Why Hick's Law Fails UX Design
Transcript
Made automatically from the audio, so it may contain mistakes. Click a line to play from there.
Picture this. You are logging into a brand new educational portal. Let's just call it the pathway portal for now. Right.
Okay.
And you're trying to figure out which course to take to, you know, advance your career. You hit the homepage, and you are immediately confronted with scenario A.
Let me guess. It's a mess.
Oh, it is a chaotic flat wall of 48 different courses. But they aren't named anything helpful. They are all titled using this like dense internal curriculum jargon that makes absolutely zero sense to an outsider.
Right. Like you're supposed to just instinctively know what they mean.
Exactly. You're staring at tiles that say things like Synergy-402 and Pathway Matrix Beta. I mean, it's totally paralyzing.
Yeah, no, it's the digital equivalent of walking into a warehouse where all the boxes are labeled with barcodes instead of, you know, actual product names. You have all the options in the world and absolutely no way to evaluate them because the system is demanding that you already speak its internal language.
Right. So you hit the back button, you decide to try a different portal. This one, scenario B, promises a quote, modern minimalist AI-driven experience.
Oh. Oh, I know where this is going.
Yeah. A designer somewhere proudly cited Hicks Law in the release notes to justify this layout. So instead of 48 courses, you just see this vast expanse of white space and three poetic mysterious buttons. Just three. Just three. Builder, explorer, and visionary.
Oh wow.
You have literally no idea what they actually mean, but you click visionary, you go through a multi-step wizard, and right at the end, like on step four, you notice they've sneaked a pre-checked 49 certification upgrade into your cart.
Of course they did.
Yeah.
I mean, both of those scenarios are disastrous, but for like entirely different reasons. One overwhelms you with raw, unfiltered complexity, and the other hides the complexity behind this opaque curtain. You know, you know.
Stripping you of your agency.
Exactly. Stripping your agency and tossing in a dark pattern just to exploit your fatigue at the end.
Okay. So let's unpack those. If you are with us for the last deep dive, you know that having built our three-tiered lens separating mathematical models, psychological effects, and context-dependent heuristics across Norman's two gulfs. We are now going to investigate what actually happens inside a human brain when facing menus, catalogs, and multi-step forms.
And that is the exact mission for today's deep dive. Looking at the massive stack of research you sent over, the articles, the historical psychology papers, your notes. We are putting Hicks Law, the concept of choice overload, and the infamous three-click rule under the empirical microscope.
Because we really need to separate real cognitive science from those user interface design myths that lead to the broken scenarios we just talked about.
Right. Absolutely.
And the sources are just fascinating. We've got historical data from the 1950s, meta-analyses, cognitive load theory. But to really get to the bottom of this, I want to start with a classic category error in design.
Okay, yeah, let's jump in.
We need to look at what human reflexes actually tell us, and more importantly what they don't tell us about complex decision making.
Yeah. To understand the misapplication of Hicks Law in user experience, we really have to travel back in time. We need to look at what William Edmund Hick in 1952 and Ray Hyman in 1953 actually measured in their laboratories.
Because I mean, designers love to throw the term Hicks Law around to justify hiding options, like in our scenario B with those three vague buttons.
Right. They use it as a shield.
Yeah. They point to the math and say, well, Hicks's law says more choices mean it takes longer to decide, so we must drastically reduce choices to make the interface faster. But the original experiment wasn't about a user reading through 48 course descriptions, was it?
Not even close. I mean, Hick and Hyman were studying choice reaction time to highly practice simple stimuli. Let's break down Hicks' 1952 laboratory setup. So he had participants sit in front of ten lamps arranged in a circle. Each lamp was paired with a specific Morse key, and the participant operated those keys with different fingers. And a pre-punched tape would activate a random lamp every five seconds.
A pre-punched tape, like an old player piano role.
Very similar. Yeah. It was a mechanical way to ensure true randomness in which lamp lit up without any human bias. So the tape would feed through a machine, and with a hole in the tape passed over a sensor, it completed a circuit and flashed a specific lamp.
Okay. So it's essentially a physical reflex test, a light flashes, you press the corresponding button with the correct finger. It sounds like a gamer executing a quick time event in a video game or like playing whack-a-mole.
That is a perfect modern equivalent.
Whack-a-mole.
Hick measured the reaction time, the exact physical time it took for the participant's sensory system to perceive the light, send a signal to the brain, and for the motor system to hit the key.
Right.
