Accessibilité

It’s mid-2024, Google has just released AI Overviews and all we hear is that search is doomed. Well, we’ve been hearing that for a while, but this time it feels a bit more real than it was before.

The current dark ages of the web, the unavoidable ranking declines and having to explain to leadership why we are not ranking ahead of some Reddit user who proclaims you should glue cheese to your pizza would genuinely be enough to push even the strongest soul to consider a career change. I’ve always said I’d leave everything and run away with the circus once the time was right, and this feels like it is the sign I’ve been waiting for.

I started in SEO when Google core updates were a yearly event rather than the order of the day, and when voice search was supposed to be the next big thing (spoiler alert: it never got to that point). ‘Fred’, one of the early core algorithm updates, had just been released, and it introduced an important part of the search journey: user experience (UX). 

For many of us, UX is a bit of a guessing game, based on ‘common sense’ and the inaccurate assumption that what works for us will be fine for everyone else. 

For me, coming from a lifelong passion and years studying the human brain, it was something I’ve always felt drawn to explore and tie with the rest of my digital expertise.

This is what I want to cover in this article - how to analyse and improve your UX even when you don’t have access to a dedicated team or the resources to create a full-fledged strategy.

UX and search

According to Don Norman and Jacob Nielsen, user experience, or UX, is “the end-user's interaction with the company, its services, and its products”.

UX is about solving problems and making things easier for our audience. For some reason, search and UX have historically been separated, with the majority of jobs I have been at having separate departments that don’t even talk to each other. 

The signs of an imminent union were always there. What the modern search landscape is making more plain than ever is that the user is your main stakeholder, not Google.

Documents leaked from the Google trial seem to suggest that user behaviour is, in fact, a ranking factor or at least a contributor to it - although they have always claimed the opposite. While the SEO community still debates on these speculations, it’s still quite clear that we need to focus on the user journey to make our brand and our products immune to the challenge of time (and core updates).

This has led to a revolution of sorts in the industry, with more and more traditional SEOs looking at the entire organic experience rather than the pure SERP on its own, a discipline that is now known as “Search Experience Optimization” or “SXO” (which Sara Fernandez has talked about extensively).

Taking on a UX approach shouldn’t feel like an added burden, since it has a benefit for us SEOs: human behaviour is, for the large part, much more predictable than Google dumpster fires. And at the end of the day, getting a click on our website is only the first part of a larger journey, which needs to be curated in order to keep a user on site and make them consume (and love) what we have to offer.

Qualitative and quantitative measures

Improving UX relies on researching users’ points of friction and designing solutions that can bridge the gaps.

UX research is based on different methods to gather data, most commonly separated in:

1. Quantitative research 

This type of research gives us an output that can be quantified: time on page, sessions, abandonment rate, average order value and any other metric that can be expressed in number form. For example, looking at our users’ engagement data on any analytics platform we use is a good example of quantitative research.

2. Qualitative research

This is anything that is based on observations rather than raw data, and its output cannot be expressed in number. Examples include interviews, heuristics assessment, recordings of someone performing a task on your website. This type of research is generally more open to interpretation than its quantitative counterpart.

From theory to practice: the doctor’s analogy

Fine, we’ve seen all the theory - but how can we put some of these concepts into practice and make sure that our users’ experience on our website isn’t cut short?

Diagnosing UX issues is similar to what I imagine is required from a doctor investigating their patients’ health complaints: analysing symptoms, comparing scenarios and detaching their own assumptions from a patient’s experience, in order to identify the underlying cause of the issue and produce a treatment.

When we carry out UX research, we also need to leave our own assumptions and biases at the door, so that we can actually see or listen to our users’ problems with an open mind. 

Here are a few steps to start doing this from the ground up.

Step 1: Find your “symptoms

Carrying on with the medical analogy, our symptoms of bad UX are the easy ones to start with. 

For anyone working on a website regardless of their job title, these might look like:

• Loss in traffic/low clicks to site

• Lower impressions

• Lower average order volume or conversions

• Higher cart abandonment

…and so on. In short, symptoms are all of the things your leadership team would give you a hard time about, because they are relatively easy to spot and are usually quite uncomfortable to deal with from a business perspective. However, most times these are only the outside manifestation of a deeper issue that you have to dig out from the inside.

So, what are the symptoms you want to resolve? This will inform the next part of the journey.

Step 2: Analyse them

Analysing is probably the most time-consuming aspect of this investigation process. You might have to go through a lot of trials and errors before you can establish the root cause to your specific issue, but luckily there are a number of mediums that can help your discovery journey and cut out a bit of the guesswork. 

