What does an analytics stack actually weigh?
Measurement is not free. A standard GA4 installation pulls gtag.js at roughly 90-100KB compressed -- several hundred KB of JavaScript once the browser decompresses and parses it. Most sites load it through Google Tag Manager, which adds its own container script of a similar order. And in the EU the stack legally needs a consent management platform in front of it: another script, another network round-trip, plus the banner's own HTML, CSS and logic.
Add a Meta pixel or a heatmap tool and a quarter of a megabyte of third-party JavaScript per page is not an outlier -- it is the common case. On a fibre connection with a desktop CPU you barely notice. On a mid-range Android phone on mobile data -- which is what a large share of your real visitors use -- every one of those scripts competes with your actual content for bandwidth and processing time.
How does this touch Core Web Vitals?
Core Web Vitals are Google's three user-experience measurements, and they are part of its ranking systems:
- LCP (Largest Contentful Paint) -- how fast the main content appears. Third-party scripts compete for early bandwidth; a preloaded tag container is bandwidth your hero image did not get.
- INP (Interaction to Next Paint) -- how quickly the page responds to taps and clicks. INP replaced First Input Delay in 2024 and it is where heavy JavaScript hurts most: long script-evaluation tasks block the main thread, and a tap that lands during one waits.
- CLS (Cumulative Layout Shift) -- visual stability. Consent banners are notorious here: inject the banner after first paint and push the content down, and you have donated layout shift to every EU pageview.
No single script ruins these scores. The pattern is accumulation -- analytics plus tag manager plus consent plus pixel, each "only" a modest cost, summing to a page that feels heavy and measures poorly on the phones your customers actually own.
Does the speed cost show up in revenue?
The link between speed and outcomes is one of the most consistently replicated findings in web performance: study after study, from vendors and independents alike, shows conversion rates falling measurably as load time grows -- commonly cited in the range of several percent lost per extra second. Slow pages also bounce more and get crawled less enthusiastically. The irony writes itself: the measurement stack exists to optimise conversions, and its weight quietly taxes them.
There is also a purely practical cost: complexity. Every script in the chain -- container, consent platform, tags -- is something that can misfire, double-fire, or silently stop firing after a site update. Lighter stacks break less.
How light can analytics be?
The counting job itself -- pageview, referrer, page URL, rough geography, a goal event -- needs almost nothing. TrackTrendy's script is 1.4KB, loads with defer so it never blocks rendering, sets no cookies, and needs no consent banner or tag manager in front of it. Plausible and Fathom are in the same low-kilobyte class. That is a 98-99% weight reduction against a typical GA4-plus-GTM-plus-CMP stack -- not by clever engineering, but by not doing the profiling work that makes advertising analytics heavy.
Note what the banner removal does for Core Web Vitals specifically: no consent script, no injected banner, no banner-induced layout shift, and none of the "did they consent yet?" gating logic. For EU sites the CMP is regularly a bigger performance item than the analytics itself.
A sensible performance checklist for measurement
- Audit what actually fires. Open DevTools' Network tab, filter to third-party requests, and list what loads on your homepage. Most owners find a forgotten pixel or two.
- Remove what you do not read. A heatmap tool nobody has opened since last year is pure tax.
- Load what remains with defer, after your content.
- Prefer lightweight, cookieless measurement for the traffic-and-conversions job; keep heavy ad pixels only where you actively run the corresponding ads.
- Re-test your Core Web Vitals (PageSpeed Insights) after the cleanup -- the change is usually visible immediately.
Faster pages, complete visitor counts, and no banner: it is one of the few optimisations where the trade-offs all point the same way. For what the switch looks like in practice, see our comparison of Google Analytics alternatives.