The RICE framework: prioritising experiments on reach, impact, confidence and effort

Your growth team never runs short of ideas. A new landing page, an email flow, a paid test, an onboarding tweak: the backlog grows faster than you could ever work through it. The real question is not what you can do, but what you should do first. The RICE framework gives you an honest, repeatable answer by scoring every experiment on four factors: reach, impact, confidence and effort. In this article you will read how to quantify all four, with extra emphasis on the two teams most often get wrong: reach and effort.

RICE originally comes from product management at Intercom, but it works at least as well for growth experiments. Its power lies in its simplicity: one number per idea, built from factors you can substantiate, instead of letting the loudest opinion in the meeting room win.

The formula in one line

You calculate the RICE score like this:

RICE = (Reach × Impact × Confidence) / Effort

The first three factors represent the expected return of an experiment. You divide those by the effort. What remains is return per unit of work: exactly the yardstick you want to prioritise on when time and people are scarce. An idea with a huge expected impact but months of work can end up lower than a small experiment that goes live tomorrow.

Let’s walk through the four factors one by one.

Reach: quantify who you really touch

Reach is how many people or accounts this experiment touches within a fixed period. That last part is crucial: pick one time unit and stick to it across your whole backlog, for example “per quarter”. Otherwise you are comparing apples to oranges.

The biggest mistake teams make: they guess. Someone says “a lot” and that implicitly becomes a high number. RICE forces you to go into your analytics and pull out a real figure. A few examples of how you substantiate reach:

• An experiment on your pricing page? Take the number of unique visitors to that page per quarter from your web analytics.
• A change in your onboarding email? Count how many new users receive that email per quarter.
• A paid campaign for a new segment? Estimate the reach based on comparable campaigns you ran earlier.

The difference between reach and your total audience matters. Reach is not how big your email list is, but how many people actually see or go through this specific experiment in the chosen period. A test that only runs for visitors already in your checkout has a smaller reach than a test on your homepage, even if your total traffic is the same.

By expressing reach in absolute numbers, ideas that feel “hot” but touch few people naturally drop back in the ranking. That is exactly the point.

Impact: how strongly does this change behaviour?

Impact is how much this experiment contributes to your goal per person you reach. Because you rarely know this exactly, RICE works with a fixed scale instead of false precision. A commonly used variant looks like this:

• 3 = massive effect
• 2 = high effect
• 1 = medium effect
• 0.5 = low effect
• 0.25 = minimal effect

You deliberately choose from these fixed values, so impact scores stay comparable between team members. Tie your impact estimate to one clear goal, for example conversion to qualified lead or activation of new users. Do not invent percentages you cannot substantiate. The scale is coarse on purpose, because you only know the real impact after the experiment.

Confidence: how solid is your evidence?

Confidence is your honesty filter. It corrects for the risk that your reach and impact estimates are hot air. Here too you use a fixed scale, expressed in percentages:

• 100% = firmly backed by data
• 80% = reasonably backed
• 50% = gut feel with some signals

Do you have hard data from earlier tests, user research or clear analytics? Then confidence can be high. Are you leaning mainly on an assumption? Then set confidence low and be honest about it. This factor makes sure a wild idea without any grounding does not land at the top of your list, however appealing it sounds. It keeps your discipline intact: if you cannot substantiate anything, that should be visible in the score.

Effort: quantify the full body of work

Effort is the denominator and it is structurally underestimated. It is the total amount of work across all disciplines: not only development, but also design, copy, data, campaign setup and analysis. You express effort in a fixed unit, usually person-months or person-weeks.

This is where things often go wrong. An idea feels small because it looks visually simple, but under the hood sits hidden work: a new tracking setup, alignment with sales, an approval round. By consistently counting effort in person-months, that hidden work surfaces. A few guidelines:

• Count the work of everyone who is needed, not only of the person writing the first line of code.
• Half a month of work from a developer plus half a month from a designer together make one person-month, not half.
• Round off deliberately coarsely. Just as with impact, you are not after pseudo-precision, but honest comparability between ideas.

Because you divide by effort, this factor has a big leverage effect on your ranking. Two experiments with the same expected return, but one costs twice as much work: that second one automatically drops to half the score. That is how RICE rewards ideas that are quick and cheap to test, and that is exactly what you want in an experimentation rhythm.

A worked example

Suppose you are torn between two experiments.

Experiment A is a redesign of your homepage hero. Reach: 8,000 visitors per quarter. Impact: 2 (high). Confidence: 80%. Effort: 2 person-months. Score = (8,000 × 2 × 0.8) / 2 = 6,400.

Experiment B is a new variant of your demo request form. Reach: 1,200 visitors per quarter. Impact: 3 (massive on this group). Confidence: 100%. Effort: 0.5 person-months. Score = (1,200 × 3 × 1.0) / 0.5 = 7,200.

On instinct, the homepage hero would probably win: bigger, more visible, more exciting. RICE shows that the form experiment scores higher, because it is almost free to test and you are certain about it. That is the kind of counterintuitive decision you use the framework for.

Where RICE fits in your growth system

A prioritisation method is not a standalone tactic. It is part of a bigger whole: the system that orchestrates SEO, CRO, content, paid and lead generation into one predictable growth engine. RICE is the filter that decides which experiments from that system come up first, so your scarce capacity goes to the highest expected return. Want to understand the broader picture? Then read our pillar on what growth marketing exactly is.

Two points of attention keep RICE honest. First: the score is a tool to structure the conversation, not a judge. If two ideas sit close together, context decides. Second: scores are not final. As soon as an experiment produces data, you update your confidence and impact assumptions for related ideas. That way your backlog gets a little smarter every quarter.

Want to dig into the experimentation rhythm itself? Our guide to the growth experiment process shows how to also run the prioritised experiments reliably, and in the ICE scoring model you can read about a lighter variant that only scores on impact, confidence and effort.

Start small and build the rhythm

You do not have to implement RICE perfectly to get value out of it. Score five ideas from your backlog this week, force yourself to pull reach from your analytics and count effort honestly in person-months. You will notice that the ranking deviates from your gut, and that is exactly the gain: decisions based on expected return per unit of work, not on who argues loudest.

Want to anchor this in a real growth engine where prioritisation, experiments and channels work together? As a growth marketing agency, we build that rhythm together with your team, from backlog to result. Get in touch and we will look together at which experiments belong at the top for you.