And what he found was that the relationship between the number of choices, meaning the number of active lamps, and the reaction time wasn't linear. It was logarithmic.
Okay, so we promised in our recap that we wouldn't read raw mathematical notation aloud on this show. So let's translate that logarithmic curve into vivid plain English. Because the math here is deeply misunderstood by almost everyone who quotes the law.
Let's do it. So if the relationship were linear, you might assume that if you double the number of lamps from four to eight, the reaction time would also double.
Right. Makes intuitive sense.
Yeah. If it takes you one second to react to four lamps, a linear law would say it takes two seconds to react to eight. But because a relationship is logarithmic, doubling the number of highly practiced choices doesn't double your reaction time. It merely adds a tiny constant fraction of a second.
So going from four lamps to eight lamps might just add, say a hundred and fifty milliseconds to your reaction.
Yes. And going from eight lamps to sixteen would add that exact same tiny hundred and fifty milliseconds again. The curve flattens out pretty quickly. And this ties directly into how they measured the information being processed using Shannon information entropy.
Right, which is measured in bits, correct?
Exactly.
See, I always hear the term bits in computer science, like megabytes and gigabits, but how does it apply to human psychology in a flashing lamp?
Well, in this context, a bit is just the amount of information needed to half the uncertainty of a situation. It's the mathematical equivalent of a binary search.
Okay, give me an example.
Imagine you were trying to guess a number between one and one hundred. If your first question is, is it higher or lower than fifty? That single question, that single bit of information cuts the possibilities completely in half.
So you've eliminated 50 wrong answers in one move.
Precisely. The brain here is essentially performing this rapid subconscious binary sorting when a light flashes. And the math shows that reaction time grows gently as the logarithm of choices increases, weighted by how likely each choice is.
Wait, weighted by likelihood. What do you mean by that?
Well, if the participant learns over time that the top lamp flashes more often than the bottom lamp, their brain actually anticipates it. That high probability shortcut actively shaves off reaction time. The nervous system preloads the response for the likely event.
Wow. So let me get this straight. Hick and Hyman were measuring people who had essentially memorized a physical reflex to a flashing light.
Right.
They were testing the limits of motor sensory pathways and subconscious binary sorting. But today, digital designers use the specific mathematical law to justify hiding options from a new user who is trying to choose a college major or buy a complex financial product.
You've hit the nail on the head regarding the category error. It is a massive failure of translation from the laboratory to the real world, validating this motor reflex curve and then applying it to unpracticed, complex choices is fundamentally flawed.
It's like comparing the time it takes to blink when a puff of air hits your eye to the time it takes to decide who you want to marry.
Yes, exactly.
I mean, they both involve the brain, but they are completely different mechanisms.
That is the exact distinction. When you were looking at 48 courses on a portal, your brain isn't just reacting to a visual stimulus like a moth to a flame. You are evaluating semantics. You're digging into your long-term memory, projecting your future career path.
Weighing personal preferences, reading text.
Yes, comprehending concepts. Hicks' law simply does not model that behavior, it models reflexes.
So designing an interface based on Hick's law when the user has to make a complex choice is optimizing for the wrong biological system. You're optimizing for the spinal cord and the visual cortex when the user is actually using their prefrontal cortex to reason and evaluate.
That is a brilliant way to phrase it. And that is why scenario B from our opening, the one with just three vague buttons, fails so hard.
Right, because the designer thought three buttons means less reaction time than 48 buttons.
Exactly. But because the buttons said builder, explorer, and visionary, the user's prefrontal cortex has to work overtime to guess what those words mean in this context. The physical reaction time might be a millisecond faster, but the cognitive decision time skyrockets because the meaning is obscured.
But wait, even if Hicks law doesn't apply to complex choices, there is another concept in your research stack that we have to tackle. Choice overload.
Yes. The famous paradox of choice.
Yeah, because we've all been told repeatedly that giving users too many options freezes them.
Right, that's the prevailing wisdom.
And the classic example everyone cites is Sheena Iyengar's jam tasting study from 1995. Let me just set the scene for this because it's such a famous story.
Go for it.
So the researchers set up a tasting booth at an upscale grocery store. On one day, they put out a massive display of 24 gourmet jams. On a different day, they put out just six jams. Customers could taste as many as they wanted, and if they stopped by, they were given a discount coupon to buy a jar.
And the results of that single study became absolute gospel in the design, marketing, and behavioral economics worlds.
