What is ICE Score?
ICE Score is a prioritization framework used by CRO and product teams to rank experiment ideas. Each idea is scored on three dimensions — Impact, Confidence, and Ease — and the scores are averaged or multiplied to produce a single priority number.
How to calculate ICE Score
ICE Score = Impact x Confidence x Ease
Each factor is rated on a scale of 1-10:
- Impact (I) — How significant will this change be if it works? Consider the potential effect on conversion rate, revenue per visitor, or another key metric.
- Confidence (C) — How certain are you that this change will produce the expected impact? Confidence should be based on supporting data — analytics, user research, session recordings — not gut feeling.
- Ease (E) — How simple is this to implement and test? Consider design, development, and QA effort. A CSS change scores high; a platform migration scores low.
For example, an idea with Impact 8, Confidence 6, and Ease 9 scores 432 (8 x 6 x 9). Compare this to an idea with Impact 9, Confidence 3, and Ease 4, which scores 108 — despite having higher perceived impact, it ranks much lower because of low confidence and high effort.
Why it matters for eCommerce and SaaS
Testing bandwidth is limited. Most eCommerce sites can run 2-4 meaningful A/B tests per month, and SaaS companies with lower traffic may manage even fewer. A prioritization framework ensures those limited test slots go to the ideas most likely to produce meaningful results.
Without a scoring system, prioritization is driven by HiPPOs (Highest Paid Person’s Opinions), recency bias, or whatever feels most urgent. ICE provides an objective way to compare ideas across different parts of the funnel and different team members.
Limitations of ICE
While ICE is simple and widely adopted, it has known weaknesses:
- Subjectivity in scoring — Impact and Confidence are often scored based on intuition rather than data, which undermines the objectivity the framework is supposed to provide.
- No revenue anchor — ICE does not tie Impact to a specific dollar estimate, making it hard to compare a checkout optimization against a landing page test.
- Actionability gap — ICE does not account for whether an idea is well-defined enough to act on immediately. A high-scoring but vague idea wastes planning time.
These limitations are why some teams, including acceleroi, have developed alternatives like the AXR Score framework.
How acceleroi approaches it
At acceleroi, we use ICE as a starting point when onboarding new clients who are familiar with it, then transition to our AXR framework which addresses ICE’s limitations by anchoring scores to revenue estimates and filtering for actionability. Both frameworks serve the same purpose — ensuring every test slot is spent on the highest-value opportunity.
Related resources
- Get a free CRO audit to get a prioritized list of your optimization opportunities
- Read our blog for guides on building a testing roadmap