Performance budget vs. design ambition — how we decide
Every Mentes project has a performance budget and a design ambition. When they conflict, here's the honest framework we use to decide.
Performance budget vs. design ambition — how we decide
The website performance vs design tradeoff is real. Anyone who tells you otherwise is selling you something — either a beautiful site that crawls on mobile, or a fast site that no one remembers.
We hold both at once. That's the position. And it creates genuine tension on every project we take on.
Here's how that tension actually gets resolved — not in theory, but in the room, with clients, under deadline.
The budget is non-negotiable. The ambition is not optional.
Every Mentes project opens with two documents on the table: a design ambition brief and a performance budget.
The performance budget is a hard ceiling, not a soft guideline. We target LCP under 2 seconds, CLS under 0.1, and INP under 200 milliseconds. These aren't arbitrary numbers. They reflect how real visitors — on real devices, on real networks — actually experience a site. Google's Core Web Vitals research has shown consistently that sites crossing these thresholds lose visitors before a single word is read. A slow site is not a premium site. Full stop.
The design ambition brief is equally non-negotiable in the other direction. Mentes doesn't do templates. We don't do drag-and-drop. Every project is built from first principles — concept, copy, design, code, and the growth engine underneath. If the ambition brief is vague, we push back before a single frame is designed. A weak creative brief is just a path to a mediocre site that also happens to be fast.
So when these two documents conflict — and they do, regularly — we don't pick a winner. We find a resolution.
Four decisions we've actually had to make
1. The heavy hero treatment on a hospitality brand
A restaurant group came to us with a clear vision: an immersive full-screen hero that filled the first scroll with motion, atmosphere, and texture. On a fast connection with a flagship device, it was exactly right — commanding, distinctive, impossible to scroll past.
On a mid-tier Android phone on 4G, the LCP was sitting above 4 seconds. That's not a rounding error. That's the bounce window.
The resolution wasn't to strip the treatment. It was to separate the experience into layers. Visitors on fast connections get the full visual. Visitors below a network threshold get a still frame — equally considered, equally branded — that loads in under a second. The hero feels the same from the outside. The architecture underneath it is adaptive.
The client never had to choose between fast and beautiful. We made that choice invisible.
2. The immersive intro that was blowing the budget on 3G
A brand founder wanted an immersive intro sequence — the kind of entrance that signals "this is not a template." The sequence was stunning in staging. It was also, on a 3G connection, loading for 7 seconds before a single brand word appeared.
This one required a harder conversation. An intro that asks a visitor to wait 7 seconds is not confidence — it's friction. We walked the client through real-world 3G data for their target market. Once they saw the numbers, the question shifted from "how do we keep the intro" to "how do we keep the feeling of the intro."
We rebuilt it. The new version arrives in under 2 seconds. The brand impression lands just as deliberately. The sequence is shorter, tighter, and — this is the part clients don't always expect — more impactful because it doesn't overstay.
Restraint is not the enemy of drama. Sometimes it's the source of it.
3. Typography at scale vs. layout shift
A professional services firm wanted a homepage dominated by large, expressive type — the kind of typographic confidence that reads as authority. The challenge was CLS: large type that shifts during load is jarring, and it tanks the score.
Most agencies would compromise on the type. We compromised on the loading sequence instead.
Font preloading, careful fallback sizing, and a considered placeholder state meant the layout never visibly shifted. The type arrived already in place — not after it. The CLS score stayed under 0.05. The typography stayed unreduced.
This one required more engineering time than the client saw. That's the right outcome. The craft should be invisible; the result shouldn't be.
4. The image-heavy portfolio page
A creative agency client needed a work page that showcased dozens of projects — high-resolution, edge-to-edge, each one a full visual statement. Naively loaded, that page would be an LCP disaster and a bandwidth nightmare on mobile.
The solution was progressive disclosure: projects load as the visitor reaches them, not all at once. Above-the-fold loads fast. Everything below it loads just in time, without layout jumps. The INP stayed under 150ms throughout — even with dozens of assets in flight.
The client asked us, after launch, why the page felt so smooth. We told them it was the result of decisions made before they saw the first design mockup. Which is true.
The two bad defaults — and why we reject both
There's a common failure mode on the "design over everything" side: agencies that treat performance as someone else's problem. They hand off a stunning Figma file and leave the load times to the developer, who inherits constraints they had no hand in setting. The result is a beautiful site that fails on the device their client's actual customers use.
There's an equally common failure mode on the "performance over everything" side: developers who treat any visual ambition as risk and quietly sand it down until the site is fast and forgettable. These sites score 100 on Lighthouse. They also score 0 on recall.
Neither is acceptable to us.
The honest answer is that performance and design are not enemies — they're collaborators under pressure. The pressure is what produces good decisions. A project with no performance constraints will always drift toward excess. A project with no design ambition will always drift toward utility. You need both pulling in opposite directions.
That tension is not a problem to be solved. It's a process to be managed.
How we surface the tradeoff early
The worst version of this conversation happens in the final week before launch, when something is either too slow or too stripped down and there's no room to maneuver. We've learned to surface the tradeoff early — during concept, not during QA.
When we're designing a treatment that could affect performance, we say so. We bring rough load estimates into the design conversation. We ask the client: who is your visitor, on what device, on what network? A luxury real estate site targeting Miami Brickell buyers on desktop can hold more visual weight than a service business targeting working families in central Florida on mobile.
The answer to "what's the right performance budget" depends on who's actually loading the page. Not on a spec sheet.
We do this because we believe clients deserve to make informed decisions — not because we think they want to hear about milliseconds. The framing is always in outcomes: this treatment adds roughly one second of load time for a third of your visitors. Here's what we can do to reduce that without changing what you see.
That's the conversation. It's not dramatic. It's just honest.
Both, in tension, on every project
There's no formula that resolves the website performance vs design tradeoff once and for all. Every project is different. Every client's audience is different. Every creative treatment has its own cost profile.
What we can offer is a consistent commitment: we will never ship something slow and call it premium, and we will never ship something fast and call it finished if it doesn't hold a visitor's attention past the first scroll.
Both standards are enforced on every project. Not alternately — simultaneously.
That's the Mentes position. It's harder than picking a side. It's also the only one that actually works.
If you're weighing a site rebuild and wondering whether you'll have to choose between a site that looks like your brand and one that performs — you shouldn't have to. See how we approach it, or look at the work and decide for yourself. When you're ready to talk, we're here.