Because the numbers were so stark. The study found that while the 24 jam display attracted more people initially, only three percent of the people who stopped actually bought a jar.
Right, a tiny fraction.
But for the sixth jam display, 30% of the people made a purchase. A huge difference. So the conclusion was blasted everywhere. Vast arrays of choices paralyze people. More options lead to fewer decisions.
It is an incredibly compelling narrative. And it certainly feels true when we are staring at a massive disorganized menu. But if we look at the empirical science over a larger timescale and across different domains, the picture changes dramatically.
Okay, so this brings us to that landmark 2010 meta-analysis by Scheibenheny, Gryfonetter, and Todd that you included in the sources. Yes. And for anyone listening who might not read academic papers all day, a meta-analysis is when researchers take a huge number of individual studies, sometimes decades of research, and combine all their data to see if the overall effect actually holds up across different conditions and populations.
Right. It's the ultimate reality check for a scientific theory. And these researchers analyzed 50 different studies on choice overload. 50. 50. Encompassing a massive 5,036 participants. They looked at studies involving everything from jelly beans and chocolates to complex financial investments in real estate.
Wow. Okay. So what did they find when they zoomed out and looked at those 5,000 people?
What's fascinating here is that the average choice overload penalty across all those studies was virtually zero.
Hold on. Zero? Come on. I am going to push back on this hard.
I figured you might.
I have literally stood in the toothpaste aisle of a pharmacy for 10 minutes, completely paralyzed by 80 different types of mint gel. Do I want tartar control? Whitening, enamel shield, breath defense. Are you telling me my lived experience is a myth? Are 50 studies calling me a liar?
No, no, they are calling you a liar. But they are redefining what actually caused your paralysis. When you aggregate all the data, the crippling paralysis we strictly associate with a sheer volume of options essentially vanishes.
Really?
Yeah. In fact, in many cases, they found a more is better effect. People often preferred and were more satisfied with larger assortments of choices.
Okay. So if the penalty for high volume is zero, why do I freeze in the toothpaste aisle? Why did the jam study happen the way it did? Why does our 48 course scenario A feel so viscerally terrible?
Because the raw number of options isn't the problem. The problem is the structure of those options, the presence of dominant choices, and the prior knowledge of the user.
Oh, interesting.
Yeah, people actually love rich massive catalogs, provided the options are well categorized and easily comparable.
So it's not the volume of the information, it's the organization.
Yes. The meta analysis revealed that whether you experience choice overload depends heavily on whether you are a preference matcher or a preference constructor.
Okay. Let's break those down. What exactly is a preference matcher?
A preference matcher is an expert, or at least someone who already knows exactly what they want before they encounter the choices. They have well-defined pre existing preferences.
Okay, so they have a goal in mind.
Exactly. For a preference matcher, a huge catalog is wonderful. If you are a coffee aficionado and you know you want a lightly roasted Ethiopian Yurgachef with floral notes and a natural process, a store with 200 coffees is a playground.
You simply scan the environment to match your exact preference to the inventory.
Right.
But if I walk into that same store and I only drink coffee twice a year and I just want something that tastes like chocolate and doesn't make me jittery, I am a preference constructor.
Exactly. You are trying to figure out what you want while you are looking at the options. You're constructing your preference on the fly.
That sounds exhausting.
It is. For preference constructors, a massive unorganized catalog is a nightmare. It requires immense cognitive effort to weed out the irrelevant options, figure out the baseline criteria for making a choice, and you know, weigh variables you don't fully understand.
So back in my toothpaste aisle, I'm a preference constructor. I don't know the chemical difference between Stannis fluoride and sodium fluoride. All 80 boxes are shouting different benefits at me, and they aren't organized in a way that helps me learn the criteria.
Exactly. So I'm not paralyzed by 80 choices. I'm paralyzed because the interface of the shelf forces me to do a biochemistry degree to make a decision.
That is the perfect distinction.
The cognitive load comes from the lack of structure, not the number 80. And if we look back at the original jam study, some researchers have pointed out that the jams in the 24 condition didn't have a clear dominant choice.
Oh, they were all weird flavors.
Yeah, they were all equally unfamiliar gourmet flavors. If one of the 24 jams had been a familiar, high quality strawberry preserve, a dominant choice, the paralysis likely would have vanished because constructors would have a safe baseline to anchor against.
So if you're designing our educational portal, your job isn't to artificially hide choices like our scenario B did with those three mysterious buttons. Your job is to organize the choices so the user's brain doesn't have to work so hard to categorize them.
