Models, Effects and Rules of Thumb
Transcript
Made automatically from the video's sound, so it may contain mistakes. Click a line to play from there.
Welcome back to The Explainer. We've tackled visual hierarchy, we've unpacked full stack digital accessibility, and now, now we're diving into a truly rigorous, scientific approach to user experience.
We're going to answer that massive question. Why do some interfaces feel completely effortless, whilst others make you want to throw your laptop out the window? Our promise today is to help you spot the exact difference between a predictive mathematical model, a psychological tendency, and a simple rule of thumb.
Let's run through the roadmap quickly. We'll start with the vibe coder trap, move into the three tiers of UX laws, and then look at bridging the two goals. After that, we'll do a multimodal pathway audit and wrap things up with a practical four-lens UX audit.
Right, part one, the vibe coder trap. So here's a brilliant question for you. Why is it that these beautiful, minimalist, AI-generated interfaces systematically fail real users?
I mean, we see it all the time, don't we? Unguided UI generators spitting out these gorgeous screens that look flawless in a static mouse demo. They look fantastic on a portfolio, but when actual humans try to use them to get real work done, they completely fall apart.
Why? It really comes down to the massive clash between static screenshot minimalism and real-world expert workflows. You see, designers often strip away borders, labels, and persistent buttons just to make things look clean.
But ISO standards tell us everything depends on the context of use. Real users doing actual work desperately need persistent visual clues. When you strip away those signifiers, you're secretly exploding the cognitive effort required.
What looks simple is actually exhausting to use. Three tiers of UX laws. Now, it is absolutely crucial to understand the epistemic difference between these three tiers.
Tier 1 is literally predictive mathematical physics. Think Fitz's law or the GOMS model. These use bounded laboratory parameters to calculate exact human performance times. Then we have tier two, which is all about descriptive psychological tendencies, like gestalt grouping or Sweller's cognitive load theory.
These shift the probabilities of behavior, but they aren't fixed constants. And finally, tier three, folk heuristics. These are your context-dependent diagnostic rules of thumb, like Nielsen's 10 heuristics or wax standards.
They are brilliant checklists, absolutely, but they aren't literal laws of physics. Let's zoom in on a tier 1 model for a second. The keystroke level model, or KLM. This is amazing because we can actually quantify the time cost of every single physical and mental action a user takes.
Take the M operator, for example, which stands for mental preparation. It takes a human roughly 1.2 seconds. So when a vibecoder hides a button so it only appears when you hover over it, they are forcing the user to incur that 1.2 second mental tax every single time they have to guess where to click.
It is literally a hidden tax on their workflow. Alright, part three. Bridging the two gulfs. The first of these gaps is called the Gulf of Execution.
Basically, this is the user sitting there thinking, right, how do I act? Picture a frustrated learner trying to enroll in a course, but they can't figure out which borderless, ultra-minimalist card is actually a clickable button. They are stranded in the Gulf of Execution, because their mental model of what a button should look like just doesn't match the system in front of them.
The Gulf of Evaluation. This is the user anxiously asking, wait, what just happened? So imagine they actually manage to click the button, but nothing happens visually. No loading spinner, no success message.
Maybe the system just did a silent background save. Well, now they are trapped in the gulf of evaluation, totally unable to tell if their action worked, or if they need to frantically click the button five more times. Cognitive scientist Don Norman sums this up perfectly.
He says, Designers of the world forget affordances, provide signifiers. We need deliberate clues to bridge those two gulfs. You see, a real physical door has an affordance, a handle you can grab, but a flat glass screen, it doesn't have physical affordances.
It only has pixels. So we absolutely must design deliberate signifiers to signal functionality so users aren't left guessing. Let's put this into practice with part four, the multimodal pathway audit.
Starting with the desktop pointer modality. Hiding enrolment buttons behind a cursor hover is a direct violation of Jakob's law, which tells us that users spend most of their time on other websites and expect yours to work the exact same way.
If you force desktop users to hover over mystery cards just to reveal standard actions, you are ripping open that gulf of execution and completely destroying their ability to scan a catalogue quickly. Now, shift over to the mobile thumb modality, and it gets even worse.
Hover states literally do not exist on touchscrees. They are just dead ends. And those tiny aesthetic 18 by 18 pixel filter icons, absolutely impossible for a human thumb to hit reliably.
We fix this gulf of execution by ditching the tiny targets and using persistent, properly sized buttons that actually meet WCAG target size standards. For screen reader users, that evaluation gap is critical. Think about what happens with silent background saves and generic unnamed graphical tiles.
They completely blow open the gulf of evaluation. Visually impaired users are left with zero idea of what action they just took. But we can easily repair this by using explicit semantic HTML roles and polite status confirmations.
That way, the assistive technology actually speaks the enrolment confirmation aloud and the gap is bridged. Finally, section five, the four-lens UX audit. Right, here is your practical hands-on action plan to stress test any interface.
Step one, classify the UX law you're using. Is it a math model, a cognitive effect, or a heuristic? Step two, measure the gulf of execution. Basically, how much mental prep time are you forcing onto the user?
Step three, measure the gulf of evaluation. Is your feedback actually clear and immediate? And step four, stress test it across every modality. Pointer, thumb, keyboard navigation, 400% zoom, screen reader, and cognitive clarity.
Do this, and you guarantee your interface actually works for real human beings. I highly recommend keeping this four lens reference sheet right at your desk. But as we finish up today's explainer, I want to leave you with a bit of a provocative question.
If we can mathematically model exactly how users physically execute actions, how on earth do people mentally choose between complex options when they're faced with uncertainty? Keep that thought in your mind, because in our next explainer, we are going to unpack the fascinating psychology of user choice, cognitive load, and Hicks Law.
Thanks so much for joining and see you next time.