The ICE framework is the original scoring system for growth marketing teams. Created by Sean Ellis — the person who coined the term “growth hacking” and author of Hacking Growth — ICE was designed to help teams at companies like LogMeIn and Dropbox decide which experiments to run first. It remains the most widely used prioritisation framework in growth teams today.
Table of contents
Open Table of contents
- What does ICE stand for?
- How to calculate an ICE score
- Worked example: scoring real marketing experiments
- When to use the ICE framework
- Strengths of the ICE framework
- Limitations of the ICE framework
- Tips for better ICE scoring
- ICE Framework vs other prioritisation frameworks
- Getting started with ICE
- Frequently asked questions
What does ICE stand for?
ICE stands for Impact, Confidence, and Ease — three factors used to score and rank experiment ideas on a scale of 1 to 10.
- Impact — How much will this move your target metric?
- Confidence — How sure are you it will work?
- Ease — How quickly can you ship it?
Each factor shapes the final ICE score differently depending on your team’s stage and velocity. Read on to see how to weight and calculate them.
How to calculate an ICE score
The ICE score is calculated by averaging the three individual scores:
ICE Score = (Impact + Confidence + Ease) / 3
This gives a final score between 1 and 10. The key is what you do before you plug the numbers in — how you define and calibrate each factor determines whether your rankings are meaningful or misleading.
ICE scoring factors explained
Use the table below as a reference when calibrating scores with your team. Agreeing definitions before your first session prevents anchoring and makes comparisons fair across different ideas.
| Factor | What it measures | Low score (1–3) | Mid score (4–6) | High score (7–10) |
|---|---|---|---|---|
| Impact | How much this will move your target metric | Negligible lift expected | Moderate, measurable lift | Significant, data-backed lift |
| Confidence | How certain you are it will work | No data, pure gut feel | Some data or precedent | Strong data, prior wins |
| Ease | How quickly you can ship it | Weeks of engineering, complex | Medium effort, some dependencies | Can ship in days, minimal dependencies |
Worked example: scoring real marketing experiments
Most teams score ICE wrong in the same way: they give Impact a 9 because an idea feels big, not because they have data. Here is how a real growth team would score four common experiment ideas — and why the highest-impact idea doesn’t always win.
| Experiment | Impact | Confidence | Ease | ICE Score |
|---|---|---|---|---|
| Rewrite homepage headline based on customer interviews | 8 | 7 | 9 | 8.0 |
| Launch referral programme with double-sided incentive | 9 | 5 | 4 | 6.0 |
| Add exit-intent popup offering lead magnet | 6 | 6 | 8 | 6.7 |
| Rebuild onboarding flow with personalised steps | 9 | 6 | 3 | 6.0 |
The homepage headline rewrite scores highest — not because it has the biggest potential impact, but because the team has high confidence (backed by customer interview data) and it is easy to execute. The referral programme and onboarding rebuild have higher potential impact but score lower because they require more effort and carry less certainty.
ICE stops teams from always chasing the biggest ideas and surfaces the experiments most likely to deliver results quickly.
When to use the ICE framework
ICE is not the right tool for every team. It works best in three specific situations — and can actively mislead you in others.
- Early-stage growth teams that need a simple, fast way to start prioritising without methodology overhead.
- High-velocity experimentation where you run multiple experiments per week and need quick decisions.
- Cross-functional teams that need a shared scoring language without bogging down in debate.
For teams running experiments at scale or needing to account for audience size, a more structured framework may be a better fit. See the comparison table below.
Strengths of the ICE framework
ICE has survived 15+ years in growth teams for three reasons — and one of them is often underestimated.
- Speed. Three factors, scored 1 to 10. You can score an idea in under a minute.
- Simplicity. No formula to memorise, no training required.
- Forces structured thinking. Even a quick ICE score makes you consider an idea from three angles before committing resources.
The third point matters more than it sounds. Read on for why.
Limitations of the ICE framework
ICE has four well-documented weaknesses. The first is widely known; the fourth catches most teams off guard.
- Subjectivity. Scores are inherently personal — two people can score the same idea very differently, especially on Impact and Confidence.
- No reach factor. ICE ignores how many people an experiment will affect. A high-impact change on a low-traffic page scores the same as one on your highest-traffic page.
- Ease is ambiguous. Does Ease mean time, cost, or complexity? Agree on a definition before your first scoring session.
- Anchoring bias. The first score shared in a group session pulls everyone else toward it.
The RICE framework was built specifically to address the reach limitation — see the comparison below.