Precisely. You have to support both the matchers who want to dig deep into the 48 courses and the constructors who need guidance on how to evaluate them.
Right. You provide a structure that aligns with the user's mental model. When options are properly categorized, people don't suffer from choice overload, regardless of how many options there are.
Yes. And this directly challenges another massive myth in the UX world, which governs how people actually navigate through these organized spaces.
Oh, I know where you're going with this. Because if I'm building a massive, deeply categorized menu to support all these options, someone is inevitably going to yell at me about the three-click rule.
The three-click rule. The old heuristic that claims users will become frustrated and abandon your website if they can't reach their final goal in exactly three clicks or less.
It is one of the most stubborn myths in digital design.
It really is.
And it forces designers to do crazy things. If you believe the three-click rule as absolute law, you can't have a deep logical hierarchy. You have to flatten everything out, shoving massive amounts of links onto the homepage or grouping completely unrelated things together just to keep the click count under the magic number of three.
And much like Hicks' law, the three-click rule has no empirical basis. Usability studies like the foundational research from UIE and Joshua Porter that you highlighted in the notes have proven this false time and time again.
I mean, nobody likes clicking 20 times to find a PDF.
True, nobody likes useless clicking. But what Porter and UIE actually measured when they track user clicks across massive websites was eye-opening. They found that user drop-off does not increase after three clicks.
Really?
Really?
No cliff at click number four.
No cliff at all. The completion rate doesn't suddenly plummet on the fourth click. And user satisfaction doesn't drop either, provided, and this is the crucial non-negotiable part, provided the user feels they are continually getting closer to their goal.
So as long as I feel like I'm on the right path, my brain is willing to invest the physical effort of clicking.
Yes. The physical cost of a click is microscopic. The mental cost of not knowing where to click is massive.
Right.
Which brings us to a brilliant concept from researchers Peter Paroli and Stuart Card. Information foraging theory.
Oh, let's dive into this because this analogy is arguably the most powerful tool in the research stack.
It's fantastic. So human beings evolved as foragers. Parolian Card took this evolutionary behavior and applied it to web navigation. They introduced the concept of information scent.
Like an animal tracking a scent.
Exactly. Think of a user navigating a digital interface the exact same way an animal forages for food in the wild.
I love this.
Okay. So the user is a predator. Let's say a wolf, hunting for a specific piece of information.
Right.
And the wolf uses scent cues in the environment to estimate where the prey is. In digital terms, the user evaluates a link or a button, what the researchers call the remote representation, to estimate the value of the actual destination page, which is the source or the prey.
So the label on the link is the scent in the air.
Exactly. If you are looking for a broken water meter report form on a city government website, and you see a link that says utility complaints, that link has a strong information scent.
You can smell your target through that link.
Yes. You will happily click it, and if the next page gives you links like electric gas and water, the scent gets even stronger. You are getting closer.
And you're happy to make five, six, or seven clicks if the scent trail is strong, continuous, and highly rewarding.
Absolutely.
So if a designer forces a deep, multi-level menu to flatten out just to satisfy a fake three-click rule, maybe by grouping utility complaints under a vague heading like citizen resources to save a click, they're essentially scrambling the scent.
Yes. They are masking the scent with noise. When you group things illogically just to flatten the architecture, or when you bury links behind vague, poetic labels like the visionary button from our scenario B cold open, you destroy the information set.
The wolf can't smell its prey.
It has no idea if visionary means taking an art class, starting a business, or upgrading to a premium account.
And what happens to our predator when the scent is lost?
You get a behavior called pogo sticking.
Pogo sticking, like jumping up and down?
Pretty much. The user clicks a link, waits for the page to load, realizes the page is wrong, clicks the back button, tries another link, hits the back button again, they are bouncing in and out of pages fruitlessly.
I hate doing that.
Everyone does. That is what causes frustration and abandonment. It is the loss of scent, not the total number of clicks.
If we're acting like predators, sniffing out links, and bad links scramble that scent, what happens when we hit a wall? The brain's energy tank must empty out pretty fast. Pogo sticking sounds exhausting.
It is exhausting. And it leads us directly to the biological limits of the brain. This is where John Sweller's cognitive load theory becomes essential.
Sweller developed this theory in the late 1980s, initially studying problem solving and instructional design, right?
Yes. And it's built on a fundamental neuroscientific reality. Human working memory is extremely limited.
