Last Updated: October 1, 2025 | Originally Published: November 24, 2022

As entrepreneurs, we rarely experience the clarity that makes our next steps obvious.

Most days we're either chasing decisions already made or delaying them indefinitely, unsure which path is right. But the quality of your decision-making framework is one of the main factors determining whether your startup thrives.

So when you've got 10+ ideas competing for attention but only resources to deliver three of them, how do you prioritise effectively?

In this guide, we'll show you exactly how to use ICE scoring – the same prioritisation framework that helped companies like Dropbox achieve breakthrough growth. You'll learn the three-factor formula and run your first ICE workshop by the end of this article. ⬇️

What is ICE scoring?

ICE scoring is a lightweight prioritisation framework that helps you sort ideas based on three factors: Impact, Confidence, and Ease.

It was created by Sean Ellis (yes, the person who coined "growth hacking") when he needed a simple way to evaluate growth experiments at LogMeIn and Dropbox. After seeing it work, he shared it with the growth community, and it quickly became the go-to scoring model for fast-moving teams.

Perfect for when you have lots of competing ideas, ICE is a prioritisation method we use with many of our clients. It's also brilliant for improving team communication, and you can run an ICE workshop in 30 minutes.

Why startups need this

Teams have no shortage of ideas. But everyone thinks their idea is good, which inevitably leads to analysis paralysis and no measurable improvement in results.

Fast-growing companies use prioritisation frameworks to surface the handful of ideas, from the tens or hundreds suggested, that will actually drive growth. The goal is finding the 10% of ideas that drive 90% of growth – essential for early-stage startups working with limited resources.

“ICE scoring forces you to have an honest conversation with your team about what you actually know versus what you're assuming. The magic is in the debates the numbers provoke. When a founder argues their pet feature deserves a confidence score of 8 but can only cite 'gut feeling' as evidence, that's the moment real learning begins.

At Lumi, we use ICE with our clients because it's the fastest way to align a team around reality. You can run a full scoring session in 30 minutes and walk away with crystal-clear priorities. Compare that to endless Slack debates or 'let's circle back' meetings that achieve nothing. 

The framework itself is deliberately simple – three factors, basic multiplication – but that simplicity is precisely why it works. Sophisticated founders don't need complicated frameworks; they need tools that help them move fast and stay honest about what they're building.”

— Milosz Falinski, Founder of Lumi Studio
Milosz has worked on multiple startups that have been acquired, including ventures in fintech, SaaS, and digital products. He specialises in helping early-stage founders make better product decisions with limited resources.

How it works

Let's say you've generated 10+ ideas for what to do next, but only have resources to deliver three of them. You're wondering which three to choose – ICE scoring helps you decide.

Step 1: List all competing ideas

Gather in one place (whiteboard, Miro, Trello – whatever works) all the ideas you're considering implementing. This could include feature requests, marketing campaigns, A/B test ideas, or MVP features.

Step 2: Score each idea on a scale of 1-10

Impact: How much will this move the needle on your key metric?

This is about potential value. Will this significantly improve your KPIs, increase revenue, boost engagement, or solve a critical problem? The higher the impact, the higher the score.

Confidence: How sure are you this will actually work?

Be careful not to score based on how they feel about an idea rather than the evidence supporting it.

Instead of asking "How confident are you?", ask "How much evidence do you have?" This slight tweak makes all the difference between scoring based on personal preferences versus being honest about what you actually know.

Confidence scale:

• 1-2: Pure gut feeling, no supporting data
• 3-4: Team opinions or stakeholder feedback
• 5-6: Customer feedback or feature requests
• 7-8: Market research or competitive analysis
• 9-10: Validated by user testing or launch data

Ease: How simple is implementation?

This measures how easy it is to build and launch – essentially the inverse of effort. Consider time required, resources needed, technical complexity, and dependencies. Think about your resource constraints realistically.

