How Fast Is Fast Enough? Setting Speed Targets by Page Type

Performance advice tends to imply a single answer to “how fast should my site be”: as fast as possible, ideally a perfect score. That sounds right and is quietly unhelpful, because it treats every page as if it has the same job. Chasing a flawless score on a deep documentation page nobody reaches from an ad is effort spent where it does not matter, while a paid landing page that is merely “okay” is bleeding money. The useful question is not “how fast is fast” in the abstract. It is “how fast does this specific page need to be, given what it is for.”

Here is how to set speed targets by page type instead of one impossible number for everything.

The Principle: Match the Target to the Stakes

A page’s speed target should be driven by two things: how much money rides on it, and how impatient its visitors are. A page that receives expensive, high-intent paid traffic from people who will bounce in seconds has the highest stakes and deserves the strictest target. A page reached by patient, already-committed users deep in your product has lower stakes and can carry a looser one. Same site, very different requirements. Once you sort your pages by stakes, the targets sort themselves.

The Tiers

Tier one: paid landing pages and key conversion pages. These are the strictest, no exceptions. You are paying for every visitor, they arrive with high intent and low patience, and every fraction of a second of delay is bounce probability you paid for. Hold these to the top standard: fast on a mid-range phone, the hero visible almost immediately, nothing that jumps or stalls. This is where a perfectionist target actually earns its cost, because the return is direct and measurable.

Tier two: your most-visited organic pages. The homepage, the top blog posts, the main product pages. High volume means small improvements compound across many visits, and organic visitors, while a little more patient than paid ones, still leave slow pages. These deserve a strong target, close to the tier-one bar, because the sheer traffic makes speed here worth real effort.

Tier three: deep and low-traffic pages. Documentation, archive pages, rarely-visited corners. These should not be slow, and the baseline hygiene still applies, but spending days squeezing the last hundred milliseconds out of a page a handful of committed users reach is effort better spent on tier one. Good enough is genuinely good enough here.

Why This Beats Chasing a Perfect Score Everywhere

The one-number approach fails in two directions at once. It lets you feel done when your average score looks fine, even though your money page might be the slow one dragging in the wrong direction invisibly. And it guilts you into over-investing in pages where speed barely moves anything. Sorting by page type fixes both: it points your effort at the pages where speed converts to revenue, and it gives you permission to stop optimising the pages where it does not. That is a more honest and more profitable use of a limited performance budget.

To set your own tiers, run your important pages through PageSpeed Insights, but read them through this lens: a mediocre score on a tier-one page is an emergency, while the same score on a tier-three page is fine. The number matters less than which page it belongs to.

The Takeaway

Fast enough is not a universal number, it is a judgement about what a page is for. Set your strictest targets where the money and the impatience are highest, ease off where the stakes genuinely drop, and stop treating a documentation page and a paid landing page as if they owe you the same performance. You will get better results from less total effort, because the effort will be aimed where speed actually turns into conversions.

If you want your pages tiered and given targets that match what each one is actually for, that is a conversation away.