Quantitative Methods

a worked method for estimating a realistic audit timeline

Quantitative research measures behaviour and opinions across large samples. Surveys collect structured data from many respondents, revealing patterns and frequencies. Analytics track user actions on existing products–page views, task completion rates, drop-off points–in aggregate. Large sample sizes allow statistical confidence and spotting edge cases that matter. The trade-off is that numbers alone do not explain why users behave as they do. A high bounce rate tells you users left; it does not tell you why they left or what would make them stay.

Qualitative Methods

ux audit scope

Qualitative research explores the reasoning behind behaviour. Interviews with small groups reveal motivations, frustrations, and mental models that quantitative data cannot capture. Contextual observation–watching people use products in their natural environment–exposes unexpected barriers. Open-ended questions yield rich narratives that uncover motivations and pain points. The limitation is that findings from small samples cannot be generalised to entire populations. Five users cannot represent all users, but their detailed feedback often highlights genuine problems worth fixing.

Mixed Approaches

Strong research programmes combine both methods. Begin with qualitative research to identify what matters, then quantify how common the problem is. Use analytics to spot unexpected patterns, then conduct interviews to understand why those patterns exist. This layering reduces bias and builds confidence in conclusions. Diary studies allow users to record behaviour over days or weeks, capturing natural patterns without researcher presence. Usability testing, covered separately elsewhere, bridges the two by observing task performance (quantitative) and asking follow-up questions (qualitative).

quick design audit

Research works best when framed around specific questions rather than open-ended exploration. Ask what problem you are trying to solve before choosing a method. Time constraints matter–a five-day research sprint requires different methods than a six-month programme. Budget limitations are real; acknowledge them and prioritise high-impact insights. Document findings clearly so recommendations connect directly to observed evidence. Research that goes unread or misunderstood wastes resources. Communicate results in forms others can act on.

an inside look at what a design audit report should contain

Inventory and Gap Analysis

Audit begins by inventorying all products and identifying their components–buttons, forms, navigation patterns, modals, cards. Compare inventory against system documentation. Do all products use system buttons or have teams created custom alternatives? Gap analysis reveals where products diverge from system. Inventory is difficult for large organisations but essential. Without it, system team works blind, assuming coverage that does not exist. Visual audits help–screenshot views of components across products reveal subtle inconsistencies. A button using system component but with a font colour tweaked for one product is a gap. These accumulate and undermine system value. Teams skip system components when they are incomplete or do not perfectly fit context. Identifying why teams work around system guides design decisions.

Documentation and Maintenance Health

Review documentation clarity and accuracy. Is guidance clear enough that a new engineer can use a component without support? Are visual examples up-to-date? Do examples reflect real usage or simplified ideal cases? Outdated documentation creates frustration and drives teams toward custom solutions. Audit component last-update dates. When was this button last reviewed? When was this form pattern last used in a product? Components that have not been touched in years may be holding legacy debt. Maintenance practices matter–do changes require system team review or can anyone update components? Is there a process for deprecating old patterns? Without process, systems accumulate dead weight.

a realistic look at what a screen inventory actually contains

Planning System Evolution

https://dnsk.work/blog/well-just-do-a-quick-design-audit-famous-last-words

Audit findings guide system roadmap. High-priority fixes are gaps teams consistently work around and outdated components causing confusion. Lower-priority work includes documenting rarely-used components and refining edge cases. Resource constraints are real; audits help system teams focus effort on maximum impact. Consider audit as recurring practice, not one-time exercise. Conduct lightweight audits annually to track progress and spot emerging gaps. Major audits every few years (when system undergoes significant change) validate whether system is meeting its purpose. Systems that do not evolve become irrelevant. Teams abandon them and build workarounds. Proactive auditing keeps systems healthy, reduces duplication, and supports product coherence at scale.