We can only hold and manipulate a tiny amount of information at a given time in our active consciousness.
Right. Think of working memory like the RAM in a computer. It is fast, but its capacity is small. If you fill it up with background processes, the whole system grinds to a halt.
And Sweller breaks down how we use this limited mental bandwidth into different types of cognitive load. What are the ones we need to care about for our portal?
The two most important types for our discussion are intrinsic cognitive load and extraneous cognitive load.
Let's unpack those. I want to know the difference down to the mechanism.
Sure. Intrinsic cognitive load is the mental effort inherently required to understand the information or make the decision. It is the unavoidable natural complexity of the task itself.
Give me an example in our portal context.
Well, if you are deciding whether to enroll in a postgraduate degree in astrophysics, that is an intrinsically complex decision. You have to hold a lot of heavy variables in your working memory, tuition cost, time commitment, calculus prerequisites, future career outcomes.
And you can't design that complexity away. It's baked into the reality of the task. If you hide the tuition cost to make the page look cleaner, you aren't helping.
Exactly. You shouldn't try to oversimplify it to the point of abstracting away necessary details. Extraneous cognitive load, on the other hand, is the unnecessary mental processing caused by bad UI, distracting design, or confusing jargon.
Oh, like the bad labels.
Yes. It's the effort you expend trying to figure out how to use the interface rather than focusing on the decision itself.
So in our scenario A, the wall of 48 jargon-filled courses, the extraneous load is off the charts. The user is burning all their working memory, just trying to decode what Synergy 402 means.
Right. They are translating.
And by the time they figure it out, their RAM is full. They have no mental bandwidth left to evaluate if it's actually the right course for their career.
Precisely. And in scenario B, the minimalist layout, the extraneous load is equally high, just in a different way. The user has to burn working memory, guessing what the vague visionary button means. And then they have to navigate a wizard that hides the necessary intrinsic information.
So what does this all mean for the visual layout? Do we want it sparse or dense? Because we've established that hiding things like scenario B is bad, but a wall of text like scenario A is also bad.
Well, it depends entirely on the user's context. And this is where we have to look at specific UI patterns. Let's contrast two very different approaches to managing cognitive load.
Okay.
First consider the gov.uk one thing per page pattern.
Oh, the gov.uk design system is famous in the UX world. It is incredibly stark, black text on a white background, extremely functional.
Yes. And there's a profound psychological and biological reason for that. They're designing for citizens who are often anxious, infrequent users.
What kind of tasks are they doing?
Think about someone applying for a passport because of a family emergency or reporting a crime or applying for government unemployment benefits.
They are stressed.
Highly stressed. And we know from neuroscience that anxiety and stress release cortisol, which actively impairs the prefrontal cortex.
Wait, really?
Yes. Their working memory capacity is literally physically reduced by their emotional state.
Wow. So their RAM is already half full of anxiety before the page even loads. How does the one thing per page pattern help them?
It ruthlessly eliminates extraneous load. By asking only one question per page, for example, a giant page that just asks, what is your date of birth? And providing absolutely no other visual distractions. They isolate the decision.
They just have to focus on that one thing.
Exactly. They use massive clear labels. They even avoid complex progress trackers on the side of the screen.
Really? No progress tracker. I feel like designers always want to show you where you are in a flow.
For a stressed user, showing a progress tracker that says step two of 45 can actually induce panic. It adds extraneous load because the user starts worrying about the 43 steps they haven't seen yet, instead of just typing in their birthday. The one thing per page approach acknowledges the limited working memory and guides the user through safely.
That makes total sense for a stressed citizen, but what about the other extreme? What about someone who isn't anxious but is a highly trained expert?
That requires a completely different pattern. The dense single screen workspace.
Okay, set the scene for this one.
Think about an admissions coordinator at a university who processes 80 complex student applications a day, or a financial trader using a Bloomberg terminal, or an emergency 911 dispatcher. For an expert user who works in the system for eight hours a day, switching pages adds extraneous load.
Because they have to hold the information from page one in their head while they wait for page two to load, their brain has to bridge the gap.
Exactly. For an expert, the intrinsic load of cross-referencing data points is high. They need to see a student's GPA, their essay score, and their financial aid status simultaneously to make a decision.
They can't be pogo sticking between tabs.
No, they need all the information visible at once. They need high-density, easily chunked information on one screen, often optimized for rapid keyboard shortcuts rather than mouse clicks. A one thing per page flow would completely break their workflow, destroy their efficiency, and frustrate them endlessly.
