Performance-Driven Design Strategies for Faster Sites
Website speed is not only a development problem. It is shaped long before anyone writes code, through the brief, the layout choices, the content you want on each page, and the way the site responds when people scroll, tap, and navigate. A heavy hero section, a complex grid, or a “just one more” animation can be the difference between a site that feels sharp and one that feels sluggish, even if the hosting is fine.
This article stays on the design side. No setup steps or tooling. Just the decisions that reduce page weight, avoid unnecessary complexity, and prevent layout shifting (when things jump around while the page loads). Speed is part of quality. It affects trust, enquiries, and search visibility, so it is worth planning for from the start rather than trying to patch later.
Speed starts in the brief, not in optimisation
Treat performance as a design constraint, the same way you treat accessibility and brand, and agree it early so the site does not inherit slow defaults.
Most slow sites are not “broken”. They are simply designed to carry too much, too soon. If the brief rewards visual impact on every screen, the web design will usually lean on big images, more sections, more components, and more motion. Then the build has to support it, even if it is not helping the business.
In practical terms, performance means three things a client can feel. First load on mobile should be quick, not a long pause on 4G. The layout should stay stable while it loads. Stable layout just means things do not jump around as images and fonts appear. And interactions should respond straight away, like tapping a menu or opening an accordion.
Projects tend to become slow before build starts in a few predictable ways:
- Overloaded pages. Homepages that try to do everything: full screen hero, video, logo walls, long feature lists, testimonials, maps, feeds, and a footer packed with extras. It reads like a pitch deck, and it loads like one too.
- Too many templates. A different layout for every service, every case study, and every campaign. Each “special” page adds design variation that usually adds more components and more assets to support it.
- Heavy visuals everywhere. Large background images behind every section, multiple decorative layers, and high resolution imagery used when a smaller, simpler approach would still look professional.
A useful performance target can be defined in plain language and written into the brief. Something like:
- Mobile first load feels fast. The page should become usable quickly on a typical phone connection, not after a long wait.
- No layout shifting. Headings, buttons, and key content should not move around as the page finishes loading.
- Interactions feel immediate. Menus, tabs, accordions, and forms should respond quickly, without lag or stutter.
That kind of target changes design decisions. It encourages a calmer layout, clearer content priority, and fewer “just in case” sections. It also helps when someone asks for a new carousel, a full width video header, or a complex animated intro. You can point back to the agreed aim and decide if it earns its place.
There are trade-offs. Keeping pages fast often means simplifying one of these areas:
- Hero sections. A single strong message, one image, and one call to action is usually enough. Layers of motion, multiple background assets, and oversized media tend to cost more than they return.
- Component variety. Reusing a small set of well designed website sections keeps the site consistent and lighter. Too many bespoke blocks increases complexity for both performance and maintenance.
- Always-on decoration. Subtle styling can carry a brand without relying on lots of large visuals. Use high impact imagery where it matters, not as wallpaper across the whole site.
My judgement call: if a page needs a long animated story to feel “premium”, it is usually masking unclear messaging. A simpler layout with better content almost always performs better, loads faster, and converts more cleanly.
Layout choices that keep pages light and stable
Good layout reduces what has to load, and it makes the page easier for the browser to keep steady and responsive.
Design has a direct impact on speed because layout decisions control how much content appears, how many components are on the page, and how much reflow happens as things arrive. Reflow is when the page has to recalculate positions while loading. That is the moment users experience as “jumping” or “wobbling”.
A common cause of slow pages is simply too many sections. Not because each section is huge on its own, but because the page repeats the same patterns over and over: headline, text, icons, image, button, then again, then again. It becomes a long scroll that looks busy and loads like a long scroll.
In most service sites, fewer, stronger sections do better. Pick a clear hierarchy. Lead with the main promise, show proof, explain the service, and make the next step obvious. If a section does not add a new decision or remove a doubt, it is usually padding.
Complex layouts can also add weight without adding value. Grid-on-grid designs, deep column nesting, and lots of “containers inside containers” tend to create more elements for the browser to handle. They also make responsive behaviour harder to predict, which is when spacing issues and odd stacking start appearing on mobile.
Simple blocks often do the job. A clean two-column section with a clear heading and one supporting visual is easier to load and easier to read than a layered layout with multiple alignment rules. You can still make it feel designed through spacing, typographic hierarchy, and a consistent rhythm.
