Rank opportunities by both reach and frequency
Aliases: severity times frequency · opportunity score · reach and frequency
What it is
An opportunity list ranked only by how damaging a problem feels, or only by how often it shows up, will be distorted. Impact-by-frequency ranking requires each item to carry two checkable quantities: who is affected and how badly if the problem occurs, and how often it occurs in the intended use. Drop either dimension and the order tracks what researchers remember most vividly, not how the problem is distributed in actual use.
Why it happens
A severe incident just witnessed is easy to file as “the biggest problem.” Friction that costs a few seconds every day rarely becomes an incident report and sinks into the background. Impact captures scale of consequence: how many people, how critical the task, whether failure is reversible. Frequency captures exposure: how many people, how many times, at which step. The two must be crossed because high-frequency low-harm and low-frequency high-harm are different kinds of risk; they do not convert into one shared feeling of importance. Ranking by drama inflates rare disasters and suppresses chronic drain; ranking by counts does the reverse.
Studying it
Define the two dimensions separately and name the evidence source. Task failures, rework, safety events, tickets, logs, and spontaneous mentions in interviews are not on the same scale. For each opportunity, record in how many sessions it appeared and the rule used to grade consequence. With mainly qualitative material, state frequency as “recurs in which tasks and which groups,” not as a percentage without a denominator. Rank once by impact, once by frequency, then inspect whether the crossed top set moves. Large movement means the prior consensus was captured by a single dimension.
Where it stops holding
A product score treats the dimensions as commensurate and roughly independent. Events that are legally forbidden, cause bodily harm, or can do irreversible damage in a single occurrence must not be pushed down the list because they appeared once in the sample—frequency describes exposure here; it does not dilute severity. High-frequency operations in logs are not automatically problems users care about. Frequency is not comparable across studies that did not use the same task set.
Applying it
- Keep at least two columns on the opportunity table: consequence (who, which task, what failure means) and frequency (in which observations, whether it recurs).
- Grade the dimensions first, then cross them; do not assign a total from impression.
- List one-off but irreversible items separately so a product score cannot bury them.
- In review, drop one column and see whether the top five change; if they barely move, the other dimension is not actually in the judgment.