Right. And this raises an important point about how we manage that load, especially when we start pre-selecting things for users to supposedly help them.
Ugh, defaults.
Yeah. Because when we design these systems, we have the power to set defaults to reduce the number of micro decisions a user has to make. But looking at your notes on cognitive load, there is a massive difference between an ethical smart default and a dark pattern.
Indeed. An ethical smart default leverages the fact that humans, under cognitive load, will often choose the path of least effort. It pre-selects the most beneficial standard or safest option for the user.
Give me an example of an ethical one.
For instance, in a retirement savings app, defaulting the contribution rate to a sensible 5% rather than zero, or defaulting a software installation to the standard folder rather than making the user specify a system path. It reduces extraneous cognitive load while acting in the user's best interest.
But a dark pattern uses that exact same psychology against the user. It weaponizes cognitive load.
It does.
Like in our scenario B cold open, where step four, if the wizard had a pre-checked $49 certification upgrade.
That is a classic dark pattern. It exploits the user's decision fatigue. By the time they reach step four, their working memory is taxed. Their information foraging instinct is just telling them click next to get the reward.
They aren't even reading anymore.
Right. The dark pattern adds extraneous load if the user actually spots it, because they have to pause and figure out the UI mechanism to uncheck it. But mostly it relies on the user missing it entirely due to cognitive overload.
It's literally setting a trap right at the feeding ground when the predator is exhausted.
Exactly. It's a breach of trust.
Okay, so we have deconstructed the myths. We know Hicks Law models reflexes, not complex choices. We know choice overload is mitigated by good structure, transforming preference constructors into matchers.
Right. And we know the three click rule is a myth, and information sent is what actually drives navigation.
Yes, and we know we must manage intrinsic and extraneous cognitive load based on the user's context.
We have the theory. Now we need the application.
Exactly. So let's put it to work. We are going to spend a significant portion of this deep dive taking all this empirical science and applying it to our broken pathway portal from the beginning. Let's completely rebuild it, focusing on multimodal accessibility, visual mechanics, and cognitive psychology.
Let's do it. We have 48 courses to present to a prospective learner. The learner needs to move through three distinct states: the discovery phase, the evaluation phase, and the checkout phase.
Starting with discovery. How do we take that chaotic wall of 48 jargon titles and apply information sent so people can actually find what they need without pogo sticking?
We use a categorized mega menu. Nielsen Dorman group research has consistently shown that for large e-commerce or educational sites, mega menus vastly outperform standard single-column hierarchical drop-downs.
And for those trying to visualize this, a mega menu is one of those large two-dimensional panels that drops down from the main navigation bar and takes up a good chunk of the screen, showing you multiple categories at once, right?
Correct. Instead of hiding the 48 courses behind a single click or spreading them across three pages, we group them into logical scent-bearing categories. We chunk the options into related sets based on the user's mental model.
Like grouping them by subject.
Yeah, perhaps organizing them by domain. Technology, business, and design. Inside those columns, we have level two and level three group headings. This allows the user to see the entire landscape at a glance.
And this leans on the cognitive principle of recognition over recall.
Exactly. And we fix the jargon. We use clear descriptive labels, no more Synergy 402. We use terms that have incredibly high information scent, like introduction to data science or advanced project management.
But the visual layout is only half the battle. The mechanics of building this mega menu at the code level are just as critical, especially when we talk about motor and sensory systems. I want to dig into the timing of the hover state.
This is a crucial, often overlooked detail. If the mega menu opens the exact millisecond the mouse touches the trigger word, you create chaos.
Because nothing is more annoying than moving your mouse diagonally across the screen to reach a search bar and having three massive menus flash open and close like a strobe light as your cursor passes over them. It's so jarring.
It creates massive extraneous cognitive load. That's why when we rebuild this, we implement a strict 0.5-second hover delay. The menu should only open if the cursor rests on the trigger for half a second.
What's the specific biological reason for half a second?
It distinguishes between a user intentionally hovering to open the menu and a user just dragging their mouse across the screen. It also accounts for mild motor tremors. If a user's hand shakes slightly, the menu doesn't violently flicker. It waits for intent.
That's brilliant for the mouse user. But we talked about multimodal accessibility in a previous deep dive. We can't assume everyone is using a mouse or looking at a screen. How does this mega menu work for someone using a screen reader or navigating via a keyboard?
