impact, Confidence, Effort	(R × I × C) ÷ E, effort divides the score	When audience size and effort differ enough between ideas that they must change the ranking.

Keep the context

Read the experiment memory guide RICE guide Create your first experiment

Reviewed 29 August 2026. Source links identify the documented method; product behavior is derived from the current Hypotize framework configuration and calculator implementation.

ICE prioritization: guide and free calculator

A relative ranking tool, not an automatic decision

The two readings, named precisely

What each letter must make explicit

Impact

Confidence

Ease

Try it on a real backlog item

Shared 1–10 inputs

One experiment, two clearly named scores

When to use ICE, and when not to

Use it when

Do not let it decide when

Common input failures

How Hypotize ranks this experiment

Pick it by the question, not the math

ICE

PIE

RICE

Score the idea here. Then put it somewhere it cannot die.