If “Pay”, “Save draft” and “Cancel” look identical, customers cannot tell which action matters most. This is more than an aesthetic flaw: someone might trigger an irreversible operation, lose data or abandon checkout. In Figma, evaluate visual hierarchy across tasks, states and realistic content—not just one idealised screen.
List the actions available on each screen and classify them as primary, secondary or potentially destructive, such as deleting a profile. Priority depends on context: confirming payment is usually primary during checkout, while saving is primary in a draft workflow. If every call to action receives a strong accent treatment, none stands out.
Review colour, position, labels, font weight, spacing and keyboard focus order. Colour alone is not an adequate description of consequences. Avoid a vague “Continue” label when the button will actually charge money; precise action names reduce ambiguity.
Define default, hover, focus, pressed, disabled, loading and error states where the product needs them. On mobile, check wrapping and separation of adjacent controls. A greyed-out disabled button is not an explanation of why an action is unavailable. Test legibility with longer localised labels.
Measure contrast against applicable WCAG criteria for the actual interface and text size. Drawing a contrast-compliant frame in Figma does not prove the implemented site is accessible; test the final behaviour too. A review of an existing website can locate ambiguous live controls, while targeted design changes can address reusable components.
Fictional account area B-83 contains two identical blue buttons: “Pay” starts a transaction, while “Save” stores form data without charging. Their equal size and placement create a risk of the wrong choice.
Positive test: the participant distinguishes payment from saving; keyboard focus is visible and errors lead back to a manageable state. Negative test: “Save” dominates the payment action or a destructive operation looks harmless. This requires a hierarchy change, not simply adjusting colour.
Provide the current screens, user scenarios, design system and frontend limitations. Request an action map, stateful components, mobile layouts and annotated development handoff rules. Visual identity supplies the brand foundation but cannot replace UX decisions. In an individual estimate request, describe the key user tasks and problematic screens; that defines scope more accurately than counting artboards.
Consider fictional screen F-83 where an operator sees four actions on an enquiry: “Save”, “Send to customer”, “Delete draft” and “Cancel”. When all controls share size, colour and position, a hurried operator may send before reviewing. Assess the frequency and consequences of each operation before styling emphasis. The most consequential action is not always the brightest; irreversible deletion chiefly needs an explicit label, distance from routine actions and confirmation.
Five hypothetical participants receive the instruction “save the draft without sending it to the customer”. Record chosen controls, errors and their explanations, without claiming that a sample of five yields statistically significant results. A positive result is clear differentiation between saving and sending without moderator prompts. If several people confuse deleting with cancelling, change labels, grouping and confirmation behaviour, then retest.
Provide developers with an annotated prototype specifying hover, focus, disabled and error states and responsive rules. The published identity commercial offer is not automatically an offer to conduct a UX audit. A quote for this work requires specific screens, scenarios and accessibility criteria.