This is vital. And it is where many complex UIs fail. Imagine a user navigating only with the tab key on their keyboard. If they hit the courses trigger and the mega menu opens, we cannot force them to hit the tab key 48 times to get through all the course links if they realize they are in the wrong menu.
That would be an enormous extraneous load. It's a digital trap.
Precisely. So we must build the menu with proper ARIA navigation landmarks.
Stop right there. For the non-developers listening, what does ARIA actually do in the code? How does a screen reader interact with an ARIA landmark?
Aria stands for accessible rich internet applications. It's a set of attributes you add to the HTML code. Think of ARIA landmarks like a highly structured, invisible table of contents that only the screen reader can see.
Oh, invisible to sighted users but read out loud.
Right. When a visually impaired user lands on the page, the screen reader uses the ARIA code to announce you are in the main navigation region. This is a menu with three subsections.
So it gives them the architectural blueprint of the page without them having to read every single word to figure it out.
Exactly. It provides structural information set. Furthermore, we build expandable trigger buttons that explicitly announce their state. The screen reader won't just say courses, it will say courses button collapsed. And when they activate it, it says courses button expanded.
And what about that keyboard trap? How do they get out without tabbing 48 times?
We engineer an instant escape key dismissal. If a keyboard user opens the menu, reads the first heading, and realizes the scent is wrong, they don't have to tab through the contents. They just hit the escape key.
Boom, menu closes.
The menu instantly closes, focus returns to the main trigger, and they can tab once to glide right over to the next major category.
It preserves their physical energy and keeps the interaction cost incredibly low. It respects their time.
Okay, so the user has successfully used the high synt mega menu to fight a category. Let's say they've navigated to the technology section. Now they are moving from discovery to the evaluation phase. They are looking at five different data science courses, and they need to choose one.
And we know from our discussion on choice overload that they might be preference constructors at this point.
Exactly.
How do we present these five courses without triggering paralysis? We definitely can't just dump all the syllabus details on them in a massive wall of text.
No. We replace the walls of text with a scannable multi-track comparison table. Remember, preference constructors need help identifying the baseline criteria for making a decision.
So we provide the structure they need.
Yes. But we must be disciplined. We cap the comparison table at exactly five decisive learner attributes. Perhaps we compare the courses strictly based on prerequisites, time commitment per week, career outcome, format, like live versus asynchronous, and cost.
Why cap it at exactly five? Why not show all 20 variables?
Working memory. Okay. If we show 20 variables, we overwhelm the brain's RAM. By limiting the comparison to the five most critical tracks, we prevent extraneous cognitive load while fully respecting the intrinsic weight of the decision.
We are giving them the exact tools they need to transition from a preference constructor to a preference matcher.
The visual scanning of a table is also powerful. You can run your eye down the cost column and instantly see the differences, rather than reading five separate paragraphs trying to extract the price from each one.
It relies on visual cortex's ability to quickly parse grid structures. But again, we must think about accessibility and responsiveness. A massive five-column table looks great on a 27-inch monitor, but what happens for a low vision user who has their browser zoomed in to 400%?
That's a great question. If I zoom in that much, a standard HTML table usually breaks. It stays rigid, and suddenly I have to scroll horizontally back and forth to see the columns. And horizontal scrolled the first column off the screen and you've forgotten the name of the course. You lose the context of the row headers.
It's terrible.
So how do we fix it at the code and design level?
We must engineer this comparison table using modern CSS grid techniques so that it reflows cleanly down to 320 pixel width, which is the equivalent of viewing a desktop site at 400% zoom.
What does a reflow table look like in practice?
The horizontal rows must collapse and stack into vertical independent cards. So instead of one row spanning across five columns, course A becomes its own discrete card displaying the five attributes vertically. Course B sits directly below it.
I see.
So at 400% zoom, the user still has a logical, vertical, coherent flow of information. They just scroll down naturally. No context is lost, off screen, and the cognitive load remains manageable.
That is a beautiful piece of engineering. We are anticipating the sensory needs of the user and adjusting the structure without degrading the information. Okay, so our user has evaluated the five data science courses using the table. They found the perfect fit.
They click enroll. Now we enter the final step, the checkout phase.
First and foremost, we ruthlessly rip out that sneaky $49 pre check certification upgrade from scenario B.
No dark patterns, no setting traps at the feeding ground.
We replace it with transparent, plain language choices. Specifically, we align this step with the W3C COGA guidelines.
Let's unpack that acronym for the listeners. What is COGA?
