The Cartographer's Compass: On the Illusion of North in Cumulative Layout Shift
We have been handed a compass. It’s called Cumulative Layout Shift, and it points unerringly toward a singular north: stability. The needle trembles, and we are told, with the certainty of received wisdom, that our goal is to make it stop. To build a page so still, so pre-ordained in its dimensions, that it could exist behind glass in a museum. This is the new orthodoxy of perceived performance, and I fear we are mistaking the tool for the destination.
CLS is a magnificent metric. It gave a name and a number to a deeply frustrating user experience—the jank, the jump, the accidental click on the wrong button. It gave us a common language to fight against the chaos of late-loading ads and unsized images. But in our fervor to chase a perfect, near-zero score, we are in danger of building maps instead of worlds. We are prioritizing the stability of the chart over the richness of the territory it represents.
Consider an interactive data visualization. A user adjusts a slider, and a complex chart animates, smoothly morphing from one state to another. To the CLS algorithm, this is movement. This is shift. It is, by its very definition, instability. But to the human using it, this is clarity. This is understanding. It is a guided narrative told through motion. By treating all movement as a failure, we risk demonizing the very tools that make the web feel alive and responsive.
Our quest for a perfect CLS score can lead us down a path of artificial constraints. We pre-allocate immense containers for the worst-case scenario, leaving acres of whitespace for a future that may never come. We sacrifice the elegance of progressive enhancement for the brute-force guarantee of pre-defined aspect ratios. We build layouts that are stable not because they are resilient, but because they are inert. We have traded the messy, beautiful process of loading for the sterile perfection of a already-loaded state.
The compass was never meant to tell us where we cannot go. It was meant to help us navigate. A shifting layout is bad when it’s unexpected and disruptive—a paragraph leaping away as an ad injects itself below. But a shifting layout is good when it is intentional and meaningful—a comment thread expanding, a modal dialog centering itself, a chart revealing its depth. The true craft lies not in eliminating movement, but in mastering its narrative. We must learn to distinguish between the earthquake and the dance.
Let us not be cartographers who only draw still lakes and frozen mountains. Let us be guides who can also map the flowing river and the shifting sand dune, understanding their motion as part of their essence. Our goal should not be a page that never moves. It should be a page that moves with purpose.
Notes & further reading
A few pages I came back to while writing this: