sonex vs Vercel Web Analytics

Vercel Web Analytics is the easiest analytics to turn on if you already deploy there, ships one of the smallest scripts we measured, and is properly cookieless. What it is not is a full analytics product, and its data is tied to the platform you host on. sonex is a standalone tool that reads revenue and keeps history far longer.

Checked against Vercel’s own documentation on 2026-08-17, plus a direct measurement of both scripts.

At a glance

	sonex	Vercel Web Analytics
Free allowance	2,000 events/mo	50,000 events/mo on Hobby
Paid model	$20/mo for 200k events	$0.03 per 1,000 events on Pro
Add-on cost	None	Plus is +$10/mo per team
Data retention	6 months free, 1 year Pro, 5 years Business	1 month Hobby, 12 months Pro, 24 months Plus
Custom events	All plans	Pro and above
Cookies	None	None
Consent banner	Not needed	Not required; no explicit legal claim published
Tracker size (brotli, measured)	2,968 bytes	1,125 bytes
Revenue from Stripe or Polar	Yes, all plans	Not documented
Funnels, retention, attribution	Yes	Not documented
Session-level views	Yes	Not documented
Works off-platform	Anywhere	Built for Vercel deployments

Where Vercel wins

The script. 1,125 bytes brotli — second smallest of the thirteen trackers we measured, and less than half of ours.

The free allowance. Fifty thousand events a month on Hobby against our two thousand. For a side project that is a meaningful difference, and it is the honest reason a lot of people never look further.

Zero setup. One package, one component, done. It is part of the platform, and platform-native tools are hard to beat on friction.

Genuinely cookieless. Anonymised data, a daily-rotating request hash, no cookies. The same structural approach as ours.

Where sonex wins

Retention. One month on Hobby is the number to sit with. Any question of the form “how did this compare to last quarter” is unanswerable on the free tier, and answering it costs $10 a month per team on top of Pro to reach 24 months. Our free plan keeps six months and Business keeps five years.

Custom events on every plan. Vercel gates custom events to Pro. If you want to count signups on a hobby project, you cannot.

Revenue. Not part of Vercel’s product. It is ours, on every plan, read from Stripe or Polar.

Depth. Funnels, retention cohorts, attribution models, journeys, session-level browsing, compare mode, city-level geography, shared dashboards, a query API.

Portability. This is the strategic one. Vercel Web Analytics is designed around Vercel deployments, and your history lives with your hosting provider. If you move — and companies do — the analytics does not come with you. sonex is a script tag on any site, and the data is exportable.

The pricing shape is different, not just the number

Vercel’s model is pure usage: $0.03 per 1,000 events, and a hard stop on Hobby after a three-day grace period rather than an overage charge.

Ours is a tier with an allowance and a low per-event overage above it. Neither is universally better. Usage-based is fairer at unpredictable volume; a tier is easier to budget. What is worth checking is the retention add-on, because $10 a month per team for history is a real cost that does not appear in the headline price.

Choose Vercel Web Analytics if

• You deploy on Vercel, want analytics in one click, and 50,000 events a month is plenty.
• The absolute smallest possible script is the priority.
• You do not need history beyond a month, or you are already paying for Pro.

Choose sonex if

• You want revenue beside traffic.
• You need more than a month of history without an add-on.
• You want custom events without upgrading.
• You want funnels, retention, attribution or sessions.
• You would rather your analytics not be tied to where you host.

Running both is entirely reasonable — Vercel’s for platform-native basics, and something with revenue and history for the commercial view. Neither sets a cookie, so there is no consent cost to doing so.

Facts checked against Vercel’s own analytics limits-and-pricing documentation on 2026-08-17. Script sizes measured the same day.