Above the fold matters most. It is what people see before scrolling. If that area relies on a large image or video to “work”, the page will feel slow even if everything else is sensible. Aim for meaningful content that stands on its own: a strong headline, a short supporting line, and a clear primary action. Then treat imagery as support, not the thing holding the layout together.
Stability is a design responsibility too. Prevent layout shift by reserving space for anything that loads after the initial content. That includes images, embedded maps, video players, testimonial sliders, and even short bits of dynamic UI like accordions. If the design assumes “it will fit somewhere”, it often ends up pushing text down as it appears.
A good rule of thumb is to design with placeholders. Decide what shape and size an image or embed will be in the layout, and keep that space consistent across breakpoints. That way the page can settle early, instead of constantly adjusting as assets arrive.
Sticky elements deserve scrutiny. A sticky header can be helpful, but it is easy to overdo it, especially on mobile where it steals valuable screen space. Add an announcement bar, then a cookie banner, then a chat widget, and suddenly the user is looking at a narrow strip of actual page.
It can also create interaction lag if multiple sticky layers fight for attention and reposition as the page loads. My judgement call: if you need a sticky header, keep it compact, and avoid stacking it with other persistent bars unless there is a clear business reason. Most “nice to have” banners are not worth the clutter.
Images and media: the biggest website design lever for speed
Most slow pages are weighed down by visuals that were added for atmosphere, not for the decision the page needs someone to make.
On most business sites, the biggest performance wins come from media choices. Not because images are “bad”, but because they are usually the heaviest assets on the page. If the design assumes lots of large visuals, the site will feel slow even when the build is solid.
A useful way to think about it is purpose. Every image should either explain something, prove something, or help someone choose the next step. If it does none of those, it is decoration. Decoration is not forbidden, but it should be rare and lightweight.
Choosing the right number of images per page is a design decision, not a content afterthought. A service page might only need one strong visual and one supporting proof image (for example, a real project photo, a screenshot, or a founder portrait). A case study may need more because the visuals are part of the evidence. A homepage often sits in the middle. It needs enough to build trust, but not so much that it turns into a gallery.
There are plenty of times when a single strong image beats multiple supporting images. If your message is clear, one image that matches it will land better than five weaker ones competing for attention. It also keeps the page calmer. It reduces cognitive load as well as load time, which is a rare double win.
My judgement call: if you are tempted to add a row of three stock photos to “break up the text”, do the opposite. Use spacing, a subheading, or a short proof point instead. You will usually end up with a cleaner section and a faster page.
Hero sections are where media choices do the most damage. Full-screen background images and videos look dramatic, but they often slow down the first impression. The user has not scrolled yet, so they feel every delay.
There are good alternatives that still feel designed. Use a strong headline, a short supporting line, and one clear action. Add a smaller, well chosen image beside the text rather than behind it. If you want texture, consider a simple colour block, a subtle gradient, or a lightweight pattern that is built into the design system rather than delivered as a huge background asset.
Video needs the same discipline. Autoplay video in the hero is nearly always a performance and attention problem at the same time. If the video is genuinely valuable, show a thumbnail with a clear play action and let the user choose. You still get the storytelling, but you do not force the download and rendering cost on every visitor.
If you do include video, think about what it replaces. A short explainer can replace a wall of text for some audiences. A looping background clip rarely replaces anything, it just adds weight. That is usually not a good trade on a service site.
Designing image ratios and crops early makes everything easier and faster. Decide the shapes you will use across the site: hero image ratio, card image ratio, case study gallery ratio, team portraits, and so on. Ratio just means the shape relationship, like 16:9 or 4:3. When you plan this upfront, you can export assets consistently, keep layouts stable, and avoid “one off” images that force awkward cropping or oversized files.
It also helps content. If your blog list uses the same crop every time, you can choose images that work in that frame. You avoid the common problem where half your thumbnails are too busy, too zoomed, or cut off at the wrong place.
Illustrations and icons can be a smart performance choice when they replace heavy photography. Simple vectors (clean shapes, limited detail) tend to scale well and keep pages feeling sharp on high resolution screens. They also give you a consistent look without relying on huge images.
But they can become heavy too. Complex vector illustrations with lots of paths, filters, or animated layers can be surprisingly expensive to render, especially on mobile. The same goes for icon libraries where you load hundreds of icons to use eight. Keep icon sets tight. Keep illustration style simple. If an illustration is basically a detailed poster, it is not “lightweight” just because it is vector.
