| name | product-prioritization |
| description | Select what to build next and show the reasons — score the problems by how much users need help, score the solutions with a simple formula, give a source for each number, and make a roadmap with no false dates. Stage P4 of /product-workflow. Simple English, no product experience needed. Use when the user asks to put a backlog in order, make a roadmap, score ideas, or invokes /product-prioritization. |
Product Prioritization
Stage: P4 of /product-workflow. It comes after /product-discovery and
/product-validation. It gives data to /product-prd.
What you need to start: a list of user problems, or a list of possible things to
build. You also need the evidence from discovery.md, analysis.md, and
validation.md.
What you make: priorities.md in the initiative folder. The layout is in
templates/priorities.md. Open the template only when you write the document. Update the
same file. The roadmap stays correct after a change of plan because it has no dates.
Condition to go forward: the user approves the selection. You write the reasons in
the document, not only the order.
This skill uses four habits: select few things, use evidence, show the reasons, and
accept that each selection can fail.
Words: see ../product-workflow/GLOSSARY.md.
How to write: see ../product-workflow/STE.md. Use simple English in the document.
Context budget: apply the "Context and artifact budget" rules in
../product-workflow/SKILL.md; load more only when it changes the current decision.
Steps
-
Select which list you put in order: user problems, or things to build. Do not mix
the two lists in one table. You cannot compare them.
To put user problems in order: for each problem, give two scores from 1 to 10. The
first score is how important the problem is for the user. The second score is how
satisfied the user is with the options today. Then use this formula:
score = importance + (importance − satisfaction, or 0 if the result is negative)
A high score means two things: the user cares, and no option helps. There is space
for you. Do not put problems in order by difficulty. If you do, you remove the
largest opportunities because they look expensive before you design a solution.
To put things to build in order: use RICE.
score = Reach × Impact × Confidence ÷ Effort.
- Reach: how many users this touches in one quarter.
- Impact: how much it changes the result for each user. Use a number from 0.25
to 3.
- Confidence: how sure you are, as a percent.
- Effort: person-weeks or person-months.
For a fast first pass, remove Reach. Use Impact × Confidence ÷ Effort. The name of
this formula is ICE.
-
Give a source for each number. A source is an interview, a research group, a test
result, or usage data. If you guess a number, Confidence cannot be more than 50%. No
source, no score. Record the source date, user segment, confidence, and limitation
beside each score. This one rule keeps the scores honest.
-
Two extra methods are optional. Use them only if they give value:
- Find which features users expect and which features make users happy. This is
the Kano survey. Ask how the user feels if the feature exists. Then ask how the user
feels if the feature does not exist. The difference puts each feature in a group:
expected, more is better, or makes users happy. Use this method for solutions only.
It sets what users expect. It does not select the winner. Use it when the team
argues about how much to spend on delight.
- Find what a delay costs. Use this method to select between two items with the
same score, when one item has a time limit.
-
Make the roadmap with three columns: Now, Next, and Later. The columns show how
sure the team is. They do not show a month. Each item is an outcome or a problem. Do
not promise a feature.
- Now: the team tested it and made a commitment.
- Next: the direction is clear. The solution is open.
- Later: the team knows the problem. The team does not work on it now.
-
Record the selection: what you selected, what you did not select and why, and
which event starts a new review. Write the decision and the reasons in memory.md.
Cautions
- The score puts items in order. The score does not make the decision. A person
reads the evidence and makes the decision. Write the reasons. A table is not
responsible for a decision.
- Scores become false in two ways. First, numbers with no source. They make a guess
look like analysis. Second, effort controls the result. Then many small easy items get
a higher rank than the one problem that you proved. Look for both before you show the
document.
- The Kano survey is for solutions only. Do not use it for problems. Use the problem
formula for problems.
Checks before you finish
- The document says which list you put in order. Each number has a source. Each number
that you guessed has 50% confidence or less.
- The Now, Next, and Later table exists. It has outcomes, not features with dates.
- The user approved the selection. The reasons are in the document. The document names
the options that you did not select and what can change your decision.
