Heuristic evaluations: Finding usability issues before users do
How to find what you might be overlooking and uncover usability gaps.
A lot of digital platforms look fine on the surface. The branding is polished, the content reads well, the homepage does what it’s supposed to do. But what users actually experience when they try to complete a task, like finding a form or navigating processes, is often a very different picture. Confusing navigation, dead-end links, inconsistent patterns, tasks that take seven clicks when they should take two. A heuristic evaluation is how you uncover that gap. It’s a structured, expert-led review of a website or digital product against established usability principles, and it’s one of the most efficient ways to understand what’s actually working and what isn’t.
Done well, a heuristic evaluation helps you find common issues before you start research, turn opinions into evidence, and create a shared vision of what actually needs fixing. Keep in mind that heuristic evaluations should be completed prior to usability testing, not as a substitute, it helps to flush out the obvious issues before involving users.
It’s not glamorous work, but it’s the kind of honest, grounded assessment that saves teams from investing in the wrong things. And yes, you can use AI to do them. But we’ll get to why that’s not the best idea later.
“Good usability is like oxygen – you don’t notice it until it’s missing.”
– Jakob Nielsen
What is a heuristic evaluation?
A heuristic evaluation is a method for identifying usability issues within an interface. Evaluators review the product against an established set of usability principles called “heuristics”, to reveal where the design helps or hinders users. It can be applied with any relevant heuristic set and is particularly effective at surfacing obvious, high-impact problems early.
A strong evaluation gives you a clear, actionable view of the interface’s strengths and weaknesses. Once complete, the findings feed directly into discovery insights and recommendations. They help:
- Prioritise issues based on severity, user impact and effort required.
- Identify next steps from quick wins to areas needing deeper exploration or testing.
- Understand risk and opportunity showing where usability barriers may be affecting user experiences, business goals or regulatory expectations.
- Align around a shared perspective of what needs attention and why.
A successful evaluation builds a grounded foundation for informed decision-making and sets the stage for focused, confident design improvements.
When and why to do a heuristic evaluation?
Heuristic evaluations are most effective early in the discovery phase of a project or early in the prototype phase for new products. Whether you’re completing an evaluation for a website, platform or product, competing it at the start helps you understand structure and flow, identify usability issues before design begins, focus research or user testing on the areas that matter most, and advise on redesign or optimisation strategy.
Beyond finding flaws and comparing against industry standards, a heuristic evaluation is also an amazing way to fully immerse yourself in a digital product. You’ll find pages, links and elements you never even knew existed. That deep familiarity pays off throughout the rest of the project.
The time required to complete a heuristic evaluation depends on the size and complexity of the platform, the number of flows or pages being evaluated, and the depth of analysis expected. As a general guide:
- Small: around 1–2 days
- Medium: 1–2 weeks
- Large or complex: 2–4 weeks
Creating a thorough deliverable like findings, recommendations and annotated screens adds additional time on top.
For larger evaluations, automated tools can help with scale. These tools can be genuinely useful, but they can’t replace a real person working through the site or digital platform. Automated tools are great at catching patterns; they’re not great at understanding context, intent or how something actually feels to use.
Above: 10 Usability Heuristics quick reference guide. A practical way to evaluate digital experiences.
Our 3 steps to complete a heuristic evaluation
Step 1. Planning & preparation
We start with the heuristics, we typically use Nielsen’s usability heuristics, making sure we understand what principles we’re evaluating against and how each one typically shows up. In this stage, we also define the criteria we’ll use to assess each principle and choose specific examples to test against.
This setup work matters. Creating clear, site-specific criteria helps with consistency and focus, ensuring the evaluation is structured without becoming overwhelming. Simply going through the Nielsen principles and cross-checking them against the digita list will have a much smaller impact than investing in a thorough planning phase. The setup is what turns a checklist into a meaningful evaluation.
Step 2. Doing the evaluation
This is the fun part. It’s where you explore, take screenshots, capture notes and mark each criteria as a pass, partial or fail. You’re working through the site, product or platform and testing heuristics in real contexts, observing how the site behaves in actual scenarios and seeing where principles are upheld or violated.