Fast sites are designed with restraint. The goal is not to remove visuals, it is to make every visual earn its place. When you treat imagery as part of the message, not as wallpaper, speed improves almost automatically.
Typography and fonts without the performance penalty
Font choices shape how fast a page feels, so aim for brand clarity and mobile readability without hauling in a suitcase of type files.
Fonts are design assets. They have a “weight” in the literal sense too, because each font file has to be downloaded before the page can fully settle into its final look. Even when the page is technically usable, swapping to the final font can make it feel slow or visually unstable.
The simplest win is to use fewer font families and fewer weights. One family that works for headings and body copy is often enough for a service site. If you really need contrast, pair two families at most. Beyond that, you are usually designing a mood board, not a website.
Weights matter as much as families. Each weight and style is typically a separate file, like Regular, Medium, Bold, and Italic. If you choose one family with four weights plus italics, you have quietly asked the browser to juggle a lot before the page looks “right”.
A common trap is trying to match print typography. Print has no loading time, so it can indulge in hairline weights, multiple cuts, and niche display fonts. On the web, especially on mobile, those choices often backfire. Thin fonts can look weak or blurry, and you end up adding more variants just to make the type feel usable.
My judgement call: if the brand font only looks good at 48px+ and falls apart for body copy, do not force it. Use it for headings only, and choose a simpler workhorse font for the reading. Your content will feel more confident, not less “branded”.
Fallback fonts are part of good web design, not an afterthought. A fallback font is the font your site shows if the main one is not available yet, or cannot be loaded for some reason. Pick fallbacks that match the general proportions of your main typeface: similar x-height (how tall lowercase letters feel), similar width, and a similar overall tone (modern, traditional, friendly, formal).
If your brand font is a clean sans serif, choose a clean system sans serif fallback. If it is a more classic serif, choose a system serif fallback. You are aiming for “close enough that nobody notices”, not a perfect match.
Type scale is where performance and readability meet. A type scale is the set of sizes you use for headings, subheadings, body, captions, and so on. If your scale is sensible, you do not need a different font weight for every step. You can often get a strong hierarchy with just Regular for body, Semi-bold for headings, and maybe one accent weight for large headings.
Keep body text comfortable on mobile. That means prioritising clear letter shapes, decent spacing, and enough contrast. If you need to rely on an ultra-light weight or a fancy condensed cut to “make it fit”, it is usually a sign the layout is too tight or the typography choices are fighting readability.
Variable fonts can help, but they are not a magic answer. A variable font is a single font file that can produce many weights and widths. Used well, it can replace several separate files and give you flexibility without a big font menu.
They add complexity when you use that flexibility as an excuse to create lots of in-between styles. If every section has its own bespoke weight and width, you are back to a design that is hard to keep consistent and can feel visually “busy”. Variable fonts are best when you still act with restraint, just with slightly more control.
In practice, the performance-minded approach is boring in a good way. Fewer families. Fewer weights. A fallback you can live with. A type scale that reads well without extra variants. When typography is planned that way from the start, the site feels faster because it becomes stable sooner, and users can read it comfortably from the first moment.
UI components and interactions: design for responsiveness
Interactive features can make a site feel polished, but every moving part has a cost in loading, stability, and attention
Most slow-feeling websites are not slow because of one “big” thing. They feel slow because lots of small interactive choices stack up. A sticky header here, an animated reveal there, a chat popup that loads late, a map that tries to render before the user even scrolls. None of these are evil on their own. But design has to choose what earns its place.
When people say a site feels “responsive”, they often mean two things. It adapts well to screen size, and it reacts quickly to user input. The second part is where interaction design can either help, or quietly get in the way.
Navigation first, not theatre
Clear navigation beats clever navigation. Especially for service sites where visitors are trying to answer simple questions: what you do, who you do it for, where you operate, and how to get in touch.
Mega menus can work for large, content-heavy sites. For most small businesses they add complexity without adding clarity. They take longer to scan, they are harder to use on touch devices, and they often turn “choice” into “work”. A simpler menu with fewer top-level items usually gets people where they need to go faster.
The same goes for animated headers. Shrinking, fading, sliding, swapping logos on scroll can look tidy in a design file. In real use it can feel jumpy, and it can cause layout shifts. Layout shift is when elements move after the page starts rendering, which makes the site feel unstable.
