A pharmacy refill app holds two principles, 'refill in seconds' and 'never guess about medication', that clash over a new confirmation step; how should ranked principles resolve it?
answer
- decided before the argument starts
- X even over Y
- higher one wins the conflict
- lower one shapes the solution
- record the decision as an example
basics
~20 sPrinciples should carry an explicit ranking agreed in advance, so the higher one wins a conflict: medication safety keeps the confirmation step. The lower principle still applies, shaping the step to be as fast as safety allows.
solid answer
~40 sWhen two principles can conflict, the set should say which wins **before** the conflict happens, usually by ordering them or writing them as 'X even over Y'. Here the ranking would put 'never guess about medication' above 'refill in seconds', so the confirmation step stays. But the lower principle is not discarded: it drives how the step is built - prefilled details, the medication name and strength shown clearly, a single confirm action - so the flow stays as fast as safety allows. The decision is then recorded with the principle it cites, so the next similar debate starts from that precedent instead of from scratch. If the ranking keeps producing outcomes that users or research contradict, that is evidence to revisit the ranking, not to ignore it case by case.
go deeper
Recall that principles can conflict and that a ranking decided in advance says which one wins, so debates do not fall back to opinion.
Explain the forms a ranking takes - ordered list, even-over pairs, tiers - and why the lower principle still shapes how the winning one is satisfied.
Resolve a real conflict end to end: apply the ranking, design the solution under the lower principle, record the decision as precedent, and read the signals that the ranking needs revisiting.
Own the ranking as a product-strategy claim: decide what evidence can change it, get leadership to commit to it, and make overrides visible rather than silent.
## Why principles need a ranking A set of useful principles will conflict, because each one names a good thing the product favours. That is a sign they are real - platitudes rarely conflict because they rarely decide anything. The problem is what happens in the conflict. Without an agreed order, the debate falls back to whoever argues longest or holds the most senior title, and the principles end up decorating a decision made on other grounds. A **ranking** fixes that by deciding, ahead of time and away from any specific argument, which principle wins when two collide. ## Ways to write the ranking | Form | Example | Strength | Weakness | |---|---|---|---| | **Ordered list** | 1. Never guess about medication 2. Refill in seconds | Simple, easy to cite | Order can be forgotten if the list is long | | **'X even over Y'** | Medication certainty even over refill speed | Names the cost inside the principle | Works best for pairs, not whole sets | | **Tiers** | Safety principles above convenience principles | Handles larger sets | Conflicts within a tier stay open | The **'even over'** form is popular because it states both goods and the direction of the trade-off in one line; it reads as 'we value the second thing too, but when forced, the first wins'. ## Resolving the pharmacy conflict The team is debating a new confirmation step before a refill is submitted. One side cites 'refill in seconds'; the other cites 'never guess about medication'. With the ranking agreed, the resolution runs in two moves: 1. **The higher principle decides whether** - the confirmation step stays, because a wrong medication or strength is a safety risk and speed was ranked below it. 2. **The lower principle decides how** - the step is designed to cost as little time as possible: - prefill everything the pharmacy already knows; - show the drug name, strength and prescriber in one scannable block; - make confirming a single, obvious action; - skip nothing that safety requires, even when it slows things down. This is the part weaker answers miss: ranking does not mean the losing principle disappears. It still constrains the solution. ## Recording and reusing the decision A resolved conflict becomes an **example** attached to the principles - 'we kept confirmation and made it one tap'. Over time these examples act like case law: the next time a similar debate appears, for example about showing a dosage change banner, the team starts from the precedent. Practical habits: - note the decision, the principles cited and the outcome in the design review record; - link the example from the principle's page; - flag decisions where the ranking produced a result the team was uneasy about. A minimal decision record for the pharmacy case might read: the question (add a confirmation step before refill submission?), the principles cited (never guess about medication, ranked above refill in seconds), the outcome (keep the step, reduce it to one prefilled screen with one confirm action), and a follow-up (measure time to confirm and abandonment at that step in the next research round). Four lines are enough; what matters is that the reasoning survives the meeting and can be found by someone who was not in the room. ## When to revisit the ranking A ranking is a claim about what matters most to users and the business. It should change when evidence says it is wrong, not when a particular argument is inconvenient. Signals to revisit: - usability research shows patients abandoning refills at the confirmation step in large numbers; - the product's vision or its users change, for example expanding into non-prescription items where the risk profile differs; - the team keeps overriding the ranking informally - a sign the written order no longer reflects the real one. The revision is done deliberately and openly, with the reason recorded, so the ranking stays credible. ## Common mistakes - Declaring all principles **equal**, which pushes every conflict back to opinion. - Ranking principles **per argument**, which makes the ranking whatever the loudest person wanted. - Treating the lower principle as **cancelled** rather than as a constraint on how the higher one is met.
- What if a conflict involves two principles in the same tier, with no order between them?Treat it as a gap the set did not anticipate. Resolve the case with the team and the product owner, record the reasoning, and then decide whether the set needs an explicit order for that pair. Repeated same-tier conflicts mean the tiering is too coarse.
- How do you stop a senior stakeholder from overriding the ranking whenever they disagree?Get the ranking signed off by that level of leadership in advance, cite it explicitly in the review, and ask that any override be recorded with its reason. An override then becomes a visible proposal to change the ranking rather than a quiet exception, which is usually enough to keep it honest.
A ranking works like a league's tiebreaker rule: it is agreed before the season starts, so when two teams finish level nobody argues about which statistic decides it.
saying these in an interview costs you the question
- All design principles should be treated as equally important
- The losing principle in a conflict can simply be ignored
- Rank principles case by case when a conflict arises
- Conflicting principles mean the principle set is badly written
- A ranking once written should never be revisited