Tips for better ICE scoring
-
Define your scoring scale. Before your first scoring session, agree as a team on what a 1 versus a 10 means for each factor. Write it down and reference it in future sessions.
-
Score independently first. Have each team member score the idea on their own before discussing. This prevents anchoring and surfaces genuine disagreements.
-
Discuss outliers. When scores diverge sharply — one person gives Impact a 3 and another gives it an 8 — the team has different assumptions. Talk through these before averaging.
-
Re-score regularly. An idea scored three months ago may need re-scoring as market conditions, team capacity, or priorities shift.
-
Use ICE for ranking, not precision. The scores produce a relative ranking, not exact predictions. Do not agonise over the difference between a 6 and a 7.
ICE Framework vs other prioritisation frameworks
ICE is the simplest prioritisation framework, but it is not always the best fit. For a detailed comparison of all the major frameworks available to growth teams, see our full guide to prioritisation frameworks.
| Framework | Factors | Best for | Complexity |
|---|---|---|---|
| ICE | Impact, Confidence, Ease | Fast, general-purpose experiment prioritisation | Low |
| RICE | Reach, Impact, Confidence, Effort | Teams comparing ideas across different audience sizes | Low–Medium |
| PIE | Potential, Importance, Ease | CRO teams prioritising page-level A/B tests | Low |
| PXL | 10 binary questions (traffic, visibility, data sources) | Reducing subjectivity in CRO prioritisation | Medium |
| DRICE | Detailed RICE + hypothesis, impact estimate, build estimate | High-stakes ideas requiring 30-min deep evaluation | High |
ICE vs RICE
The RICE framework was created by Sean McBride at Intercom specifically to fix ICE’s biggest blind spot: it ignores how many people an experiment will reach. RICE multiplies Reach, Impact, and Confidence, then divides by Effort — so a change to a high-traffic page will score materially higher than the same change on a low-traffic page. Use RICE when your backlog contains ideas that target vastly different audience sizes.
ICE vs PIE
The PIE framework was developed by Chris Goward at WiderFunnel for conversion rate optimisation. PIE replaces ICE’s Confidence factor with Importance (how valuable is the traffic to this page?), making it better suited to CRO teams deciding which pages to test next. If your experiment backlog is mostly page-level A/B tests, PIE may be a more natural fit.
ICE vs PXL
The PXL framework from Peep Laja at CXL replaces the 1–10 subjective scale with 10 binary (yes/no) questions based on evidence sources — whether the change is above the fold, whether it is supported by heatmap data, whether it targets high-traffic pages, and so on. PXL is more time-consuming than ICE but significantly reduces scoring subjectivity.
Getting started with ICE
Pick your top 10 experiment ideas, score each one using ICE in a team session, and run the highest-scoring experiment that week. Do not overthink the methodology — the value of ICE is in building the habit of structured prioritisation, not in achieving perfect scores.
Growth Method is the only work management platform built specifically for growth teams, with ICE scoring built in. Book a call to learn more.
If you are looking for a framework to structure your overall marketing strategy — rather than prioritise individual experiments — the RACE framework is a strong complement to ICE. It maps the full customer journey across Reach, Act, Convert, and Engage stages, giving you the strategic context to decide which stage to focus your experiments on first.
Frequently asked questions
What does ICE stand for?
ICE stands for Impact, Confidence, and Ease — three factors scored 1–10 and averaged to produce a prioritisation score for each experiment idea.
How do you calculate an ICE score?
Add your Impact, Confidence, and Ease scores together and divide by three: ICE Score = (Impact + Confidence + Ease) / 3. A score of 7 or above is generally considered high priority. Define what each score means for your team before your first session to prevent anchoring.
What is the difference between ICE and RICE?
RICE adds a Reach factor to account for how many people an experiment affects, making it more suitable for teams comparing ideas across different audience sizes. ICE is faster and simpler, and works best when most ideas target a similar audience size. See the full RICE framework guide for worked examples.
When should you use the ICE framework?
ICE works best for early-stage growth teams, high-velocity experimentation programmes running multiple tests per week, and cross-functional teams that need a shared scoring language without complexity overhead. If your team is scoring more than 20 ideas per sprint or working across very different audience segments, consider upgrading to RICE or DRICE.
What is the difference between ICE and PIE?
PIE (Potential, Importance, Ease) was developed specifically for CRO and A/B testing on existing pages, replacing ICE’s Confidence factor with Importance — a measure of how much traffic and revenue the page already drives. ICE is more general-purpose and suited to any growth experiment backlog, not just page-level conversion tests. See our PIE framework guide for details.