If I have to choose, I would rather ship a plain header that stays put and stays readable than a clever one that changes state constantly. Calm is a feature.
Motion with a job to do
Motion is best when it explains something. A button press state. A menu opening so users understand where it came from. A small transition that confirms an action worked. That is motion as feedback.
Motion can also support hierarchy. For example, a subtle reveal that guides the eye down a page. But “subtle” is doing a lot of work there. If everything animates, nothing feels important. And constant animation makes a site feel heavier even if it technically loads quickly.
Decorative animation is the first thing I cut when performance matters. Not because it is wrong, but because it rarely helps the user finish a task. If it does not improve understanding or confidence, it is probably just noise.
Carousels, parallax, background video: be selective
Some design patterns are popular because they look like “more website”. In practice, they often add weight and distraction.
Carousels are a common example. They hide content behind controls and timing. Users might not wait, and they might not notice there is more than one slide. If the information matters, show it. If it does not matter, do you need it at all?
Parallax effects can be nice in small doses, but they are easy to overuse. They also tend to make scrolling feel less direct, especially on lower-powered mobiles. If your message relies on parallax to feel “premium”, the underlying layout probably needs more attention.
Background video is the biggest judgement call. It can make sense for certain brands, like a venue or a product with strong visual context. Even then, it is often better as a simple, well-shot still image with an optional play button. People came for information, not a screensaver.
Form design: fewer fields, clearer states
Forms are where “interaction” meets revenue. They should feel straightforward. The fastest form is the one people actually complete.
Start by asking what you truly need at first contact. Name, email, and a message is often enough. If you need qualification, try one or two well-chosen questions, not a mini application form.
Validation should be clear and calm. Validation is the guidance that tells users what is missing or formatted wrong. Make it obvious where the issue is and how to fix it, without flashing banners or complex interactive widgets that take over the experience.
Be wary of fancy inputs that look modern but slow people down. Date pickers, multi-step wizards, and clever dropdowns can be useful in the right context. They can also introduce friction when a simple text field or a small set of radio buttons would be faster and clearer.
Maps, chat widgets, booking tools: on-page or linked out?
Third-party tools are not automatically bad. They can save time and add real value. The decision is about placement and timing, not purity.
Maps are a good example. If your business depends on footfall, a map is useful. But many sites embed a heavy interactive map on every contact page when a clear address, opening times, and a simple “Get directions” link would do the job. If users only need the map when they choose to navigate, let them click out to it.
Chat widgets can help when you have someone available to respond, and when the questions are genuinely time-sensitive. If it is mostly used as another contact form, it can be redundant. Consider whether a clear call to action and a fast form would serve better.
Booking tools are often worth it, but they do not always need to load on every page. If the booking experience is complex, it can be cleaner to send users to a dedicated booking page, or even a trusted external booking link, rather than making your whole site carry that weight. The goal is a confident path to conversion, not every feature everywhere.
The inside truth is simple: fast sites are designed to be fast. Interaction design is part of that. Choose fewer components, make each one earn its place, and prioritise clarity over novelty. The site will feel more professional because it will feel more in control.
Designing for content reality, not mock-up perfection
A fast design should still work when real people add real content in WordPress over time.
Most slow sites are not slowed down by one big mistake. They get slower as the design breaks under real content. Longer page titles. New service lines. Extra testimonials. A few more images per case study. Then someone adds another plugin to “fix” the layout and you are carrying more weight forever.
The fix is not to design for a perfect mock-up. It is to design for the messy reality of a business that grows. That is a performance decision as much as a visual one, because a design that holds up does not need extra widgets, heavy page builders, or one-off templates to keep it looking tidy.
Plan for content growth from day one
When I review designs for performance risk, I look for the pressure points. These are the places that will break first when content changes.
- Long titles and headings – especially on mobile. If your heading style only works for two short words, it is fragile.
- Varied image sizes – clients will upload photos from phones, suppliers, and old brochures. Design the layout to cope with different crops and orientations.
- Extra sections – a new trust badge strip, a new FAQ, a new service. The page should be able to add one more block without collapsing into visual noise.
This is where mild restraint helps. If the design relies on lots of overlapping effects, tight spacing, or “perfect” image crops, it usually becomes expensive to maintain. It also pushes people towards heavier tools to keep the look intact.
