Framework · DesignOps & capacity · 2 min read
RICE scoring
A scoring formula from Intercom (Reach × Impact × Confidence ÷ Effort) that lets you compare backlog items on one scale and explain why one goes first.
RReach
How many people or events it affects in a set time period.
e.g. 1,200 users per quarterIImpact
How much it helps each person it reaches.
3 massive · 2 high · 1 medium · 0.5 low · 0.25 minimalCConfidence
How sure you are about your estimates. Penalises guesses.
100% high · 80% medium · 50% lowEEffort
Total work from everyone involved: design, research, build.
Person-weeks or person-months, used the same way for every itemRICE score = (Reach × Impact × Confidence) ÷ Effort
Use it when
- Several teams ask for UX work at the same time and you need to show why one request goes first.
- Your backlog has a mix of big redesigns and small fixes, and the big ones always win the discussion.
- A stakeholder pushes a favourite idea, and you want to talk about reach, impact and evidence instead of opinions.
How to use it
- Fix the units before you score. Agree the time period for Reach (for example, users per quarter) and the unit for Effort (person-weeks or person-months). Use the same units for every item. Output: a one-line scoring rule at the top of your sheet.
- Estimate Reach and Impact. Reach is how many people or events the item affects in that period; use real data where you have it. Impact uses the fixed scale: 3 massive, 2 high, 1 medium, 0.5 low, 0.25 minimal. Output: two numbers per item.
- Be honest about Confidence. Use 100% when you have data, 80% when you have some evidence, 50% when it is mostly a guess. Below 50% means you need research first, not a score. Output: a confidence value and a note on the evidence behind it.
- Estimate Effort with the people who do the work. Count all work: design, research, build and testing. Ask the designers and developers, not only the requester. Output: an effort estimate per item.
- Score, rank, then discuss. Calculate (Reach × Impact × Confidence) ÷ Effort and sort. Bring the ranking to the people who decide, and check what the score can't see: dependencies, deadlines, compliance. Output: a ranked list and an agreed priority. About 30 minutes for 10 items.
Worked example
Three fictional items from a UX backlog on an internal operations platform. Change any estimate and watch the scores and the ranking update.
| Rank | Backlog item | Reach per quarter | Impact | Confidence | Effort person-weeks | Score |
|---|---|---|---|---|---|---|
Watch out for
- Inconsistent units. If one item counts Reach per month and another per year, the ranking is meaningless. Write the units down and check them before you compare.
- Inflated Impact. When every item is "massive", the scale stops working. Ask for the evidence behind any 3, and compare items with each other, not in isolation.
- Treating the score as the answer. A deadline, a legal requirement or a dependency can put a lower score first. Say so openly, then record why.
- Scoring alone. Estimates you make by yourself look like your opinion in a number. Score with the requesters and the people doing the work.
Template
Copy this into a sheet and add one row per backlog item.
Reach period: per quarter Effort unit: person-weeks
Item | Reach | Impact | Confidence | Effort | Score | Evidence
------------------|-------|--------|------------|--------|-------|---------
| | | | | |
| | | | | |
| | | | | |
Impact: 3 massive · 2 high · 1 medium · 0.5 low · 0.25 minimal
Confidence: 100% high · 80% medium · 50% low
Score = (Reach × Impact × Confidence) ÷ Effort
Saved only on this device.