These are some of the diagnostic tools you have at your disposal, based on the “symptom” you’re investigating:

1. Google Search Console

2. Google Analytics

3. Surveys and CX data

4. Eye-tracking data and heatmaps

5. Live user testing

Google Search Console (GSC)

If you’re an SEO, you’re probably familiar with GSC, which is great to start gathering some useful information.

If you’re dealing with a loss in traffic, for example, it’s easy to get stuck in looking at clicks only; however I have found that Click Through Rates can be quite telling, especially if we sort the data by ascending average CTR and combine them with monthly search volume to identify priorities.

You may find that there are queries that your website ranks for, but for which you don’t get many clicks. Once you rule out technical issues on the page, low CTRs could indicate different things, including (but not limited to):

• Cannibalisation for the same query

• Poor match of the result to the query of the user

• Incorrect or missing localisation

Take this example: New Balance 550 is a query that has over 470K searches a month, but on this particular website, CTR are really low.

Once we dig into it, we can see that a number of different URLs are competing for the very same query, and the intended page that we want to get clicks on is overperformed by a several duplicates that are less of a good match:

A user encountering these many results for the same query can clearly be confused on which one is the right match, so consolidation is what we aim for to improve their experience. 

In this particular case, improving UX (and SEO) would consist in providing the right localisation and differentiating the title for each one of the alternative pages, implementing hreflang tags correctly and reviewing canonicals for filtered product pages.
However, sometimes it really is as easy as changing the meta title and the description to something clear and concise that matches the query.

Like SEO doesn’t end when the user lands on a page, UX starts even before they interact with the site at all.

Google Analytics

Yes, GA4 is the spawn of the devil and we all hate it, but the user behaviour metrics hidden in there can hugely help your UX ‘diagnosis’.

I won’t expand too much onto this one because this guide tells you much more I could ever cover in this article, but some of the things you might want to look at in your ‘bad UX’ symptoms exploration are:

• A spike in bounce rates (yes, I am bringing them back from oblivion - they might indicate poor match to intent, or a soft 404 page)

• Higher internal searches (this might mean categories or products are hard to find in your navigation)

• Average time on page and deviations from it

• Low form completion rates

• Points of abandonment

The list goes on. Luke Hay has written an entire book on the topic, and while it was released pre-GA4, many processes still apply today, so it’s another invaluable resource specifically for your UX journey on Google Analytics.

Surveys and CX

Team work makes the dream work! 

What do your customers actively complain about all the time? What are some of the obvious questions that keep coming up to your customer service agents? Even if you only have a chatbot and not an actual team, have a look at the chat log and see if you can identify some common patterns (even ChatGPT can help categorise frequent topics!).

This gives you a list of pain points to address proactively. 

This study from Baymard, 2024, is one of the best examples of this: most of the reasons for cart abandonment listed below are, in fact, things that have for sure come up in customer service.

Eye-tracking data & heatmaps

You might write amazing content and provide all the info your customers are looking for, but they won’t be able to appreciate it if they don’t actively come across it.

Attention is a precious commodity in the land of 24/7 stimuli. We don’t process what we don’t see (well, sometimes we implicitly do, but that’s a topic for another article) and eye-tracking recordings can help us find areas that our gaze doesn't reach.

There are free tools with limited runs you can use for eye-tracking studies, even remotely as the calibration and recording are powered via webcam, and the ones I have most recently tested are RealEye and GazeRecorder. Eye-tracking data inform where our users are (or are not) looking, and interaction heatmapping tools can integrate that information with clicks, mouse scrolls and position data, uncovering not only areas that might not be getting enough attention, but also elements that don’t actually work (the infamous role of rage clicks).

Some tools that offer heatmapping are for example Mierica Heatmap, Lucky Orange and Mouseflow, and a great resource where you can see all this data used in practice is Kinichi Suzuki and Rio Ichikawa ‘s presentation at BrightonSEO in 2023.

Image source: Hubspot

Live user testing

This is, by far, the most time consuming and less scalable option - but probably the most rewarding as well. At the end of the day, tools help us infer human behaviour from data points, but nothing beats going to the source and talking to an actual person about their real frustrations. 

This book by Stephanie Marsh offers great guidelines to start user testing in a controlled environment with different groups, but don’t let scientific rigor be a limit to speak to your users: even when you don’t have the option to run a fully fledged testing session with multiple participants, a zoom call with someone outside your team can absolutely be enough to get some valuable insights on things you can further investigate and improve.