Use modular pages, with limits and guidance
WordPress is built for editing. People will add content. The goal is to make that safe and predictable without making the site feel locked down.
A modular page design means pages are made from reusable sections. In WordPress this often maps to blocks or block patterns – pre-built layouts you can insert and edit.
What matters is the rulebook that comes with the modules. Not a long PDF. Just a few clear, practical constraints.
- Limit the number of section types per page. For example: hero, proof, services, process, FAQ, contact. If someone needs seven different “feature grids”, the design system is already losing control.
- Give sensible max lengths for headings and intro text. Not to police wording, but to stop layout strain and reduce the temptation to add workarounds.
- Be clear about what not to add, like stacking three full-width image banners, or dropping in multiple third-party embeds on a single page.
- Provide an “escape hatch” module for unusual cases, such as a simple two-column content block, so people are not forced into custom layouts.
The judgement call here is to prioritise the modules that get used repeatedly. I would rather have six well-designed, well-tested sections than twenty “options” that create inconsistent pages and performance overhead.
Avoid template sprawl
Template sprawl is when every page becomes its own special case. A different layout for each service. A different header for each campaign. A different content block for each new idea. It looks flexible at first. It becomes slow and brittle later.
Fewer page types, with clear rules, usually performs better and stays cleaner. You want a small set of templates that cover most needs: a standard page, a service page, a case study, and a post. Then you use modules to vary the content within those rules.
This also improves editing speed. People know where things go. They do not need to guess which of eight similar templates to choose, or rebuild a page from scratch because the “right one” is missing.
Design components that degrade gracefully
Real content is incomplete sometimes. A team member does not have a headshot yet. A case study is missing a result metric. A testimonial is still being approved. If the design collapses when one field is empty, it creates pressure to patch it with extra plugins, custom code, or awkward filler content.
Graceful degradation means a component still looks intentional when optional content is missing. It might shorten, simplify, or fall back to a default, but it does not break.
- Cards and grids should cope with different text lengths without forcing fixed heights that cut content off.
- Image areas should have a sensible fallback, like a solid colour or a subtle pattern, rather than leaving a broken gap.
- CTAs should still work if there is one button instead of two.
- Stats and results should have an alternate layout for when there are fewer numbers, or none at all.
None of this is about blaming people for “wrong” content. It is about designing a site that can be edited confidently without dragging performance down over time. If the structure is sound, the site stays fast because you are not constantly adding weight to compensate for a layout that only ever worked in the mock-up.
Performance-friendly SEO and trust signals (without clutter)
Build search visibility and credibility through clear structure, sensible proof, and links that feel part of the page.
A lot of “SEO design” advice adds more stuff. More boxes, more badges, more widgets, more movement. That usually makes pages heavier, harder to scan, and less trustworthy. The better approach is quieter. Make the meaning obvious. Make the page easy to follow. Then add proof where it supports the decision.
Start with a clear page hierarchy. In plain terms, that means the page has an order people can understand at a glance. A single main message at the top, then a few sections that answer the next questions. When the structure is clean, you need less filler text, fewer repeated CTAs, and fewer “just in case” panels.
A practical pattern that works well is: what you do, who it is for, why it matters, proof, process, FAQs, then contact. Not because it is magic, but because it matches how most people scan. It also reduces the temptation to create extra sections that exist only to pad the page.
Use proof elements in a controlled way. Reviews, accreditations, case studies, awards, logos. They help, but only when they are curated and placed with intent. Too many signals at once can look desperate, and it adds weight through extra images, scripts, and embeds.
My judgement call: pick one primary proof type per page, then support it with a secondary one. For example, on a service page you might lead with two or three short testimonials, then link to one relevant case study. On a case study page, the case study itself is the proof, so you do not need a carousel of reviews on top.
Avoid heavy pop-ups and intrusive banners. Yes, they can drive clicks. They also interrupt reading, slow down interaction, and make a site feel lower quality. If you need a prompt, design it as part of the page. A simple in-content callout will often do the job with less friction.
This matters for perceived performance too. A page can load quickly but still feel clunky if things keep sliding in, shifting the layout, or blocking the content. Perceived quality is part of trust, and trust affects whether people stay long enough to take action.
Internal linking is another place where design affects both speed and SEO. Internal links are simply links between your own pages. They help people find related information, and they help search engines understand the site structure. You do not need a plugin widget to do this well.
