A checkout error, searches with no results and an unpopular homepage banner can all enter the same backlog. Their counts do not tell you which deserves attention first. Prioritise the user task at risk, the strength of the evidence and the next useful action. The largest visible number may describe exposure rather than a problem you can solve.
Describe the failed task before proposing a redesign
“Improve the product page” names a location, not a problem. “Customers cannot check whether this item fits their device” identifies a task and missing information. Record who needs the answer, where they get stuck and what happens next. A compatibility question may require clearer product information. A missing delivery option may require a technical repair. A new layout is only one possible response.
Separate observations from explanations. An error log can show that a valid address receives no delivery option. A recording may show someone reopening delivery information. Neither establishes that every abandoned checkout was lost because of shipping. Keep the observed fact, suspected cause and proposed response in separate sentences. That makes disagreements easier to resolve before work starts.
For wider planning context, Web ostrý jako břitva by Jan Řezáč is a Czech-language book about website design. Its publisher presents it for both people commissioning websites and people creating them. The prioritisation method here focuses on one practical decision: which documented obstacle should receive the next available effort.
Compare reach, severity and evidence separately
Jakob Nielsen’s guidance on usability severity considers how often a problem occurs, how hard it is to overcome and whether it keeps recurring. Those questions help explain why a task blocker can deserve attention before a cosmetic concern. Severity still requires judgement. Record the conditions and the available workaround rather than attaching a dramatic label to every inconvenience.
