Instant.Page makes pages feel faster by speculating on the next navigation: once it is reasonably confident a link is about to be clicked, it preloads the target. The trigger on desktop is either 65ms of hovering over a link or a mousedown; on mobile it's touchstart. By the time the actual click completes — likely a few hundred milliseconds later — the page is already on its way.
Measuring the benefit empirically
Integrating Instant.Page and counting it as a performance win is one matter; confirming it with data on your own site is another. Tim Kadlec ran exactly that A/B test, using Netlify's performance-neutral split testing to send half of visitors a build with Instant.Page and half a build without.
The critical part is attribution: the same halves were reported to SpeedCurve via the LUX API, so performance charts could be broken down by which variant a visitor received.
Why one site's results don't generalize
Kadlec reports that the outcome mostly looks favorable, but flags his own site as a weak sample:
It's also worth noting that even if the results do look good, just because it does or doesn't make an impact on my site doesn't mean it won't have a different impact elsewhere. My site has a short session length, typically, and very lightweight pages: putting this on a larger commercial site would inevitably yield much different results.
That caveat leaves the question open for heavier properties, where the same test would be worth repeating.



