CMP Performance Benchmark 2026
How heavy are leading consent management platforms before a visitor even makes a choice? We measured the initial CMP payload on clean pages in Chromium, using empty storage, HAR capture and raw CDP request data.
Key results
This benchmark focuses on the initial CMP load: the network requests triggered by the CMP before the visitor changes preferences or interacts with the banner.
| CMP | Total initial transfer | Initial requests | Notes |
|---|---|---|---|
| Lean Cookie Consent mono | 5,256 B CDP / 4,956 B Brotli body | 1 | Standard monolingual config |
| Lean Cookie Consent bilingual | 5,442 B CDP / 5,142 B Brotli body | 1 | IT+EN config |
| Lean Cookie Consent 7 languages | 6,944 B CDP / 6,644 B Brotli body | 1 | EN, IT, ES, FR, DE, NL, PT |
| CookieYes | 35,728 B | 12 | Public vendor config |
| CookieScript | 43,704 B | 2 | Public vendor config |
| iubenda | 120,549 B | 5 | Public/sample config |
| Cookiebot | 153,456 B | 5 | Public vendor config |
| OneTrust | 213,417 B | 14 | Public vendor config |
Metric definitions
CDP initial transfer
Chromium's encodedDataLength for the initial CMP resources. It can include transfer overhead.
Brotli HTTP body
The compressed response body size exposed by the server through Content-Length.
Initial requests
Every CMP-originated request during initial load, before preference interaction.
Methodology summary
Test date: 2026-08-08. For each measurable product, we created a clean HTML fixture with only the CMP snippet and a neutral body, used a fresh Chromium context, started with empty cache and storage, loaded the fixture once, and captured HAR, screenshot, raw CDP request table and initial network requests.
For competitors where a disposable public customer account was unavailable, the benchmark used public vendor-owned configuration extracted from official vendor sites and loaded into a clean intercepted page on the same vendor host. This is reproducible, but may not represent a minimal customer configuration.
Reproducibility
The normalized evidence pack for this benchmark is available as CSV and JSON. The public methodology explains the fixture, metrics, competitor configuration handling and caveats. A short README lists supported claims and claims to avoid. Last updated: 2026-08-16.
Methodological classification
This benchmark uses a practical classification to keep comparisons fair. It is a methodological distinction for this test, not a universal legal or regulatory definition.
Cookie banner
Notice UI and local choice storage.
Consent library
Developer component with some blocking or preference handling.
Full CMP platform in this comparison
Configurable banner, preference center, prior blocking, Consent Mode v2, consent evidence, versioning and platform/workflow capabilities such as multi-site management.
Caveats
- Results vary by configuration, language count, product version, region, cache state and optional assets.
- The around-5 KB claim applies to Lean's standard mono/bilingual Brotli body, not CDP transfer.
- Axeptio was measured during research but is not used in the main headline because the public configuration loaded a large video asset immediately.
- Benchmark numbers should be cited from this page as the primary source.
FAQ
Is Lean Cookie Consent about 5 KB?
Lean's standard monolingual and bilingual configurations are about 5 KB when measured as Brotli HTTP body size: 4,956 B and 5,142 B respectively. Chromium CDP transfer is higher because it includes transfer overhead.
Does this benchmark support an absolute lightest-CMP claim?
No. It supports a narrower and more defensible claim: Lean Cookie Consent is one of the lightest full-featured CMPs we were able to verify in this benchmark set.
Why not compare only JavaScript file size?
Because the initial user-facing cost of a CMP depends on the resources required to render and operate the banner, not only on the nominal size of a single JavaScript file.
A full CMP does not have to be heavy.
Lean Cookie Consent gives agencies and developers a one-request CMP with Consent Mode v2, blocking and consent evidence workflows.