Step 3: Establish the root cause

The analysis step should have helped already steer you in the right direction and avoid most of the guess-work. At this point, you should already have some hypothesis to test, but from my experience, here are some of the root causes that are often overlooked and you probably need to consider:

• The experience on mobile is not the same as on desktop, and was not tested properly

• Text size, fonts or format make the content hard to read (especially on mobile)

• Page load times are long

• The site is not responsive

• The site is not accessible to screen readers

• Your product, services or mission are unclear

• Crucial pieces of information are missing from a transactional page (e.g. pictures)

• Your navigation is erratic, or not clickable

• You have intrusive and aggressive pop-ups

• Sticky navigation or elements are covering selectors

• The pagination is broken and prevents users from navigating to different options

…and much more.

Here are a few practical examples I’ve encountered recently:

1.Booking a hotel room in the South of France

Wow, I can upgrade the room for a small price? Awesome! 

But wait - what perks do I actually get in return?

The only indication I have to make a decision is the size of the room. The ‘Room Information’ button doesn’t expand, and no other details on the page justifies why I should pay more for this superior room - so I decided to go for the standard instead.

Bottom line: for a failed upgrade or low AOV, make sure that your users have all information at hand to inform their decision.

2.Choosing a last minute gift on Fortnum and Mason’s website

I wanted to buy my friends a nice hamper to celebrate their engagement, and I took advantage of my train journey to do some research on Fortnum and Mason’s website.

However, once I started sorting the hampers by price, the pagination was no longer working, so I could only see a fraction of their offer and not the full range they had available. This likely doesn’t happen on their desktop site, but I wouldn’t know since I abandoned my journey there and purchased my gift from a competitor instead.

Bottom line: Always test your mobile website too! A lot of user journeys are fully carried out on a different device than desktop.

3.E-dreams: a 9-steps journey to cancel my trial membership!

I will save you the screenshots I took from this endeavour of mine just because it’s nine of them and we would’t have the space for it. We’re all into a good deal and I was guaranteed one with a trial, timed membership on E-dreams. We know that a lot of these apps bank on the fact you’ll probably forget to cancel on time, but even when you’re me and you have calendar alerts for everything, you might have a hard time trying to get out of it. The 9 steps bank on loss aversion and a several reverse psychology tricks stolen directly from a Gen-X parent’s playbook, which I do get - but I do not condone. This opens a bigger conversion of dark patterns in UX and the several losses you might face as a business by employing these practices - users tend to remember these frustrations and are more likely not to consider your brand at all next time.

Bottom line: Just don’t. That’s the bottom line.

Step 4: Prioritise and fix

Once you’ve found your root cause, you will need to plan how to fix it, which is likely going to require collaborating with other teams and prioritising your initiatives.

Similarly to when a doctor prescribes a treatment, it is more likely they will focus first on a debilitating stomach ache rather than athlete’s foot, and we should also prioritise what issues we want to treat first when we uncover them in our UX research. However, if I am a runner training for the upcoming Olympics, my doctor might decide that treating my athlete’s foot is at least as important as focusing on my stomach: this is to say that while there are general guidelines to inform priorities, yours might vary depending on what your overall goal is.

The size of your website and company will determine how much effort you put into the prioritisation process. Enterprise companies with dedicated engineering teams will likely work in sprints and will require you to slot your project into their backlog.

To inform priorisation, you can start by asking yourself (and answering truthfully):

• What’s the time/effort involved in this fix?

• How critical is this fix to the success of the business? 

• This fix resolves a blocker to navigation or conversion = urgent

• This fix is a nice to have feature = not urgent

• What’s the impact or ROI of me fixing this issue to the wider business? 

Based on these answers, filling out this matrix of prioritisation will help your (and your team’s) planning a roadmap to improve UX.

Image source: LogRocket.com

Final words

You don’t need to be a UX professional to investigate and improve user experience, and your title shouldn’t be an excuse not to deliver better content or products to your audience.

If you’re flying solo, start with some of the tools and resources provided in this article, and expand by looking at how other websites are doing UX, and what professionals in the UX space are saying (Stéphanie Walter regularly shares informative and easily digestible nuggets on LinkedIn).

If you work on enterprise websites, make sure you connect with your UX team and have regular catch-ups. You will likely find overlaps and common missions between your department’s projects and theirs. 

UX and search are more interconnected than ever right now, so getting on board the “user delight” train regardless of your background is not only a way to improve our websites, but to future-proof our career in digital.