💡 It's brilliant to do this with your growth team, discussing each score and creating consensus on each component. It's an easy way to let your team contribute to decision-making and align on priorities – crucial for team alignment.

Step 3: Calculate the scores

Multiply all three scores together (not add – more on why this matters later).

ICE Score = Impact × Confidence × Ease

Example:

• Idea A: Impact 9, Confidence 5, Ease 7 = 315
• Idea B: Impact 7, Confidence 8, Ease 9 = 504
• Idea C: Impact 10, Confidence 3, Ease 6 = 180

Idea B wins despite lower impact because the higher confidence and ease reflect a more realistic shot at success.

Step 4: Prioritise those ideas

Look at the top-scoring ideas. Are they the ones you expected? Are they the ones you believe will have the highest impact on your business?

Will you pick the top 3 and move forward, or will you challenge the scoring?

How to run an ICE workshop

Let's get practical. Here's exactly how to run your first ICE session ⬇️

Pre-workshop prep

Before gathering your team:

✅ Define your primary goal clearly – What single metric are you trying to improve? (e.g., "Increase trial-to-paid conversion" or "Reduce churn")

✅ Gather your ideas – Collect 5-15 competing ideas in one place (more than 15 and you'll run out of time)

✅ Invite the right people – Include those who understand business impact AND technical feasibility (usually 3-5 people max)

✅ Prepare your template – Set up a simple spreadsheet or use a tool like Miro/Notion

💡 Keep the group small. More than 5 people and the workshop becomes a debate society.

Workshop agenda

Frame the session

• Explain the ICE framework briefly
• Remind everyone of the primary goal
• Set the rule: Score based on evidence, not opinions

Score Impact

• Go through each idea and discuss: "If this succeeds, how much will it move our key metric?"
• Have each person score independently first (prevents groupthink)
• Discuss any scores that differ by more than 2 points
• Agree on a final Impact score

Score Confidence

• For each idea: "How much evidence do we have that this will work?"
• Use the evidence-based scale (see detailed guidelines below)
• Again: independent scoring first, then discussion

Score Ease

• For each idea: "How simple is this to implement?"
• Consider time, resources, technical complexity, dependencies
• Get input from technical team members for accuracy

Calculate & discuss

• Multiply the three scores (Impact × Confidence × Ease)
• Sort by highest to lowest
• Discuss: Are the top 3 surprising? Do they align with strategic priorities?
• Decide on next steps

Post-workshop

✅ Document everything – Record scores and rationale in a shared location

✅ Assign owners – Who will execute on the top ideas?

✅ Set review date – When will you reconvene to assess results?

✅ Share with stakeholders – Brief anyone who needs to know

Why ICE works (and when it doesn't)

What makes ICE powerful:

Speed: Unlike complex prioritisation methods, ICE is fast. Perfect for early-stage startups or innovative workplaces where you need to keep momentum and avoid decision paralysis.

Simplicity: Only three factors to consider. Everyone on your team, technical or not, can grasp it immediately – no complicated scoring system to learn.

Flexibility: Works for A/B tests, feature releases, bug fixes, marketing campaigns. Whatever you're prioritising.

Data-informed thinking: Encourages teams to justify scores with evidence rather than opinions, cutting through subjective debates and eliminating analysis paralysis.

When ICE works best:

✅ Early-stage startups with limited resources
✅ Growth experiments where speed matters
✅ Marketing campaign prioritisation
✅ MVP feature prioritisation
✅ A/B testing backlogs
✅ Small team decisions (1-10 people)

When to consider alternatives:

ICE isn't meant for long-term strategic planning. If you're planning major initiatives across multiple teams or need to account for different audience sizes, consider RICE (which adds "Reach") or other frameworks like OKRs or Growth Mapping.

ICE vs other frameworks

RICE (Reach, Impact, Confidence, Effort)

Adds "Reach" to account for how many users an initiative affects. More comprehensive but also more complex.