Each time we do this, we:
- Interact with the specific pages, flows or components identified in planning
- Capture instances where the criteria is met or broken
- Note patterns, inconsistencies or recurring friction points
- Mark each example and each heuristic as pass, partial or fail
We also run through real user scenarios, completing tasks as each user type would. We’re looking for the same things throughout, usability issues, friction points and blockers, each mapped back to the relevant heuristics.
Step 3. Synthesis & recommendations
Once the evaluation is complete, it’s time to condense everything into a clear set of recommendations. This is where the the data is turned into something meaningful for the client. The synthesis highlights where the biggest issues are, whether something is a quick fix (like updating links or improving tagging) or a bigger structural problem that needs deeper work.
For client projects, this step is incredibly valuable. It provides a clear direction and makes it easy to see why specific decisions are being recommended. No more guessing, no more debating preferences, just evidence pointing the way forward.
What you get out of it?
A well-executed heuristic evaluation delivers tangible, usable artefacts, not just a list of problems. Typical outputs include:
- A prioritised issue list with severity ratings
- Design recommendations mapped to specific heuristics
- Annotated screens showing exactly where issues occur
- A clear path to quick wins
- Patterns and systemic insights that reveal deeper structural issues
- Strategic opportunities for longer-term improvement
These outputs give teams something concrete to work with, whether that’s fixing things immediately or building a case for larger investment.
Above: Working through digital experiences together to find what works and what doesn’t.
What a heuristic evaluation can’t do?
It’s worth being honest about the boundaries. A heuristic evaluation is an expert-led assessment, and while it’s powerful, it has limits. It can’t tell you what users actually do in practice, it can’t replace behavioural research or usability testing, it won’t answer preference or desirability questions, and it may uncover symptoms rather than root causes.
That’s why we see heuristic evaluations as a starting point, not the full picture. They’re best paired with user research, analytics and testing to build a complete evidence base. The heuristic evaluation gives you the sharp, structured view; research fills in the human story behind it.
Why you shouldn’t get AI to do the bulk of a heuristic review.
Now we get to AI. As mentioned, there are tools that can scan a site holistically, and they’re great for pulling out specific, repeated issues. But using AI to do the entire evaluation is not recommended.
Actually doing the review yourself is what gets you deeply integrated into the site. You understand the pages, the flows, how a user thinks and feels as they move through the experience. A heuristic evaluation is fundamentally an exercise in empathy. Thinking like someone else, noticing what frustrates or confuses, catching the things that only a human would catch. AI isn’t a human so unsurprisingly, it’s nowhere near as good at making judgements from the perspective of a real person navigating a real experience.
Bringing it to life: a recent case study.
To bring this to life, in a recent project we completed a heuristic evaluation for a well-known organisation with a site that looked polished and professional on the surface. You’d assume it was doing its job well. Our evaluation told a different story. Every heuristic we assessed came back as either a partial or a fail. Common user scenarios also landed at partial or fail. Secondary sites needed consolidation or significant enhancement, and the overall digital ecosystem was more fragmented than anyone had realised.
What really stood out, though, was what happened once we read and acted on the actual content as part of the evaluation. At first glance, the UI suggested everything was working. But once we stepped through real tasks and followed the copy as a user would, a series of issues surfaced, inconsistent instructions, unclear pathways and missing or duplicated information. These weren’t immediately visible until we engaged with the content layer, and the evaluation made those problems impossible to ignore.
While those findings can be confronting, that’s exactly what makes a heuristic evaluation so valuable. It turns assumptions into evidence. It gives teams something tangible to act on, rather than debating opinions about what might or might not be working. And the good news is that a lot of what we uncovered were relatively simple changes, the kind of fixes that could turn a fail into a pass without a massive overhaul. The organisation now has a clear, prioritised path forward, grounded in real evidence.
Doing heuristic reviews with Avian.
None of this is especially hard to do but it does take care, structure and genuine expertise to do well. If you’re looking at a website or digital service and wondering whether it’s really working for your users, a heuristic evaluation is one of the best places to start. It’s fast, it’s evidence-based, and it gives you a clear picture of what to fix and where to focus.
We’d love to help. Whether you need a quick review or a comprehensive evaluation as part of a broader discovery, we’ll bring the rigour, the empathy and the practical recommendations to set you up for success.
Sound like the kind of help you could use? Get in touch.