They matter, but not in the way the tooling implies. Core Web Vitals are a confirmed ranking input for organic results and a weak one — Google has consistently described them as a tiebreaker between comparable pages rather than a primary factor. For map pack position specifically, the evidence for direct influence is thin; proximity, relevance, and prominence dominate.
So the honest case for fixing them is not ranking. It is that a local service site is mostly read on a phone by someone who needs the service now, and a page that takes eight seconds to show content loses that person before your ranking ever gets a chance to matter.
Where the real cost sits
A homepage scoring poorly on mobile speed is measurably losing calls. That loss does not appear in a rankings report — it appears as traffic that arrives and leaves. In audits I run, slow mobile performance is one of the most common findings on local sites, and it is almost always caused by the same three things: unoptimized hero images, a page builder loading scripts it does not need, and third-party embeds firing before content renders.
What to actually fix, in order
Images first. Correctly sized, modern format, lazy-loaded below the fold. This alone resolves most local-business speed problems.
Then scripts. Chat widgets, tracking tags, and review carousels that block rendering. Defer what can be deferred and remove what nobody uses.
Then layout shift. Reserve space for images and embeds so content does not jump while loading. This is the cheapest of the three and the most annoying to users.
Test on a real phone on cellular data rather than trusting a lab score. And ignore the impulse to chase 100 — the difference between a 45 and an 80 is worth real money, while the difference between 90 and 100 usually is not.
The Core Web Vitals guide for small business covers the diagnosis and the fixes in detail.