Common Frameworks

The “I like, I wish, I wonder” framework invites three types of feedback. I like–what works well–builds confidence. I wish–what could improve–surfaces problems. I wonder–what if questions–spark exploration. This framework balances positive and constructive feedback. Asking “I wonder if using bigger type would make this clearer” is gentler and more generative than “the type is too small.” Audience-centred critique focuses on specific user personas and contexts. Rather than generic feedback, critique references the actual user: “Sarah is a busy parent; will she understand this within two seconds?” Intent-based critique asks the designer to state their goal first. “I am trying to reduce checkout friction” focuses discussion on whether the design achieves that goal rather than debating visual choices. Design debt reviews focus on feasibility and maintenance burden. “How will this pattern scale if we add more items to the list?”

visit website

Running Effective Critiques

a structured method for comparative design audits

Prepare before the meeting–designers present work and communicate the goal and constraints. Allocate time proportional to design maturity: five minutes for early sketches, thirty minutes for high-fidelity work. Ensure diverse perspectives–different roles (engineering, product, support) catch different issues. Establish psychological safety; harsh or personal criticism kills engagement. Distinguish between “this does not meet the goal” (valid critique) and “this is not my taste” (preference). Document feedback rather than discussing everything in real-time. Synthesis after the meeting is easier than trying to act on scattered verbal feedback. The designer should leave understanding what feedback is actionable and what was merely opinion.

Integrating Feedback

Feedback does not mean doing everything everyone suggests. The designer synthesises input and makes decisions. Contradictory feedback is common; the designer interprets it in light of goals and data. If critique suggests two opposing directions, testing with users breaks the tie rather than arguing. Over-relying on critique can lead to design by committee–everything feels like a compromise. Critique is input, not decision-making. Regular critiques throughout a project maintain alignment. One critique at the end when major changes are impractical frustrates everyone. Critiques at sketching, wireframing, and high-fidelity stages let teams course-correct progressively. Documenting critiques and resolutions helps future teams learn. Why was that approach chosen over alternatives? Critique records preserve that thinking.

Planning and Recruiting

Define tasks that reflect real usage goals. Instead of asking participants to “explore the website,” ask them to “find the return policy and check if they can return items purchased two months ago.” Recruit participants who represent your actual audience–demographics, technical skill, domain knowledge all matter. Five to eight participants are usually sufficient for qualitative insight; larger numbers are needed if results will be statistically analysed. Screening questions help ensure recruits fit the target profile. For a business tool, recruit operations staff who use spreadsheets daily rather than general consumers, since their mental models differ substantially. Schedule one-hour sessions; longer durations exhaust both participant and facilitator. Offer modest incentives–£20 to £40–to respect participant time.

Running Sessions

what clients actually get from a design audit

Brief participants on their role: they are helping evaluate the design, not being evaluated. Ask them to think aloud–narrate their thoughts, reactions, confusion–but avoid leading questions that suggest correct paths. Silence after a participant gets stuck is uncomfortable but necessary; jumping in with help prevents you learning where design clarity fails. If a participant abandons a task in frustration, note the exact step where confidence dropped rather than only the final outcome, since that moment of hesitation often reveals the real flaw. Record sessions (video, screen, audio) so you can review later. In-person testing captures body language and environmental factors; remote testing is faster and reaches geographically dispersed participants. Both reveal usability problems equally well.

Analysis and Documentation

Review recordings and note where participants succeeded, struggled, or diverged from expected behaviour. Group similar problems and prioritise by frequency and severity. A participant unable to complete a core task signals a critical flaw; minor hesitations about wording are lower priority. Document findings with specific examples: rather than “navigation is confusing,” write “three of five participants looked for the settings menu in the top-right corner but it was in a hamburger icon on mobile.” Tagging each observation with the task and participant number makes it far easier to revisit raw footage later when stakeholders ask for supporting evidence. Recommendations follow from evidence. Watching participants teaches you faster than reading a report. When possible, show video clips to stakeholders–seeing real confusion changes minds better than summarised data.