➡️ When to use RICE: When working across multiple teams or when you need to differentiate between features that affect many users slightly vs. few users significantly.

PIE (Potential, Importance, Ease)

Similar to ICE but emphasises "Potential" (probability of success) more explicitly.

➡️ When to use PIE: Best for conversion rate optimisation and A/B testing prioritisation.

MoSCoW (Must Have, Should Have, Could Have, Won't Have)

Categorises ideas into buckets rather than scoring them numerically.

➡️ When to use MoSCoW: Projects with fixed deadlines or when involving non-technical stakeholders.

👀 For early-stage startups, ICE hits the sweet spot between simplicity and usefulness. As you mature and have more data, consider RICE for more nuanced decisions. Don't let perfect be the enemy of good.

Provoking your instincts

Something brilliant about ICE is that it reveals your biases.

You've probably had a 'secret favourite' amongst your ideas. The scoring process will expose this – you'll either be relieved it scores high or frustrated when it ranks low.

Instead of dismissing these feelings (or the method), get curious. What's really behind your preference? Is it based on solid evidence or emotional attachment?

ICE won't give you simple answers to complex challenges, but it provides an alternative, impartial point of view. It forces you to make your trade-offs explicit rather than keeping them hidden in your head.

Sometimes you may need to prioritise a lower-scoring initiative for strategic reasons. That's perfectly fine. The value of ICE lies in making these decisions explicit and discussable within your team.

The bottom line

• ICE scoring transforms subjective debates into structured decisions using Impact, Confidence, and Ease
• Evidence-based confidence is crucial – score based on data, not feelings
• Use ICE for relative prioritisation, not absolute measurements
• Companies like Dropbox and Airbnb used ICE to drive breakthrough growth
• Works best for early-stage startups and growth experiments
• Scores should guide decisions, not make them – apply judgment and strategic context
• Run ICE workshops quarterly to maintain momentum and alignment

🤝 Need help implementing ICE scoring or building a data-driven decision-making culture in your startup? Book a free 30-minute chat with our team. We've helped 50+ startups move from analysis paralysis to confident execution.

FAQs

How often should we update our ICE scores?
Review quarterly, or whenever significant new information emerges. As you gather evidence through experiments and customer feedback, your Confidence scores should improve. Document your scores so they act as reference points for future prioritisation and help new team members understand your decision-making history.

Can different team members score the same idea?
Absolutely. According to multiple studies, the average guess of a group is astonishingly accurate. Aggregate their scores and calculate the average – perfect for growth teams working collaboratively.

What if our highest-scoring idea conflicts with our strategy?
Sometimes you'll need to prioritise a lower-scoring initiative for strategic reasons (entering a new market, retaining a key client, regulatory requirements). That's fine. ICE makes these trade-offs explicit, not automatic. When you override the scores, document why – this builds institutional knowledge about your strategic priorities.

Should we use a 1-10 scale or 1-5? 

Most teams use 1-10 for better granularity, but 1-5 works if your team prefers simplicity. Some teams even use 1-3 for maximum speed. The key is consistency within your team – everyone needs to understand what each score level means.

How is ICE different from a simple effort vs. impact matrix?
ICE adds Confidence as a critical third dimension, letting you account for uncertainty in your estimates. It's an evolution of classic impact/effort analysis.

Can ICE work for B2B and B2C products? Yes, but adapt your evidence sources. B2B might use sales call feedback and customer advisory boards for Confidence scores. B2C might use analytics data and user testing. The framework is the same; only the validation methods change.

Should we multiply or add the scores? Always multiply (Impact × Confidence × Ease). Multiplying creates bigger differences between options and better reflects reality: a project with one very low score (like Ease = 1) probably shouldn't rank high, even if Impact is 10. Adding would hide this problem.

About the author: Milosz Falinski is Founder of Lumi Design, a design strategy expert and startup veteran. Businesses Milosz has worked on tend to be acquired.