TL;DR
- The popular explanation for why a slick website feels good — dopamine, reward circuits, mirror neurons — is not something we could trace to a source, so it is not the argument here.
- The measurable version is better: there are published thresholds for how quickly an interface has to answer before it stops feeling like it is working.
- Google publishes one of those thresholds as a ranking-relevant metric, which means you can check your own site against it this afternoon.
- Micro-interactions earn their place when they confirm something happened. When they are only motion, they cost you time you cannot afford.
There is a genre of web design writing that explains a good interface in terms of brain chemistry. Contrast triggers a reward response, animation confirms it, and the visitor is chemically nudged towards the button. It is a nice story. It is also the part of this article we have removed.
Why the dopamine framing is not here
We went looking for a primary source — a named study, a figure we could quote in the terms its authors used — for the claim that particular colour contrasts or hover animations produce a measurable neurochemical response in web users. We could not fetch one. What we found were articles citing articles, none of which ended at a paper.
That does not mean interfaces have no emotional effect. It means the neurochemical explanation is doing no work that the observable explanation cannot do better. And the observable version has numbers you can hold your own site against.
People do not stay because a page rewarded them. They stay because it answered them quickly enough that they never thought about it.
The three limits that have not moved in thirty years
The foundational work here is older than most of the sites it applies to. Writing in 1993, Jakob Nielsen set out three response-time limits: 0.1 seconds is the limit for a user to feel the system is reacting instantaneously; 1.0 second is the limit for their flow of thought to stay uninterrupted, even though they notice the delay; and 10 seconds is the limit for keeping their attention on the task at all. Beyond that, they go and do something else.
Those are not thresholds for page load. They are thresholds for every interaction: the menu opening, the filter applying, the form telling you the postcode is wrong. Most sites are judged on the second and third of those far more often than the first, and most site owners only ever measure the first.
What each limit asks of your design
- Feels instantAnything the visitor triggers with a finger — a tab, a toggle, a dropdown — should respond before they can register a wait. If it cannot, it needs a state change that can.
- Still in flowA search or a filter can take a beat, but the interface must acknowledge the input immediately. Silence is what breaks concentration, not duration.
- Attention goneAnything genuinely slow — a quote calculation, a large upload — needs honest progress, not a spinner. A spinner that could mean five seconds or fifty is worse than a number.
Google has put a number on “instant”
The modern equivalent is Interaction to Next Paint, which measures the delay between a user’s click, tap or key press and the next frame being drawn. Google’s web.dev documentation sets the thresholds plainly: an INP at or below 200 milliseconds is good responsiveness, above 200 and up to 500 milliseconds needs improvement, and anything above 500 milliseconds is poor. Crucially it looks at every interaction across a visit, not just the first one.
This matters for a practical reason. Most WordPress and ecommerce sites pass a load-speed check and then fail on interaction, because the weight is not in the page — it is in the six tracking scripts, the chat widget and the page builder’s runtime all competing for the main thread every time somebody taps something. That is invisible in a homepage speed test and obvious in field data.
What responsiveness is actually worth
The strongest money figure available comes from Milliseconds Make Millions, research commissioned by Google and carried out by 55 and Deloitte across 37 brands using 30 days of data from late 2019. A 0.1 second improvement in mobile site speed moved 3.2% more retail shoppers from the product listing to the product page and 9.1% more from the product page to the basket. Retail consumers spent 9.2% more, and travel bookings rose 10%.
Two caveats worth saying out loud, because the figure gets quoted without them. It is retail and travel data, and it is mobile. If you sell a considered B2B service, do not expect the same elasticity — the buying cycle absorbs a lot of friction that an impulse purchase will not. What does transfer is the direction and the surprising size of it: a tenth of a second is not a rounding error.
Where micro-interactions earn their place
This is where the dopamine framing does real damage, because it encourages motion for its own sake. The useful test is simpler: does this animation tell the user something they did not already know?
Feedback worth building
- A button that visibly changes state the moment it is pressed
- Inline validation that says what is wrong while the field still has focus
- A basket or enquiry count that updates without a page reload
- A loading state that shows how much is left
Motion for its own sake
- Elements that fade in as you scroll past them
- Hover effects that move layout underneath the cursor
- Carousels that advance while someone is reading
- Page transitions that add a delay to hide a delay
The left-hand column reduces uncertainty. The right-hand column adds JavaScript to the main thread, which is the same thread being measured by the responsiveness metric above. It is an unusually direct trade: decorative motion is bought with the budget you needed for feedback.
What to change first
In rough order of effort against payoff, and none of it requires a redesign:
None of that is glamorous, and none of it will be described by anyone as neuro-design. It is the difference between a site that feels considered and one that feels like it is thinking about it.
Good to know
FAQs
Colour affects legibility, contrast and where the eye lands, all of which are worth getting right. The stronger claims — that a particular hue triggers a measurable neurochemical response in shoppers — are not something we could trace to a primary source, so we would not build a strategy on them.
It is Google’s measure of how quickly a page responds to a click, tap or key press, across the whole visit rather than just the first interaction. Google publishes explicit thresholds for good and poor responsiveness, so it is one of the few design questions with a published pass mark.
Speed tests usually measure loading on a fast connection and a fast machine. Sluggishness after the page has loaded normally comes from scripts competing for the browser’s main thread every time somebody interacts. Field data from real visitors will show it where a lab test will not.
Not inherently, but they compete for the same budget as responsiveness. Animation that confirms something happened is worth the cost; animation that only fades content into view is usually paid for out of the time you needed to answer the user.
Fast enough that an interaction feels immediate rather than acknowledged. The published limits are strict — a fraction of a second for something to feel instantaneous — and they have not changed with hardware, because they describe people rather than machines.
Next step
Fast where it counts?
We build and rebuild websites for businesses across Brighton and London, and we measure the interactions, not just the homepage score.