Design internal links into the content. Add a “Related services” row with 3 to 5 links at the end of a service page. Add contextual links inside FAQs where it makes sense. Add a short “Next step” section under case studies that points to the relevant service and contact page. Keep it consistent, and it becomes part of the system rather than an add-on.
Calls to action should work without extra widgets, sticky bars, or chat tools. A clear primary button near the top, then one repeat near the bottom is usually enough for a service business. If you are adding more CTAs because people are not converting, it is often a clarity issue, not a button placement issue.
None of this guarantees rankings. SEO depends on many factors, including what you offer, how competitive the search is, and how well your content matches intent. But a clean hierarchy, controlled proof, and simple linking creates a site that is easier to understand, easier to maintain, and less likely to get weighed down by “optimisations” that backfire.
How to run a design process that protects performance
Make speed a design requirement early, then review it like any other sign-off.
Most slow sites do not happen because someone wrote “bad code”. They happen because the design kept expanding until the build had to carry too much weight. That is why performance needs a seat at the table from the first layout decisions, not as a last-minute rescue job.
A practical process is simple: decide what matters, design the core templates, review against a short checklist, then only add extras when you can explain the trade-off. That is not restrictive. It is how you keep control.
Start by agreeing a media budget in principle. Not numbers and megabytes. Just boundaries everyone can understand. For example: “Most pages rely on strong photography and short copy. Video is used only on one or two key pages, and only when it adds real explanation.” Or: “Case studies will have a small gallery, but we are not building photo-heavy stories for every project.”
This one decision prevents a lot of late surprises. If someone expects a full-width hero video on every service page, you want to find that out during design, not after the content team has sourced ten different clips.
Next, design key templates first. Home, service, case study, contact. These are the pages that set the rules for everything else. When those templates are solid, most future pages become “instances” of the system, not one-off inventions that need new sections, new scripts, and new layout tricks.
My judgement call: if a new layout element cannot be used in at least two places, treat it with suspicion. One-off exceptions are where performance and consistency usually get chipped away.
During design review, use a simple performance checklist. It is not about policing creativity. It is about spotting expensive decisions while they are still easy to change.
- Layout – Is the page built from a small set of repeatable sections, or is every screen a unique composition?
- Media – Are images doing clear work, or are they filler? Is video used only where it explains something better than text?
- Fonts – How many font families and weights are we using? Each extra variant adds load and complexity.
- Interactions – Do animations support understanding, or are they decoration? Avoid movement that shifts content after it loads.
- Third-party embeds – Are we embedding external widgets that bring their own scripts? An embed is any content pulled from another service, like a booking tool, map, or social feed.
Make this checklist part of the sign-off, like approving brand colours or copy tone. If a page fails the checklist, it does not mean “no”. It means you either simplify the design or you accept the compromise knowingly.
Stakeholder requests are the other common cause of bloat. The request is often reasonable in isolation: “Can we add a chat widget?”, “Can we show our Instagram feed?”, “Can we add a scrolling logo strip?”, “Can we embed the whole PDF?” The problem is that these things stack up.
A good way to handle it is to ask two questions in plain terms:
- What problem does this solve for the user? If it is not a user problem, it is usually a preference.
- What is the lightest way to achieve the same outcome? There is often a simpler alternative that reads better and loads faster.
Then offer alternatives. Instead of an Instagram embed, use a small curated grid of images that links out. Instead of a heavyweight pop-up, use an in-content callout. Instead of a map embed on every page, use an address and a link to open directions. These options keep the page stable and still meet the business need.
Finally, write down what you are saying “no” to, for now. That sounds formal, but it is practical. It stops the same feature coming back every meeting, and it keeps the team aligned on what the site is meant to be: clear, fast, and easy to use.
FAQ
Words from the web design experts
We often see fast sites being slowed down by design that grows without a plan. A common problem is treating the layout as decoration first, then trying to make it quick later. One thing we do early is check load behaviour on a real phone, because it shows you immediately which sections feel heavy, jumpy, or simply unnecessary.
If a design choice needs lots of assets or constant motion to feel “finished”, it is usually the wrong trade for a performance-driven site. Keep the structure doing the work, keep the visuals disciplined, and be willing to remove anything that does not help someone understand, decide, or act.
You might like these too
These sit in the same category as the one you are reading. They follow the same thread and offer a bit more depth. Have a look if you want to go further.