COGA stands for the Cognitive and Learning Disabilities Accessibility Task Force. It's a specific set of guidelines aimed at making web content usable for people with cognitive and learning disabilities, like dyslexia, ADHD, or memory impairments.
But as with all accessibility features, applying Koga guidelines makes the interface vastly better for everyone, right? Especially users experiencing temporary cognitive load from stress or fatigue.
Exactly. So what does a COCA compliant checkout look like? It relies heavily on clear, unambiguous language, and clear visual affordances. We don't use double negatives in checkboxes like uncheck this box if you do not want to opt out of the upgrade.
Oh my goodness, my brain hurts just hearing that sentence. That is an extraneous load nightmare.
Right. Instead, we use a simple binary choice with plain language. Do you want to add the official certification for $49? Yes or no? We use large distinct buttons for yes and no. The user must make an active, clear choice to proceed.
And by doing that, we aren't tricking them into a purchase, but we also aren't hiding the option from them. We are just making it a conscious decision.
Exactly. By forcing an active informed choice rather than relying on deception in pre-checked boxes, we completely eliminate post-choice regret and anxiety. The user feels in control, which is the ultimate overarching goal of reducing cognitive load.
They navigated the complexity, they evaluated the options, and they made the final choice on their terms. This has been a massive paradigm shift. Over the course of this deep dive, we've moved from treating users like physical reaction time machines who just press flashing lamps to respecting them as intelligent foragers navigating complex information landscapes.
The shift from Hicks flashing lamps in 1952 to Parole's scent tracking predator in the 1990s, and finally to modern cognitive load management is really the evolution of user experience design in a nutshell.
We stopped designing for reflexes and started designing for human comprehension.
Perfectly said.
To bring this all home for you, the listener, and to let you practice applying these concepts, we have a diagnostic exercise based on the notes. We call it the 19 field intake dual refactor exercise.
Here is the setup for the exercise. Imagine you have a standard, legally necessary 19 field intake form for this pathway portal.
Sounds thrilling.
Oh, it's standard stuff. It asks for standard data name, address, previous education history, career goals, data birth, etc. All 19 fields are strictly required by the system.
Right.
Your task is to sketch or wireframe this exact same form twice, optimizing for completely different cognitive loads based on context.
Okay. Challenge one. Design this form as a calm, one thing per page journey for a highly stressed, first-time mobile applicant who is anxious about returning to school. How do you break up those 19 fields? How do you label the steps without causing panic?
And challenge two. Design those exact same 19 fields as a dense single-screen keyboard workspace for an admissions coordinator who processes 80 of these applications a day. How do you visually group the data on one screen? How do you minimize their need to switch contexts or even touch the mouse?
It's a brilliant exercise because it proves that there is no single right or perfect UI layout. The interface must always bend to the context and cognitive state of the user.
This raises an important question. And if we connect this to the bigger picture of where digital design is heading, well, we know information sent guides users based on their prior knowledge. We know preference matchers navigate differently than preference constructors.
Sure.
Think about this for a moment. What if adaptive systems could subtly adjust the visual hierarchy of an interface based on the scent trail you've left behind, without hiding the other options from you?
Oh. Wow. Let me wrap my head around that. So if the systems analytics sees me lingering on data science articles, reading course descriptions, and behaving like a wolf tracking that specific scent, it doesn't delete the design courses because that would violate user control and trap me if I changed my mind.
Right. No hiding options.
But it subtly elevates the data science links in the visual hierarchy.
Precisely. It could make the typography slightly bolder, or move the relevant categories higher up in the mega menu. It strengthens the specific scent based on your unique foraging history, dynamically reducing your extraneous cognitive load in real time, while still providing the full landscape of choices.
That is a fascinating thread to pull on. The interface acting as a true partner in the foraging process. But that will have to wait, because next time in S11E02, we are moving from the mind back to the body. Now that we understand how the brain chooses and evaluates information, we are going to investigate what happens when the hand actually moves to click, tap, and steer. We'll be diving deep into Fitz's law and the steering law, examining the raw physical mechanics of targeting.
It's gonna be a great one.
I can't wait to get into the physics of it. But until then, the next time you are staring at a chaotic menu, or standing paralyzed in the toothpaste style, or trying to uncheck a sneaky add-on in a checkout form, remember those flashing lamps in 1952.
You aren't failing to react fast enough.
You're just looking for a scent in a world designed for reflexes. Take care.