Skip to main content

Full results on TanStack Start

Only c15t, Cookie Control, Didomi, Enzuzo, iubenda, Ketch, OneTrust, Osano, Usercentrics and CookieYes ran on TanStack Start. See every CMP on Next.js

GoodNeeds improvementPoor
Framework
Every CookieBench v3 metric for ten consent managers on TanStack Start, Mobile profile. One row per metric, one column per consent manager.
Metricc15tCookieYesOsanoCookie ControliubendaEnzuzoOneTrustUsercentricsDidomiKetch
Core Web VitalsLCP, CLS and TBT come from the page load. INP comes from the Accept click. Google rates LCP good up to 2.5 s and poor above 4 s. It rates INP good up to 200 ms and poor above 500 ms. It rates CLS good up to 0.1 and poor above 0.25. It does not rate TBT.
LCPLargest Contentful PaintNeeds improvement: 3.74 s+416 ms vs no CMPNeeds improvement: 3.51 s+176 ms vs no CMPNeeds improvement: 3.75 s+428 ms vs no CMPNeeds improvement: 3.82 s+496 ms vs no CMPNeeds improvement: 3.94 s+616 ms vs no CMPPoor: 4.10 s+768 ms vs no CMPNeeds improvement: 3.80 s+472 ms vs no CMPPoor: 4.18 s+848 ms vs no CMPPoor: 4.61 s+1.28 s vs no CMPPoor: 5.08 s+1.75 s vs no CMP
INPInteraction to Next Paint, for the Accept clickGood: 48 msGood: 24 msGood: 128 msGood: 24 msGood: 24 msGood: 72 msGood: 24 msGood: 16 msGood: 16 msGood: 32 ms
CLSCumulative Layout ShiftGood: 0.00Good: 0.00Good: 0.00Good: 0.00Good: 0.00Good: 0.00Good: 0.00Good: 0.00Good: 0.00Good: 0.00
TBTTotal Blocking Time during load24 ms24 ms+1 ms vs no CMP172 ms+148 ms vs no CMP214 ms+190 ms vs no CMP75 ms+51 ms vs no CMP224 ms+200 ms vs no CMP140 ms+116 ms vs no CMP112 ms+88 ms vs no CMP316 ms+292 ms vs no CMP1.00 s+981 ms vs no CMP
Page loadGoogle rates TTFB good up to 0.8 s and FCP good up to 1.8 s. Content paint is the paint time of the article image. We time it even when the banner was the largest element.
TTFBTime to First ByteGood: 16 ms+8 ms vs no CMPGood: 13 msGood: 10 msGood: 8 msGood: 9 msGood: 9 ms+1 ms vs no CMPGood: 9 msGood: 7 msGood: 9 msGood: 9 ms+0 ms vs no CMP
FCPFirst Contentful PaintGood: 268 ms+28 ms vs no CMPGood: 248 msGood: 244 msGood: 252 msGood: 248 msGood: 248 ms+8 ms vs no CMPGood: 248 msGood: 248 msGood: 252 msGood: 244 ms+4 ms vs no CMP
Content paintThe article image painted3.74 s+416 ms vs no CMP3.51 s+176 ms vs no CMP3.75 s+428 ms vs no CMP3.82 s+496 ms vs no CMP3.94 s+616 ms vs no CMP4.10 s+768 ms vs no CMP4.42 s+1.10 s vs no CMP4.18 s+848 ms vs no CMP4.61 s+1.28 s vs no CMP5.08 s+1.75 s vs no CMP
LCP elementWhat the browser picked as the largest element
CMP requestedFirst request for the CMP’s script or config
CMP loadedThe download of the CMP’s main script is complete
Consent knownThe CMP reports the visitor’s consent state1.81 s
Banner visibleThe banner is on screen250 ms1.19 s2.00 s2.38 s2.61 s2.67 s3.79 s4.07 s4.11 s7.85 s
Screen coveredShare of the viewport the banner covers
Click responseTime until the screen updates. We rate it like INP: good up to 200 ms, poor above 500 ms. The browser reports these times in 8 ms steps. It does not report clicks under 16 ms.
AcceptAccept all, from the bannerGood: 48 msGood: 24 msGood: 128 msGood: 24 msGood: 24 msGood: 72 msGood: 24 msGood: 16 msGood: 16 msGood: 32 ms
RejectReject all, from the bannerGood: 32 msGood: 24 msGood: 104 msGood: 24 msGood: 24 msGood: 56 msGood: 24 msGood: 16 msGood: 16 msGood: 32 ms
SaveSave, in the preferences dialogGood: 32 msGood: 80 msGood: 40 msGood: 48 msGood: 40 msGood: 16 msGood: 104 msGood: 32 ms
Next clickA click on the page 500 ms after AcceptGood: 16 msGood: 16 msGood: 16 msGood: 16 msGood: 16 msGood: 56 msGood: 16 msGood: 16 msGood: 16 ms
After AcceptMilliseconds from the click. Main thread is the long-task blocking time after it.
Banner goneBanner and any overlay removed
Consent appliedThe CMP’s consent state matches the choice
Scripts ranEvery script that the choice lets run is complete
Main threadLong-task blocking time after the click
After RejectReject lets no scripts run, so there is no script time to report.
Banner goneBanner and any overlay removed
Consent appliedThe CMP’s consent state matches the choice
Main threadLong-task blocking time after the click
After SaveFrom the Save click in the preferences dialog.
Banner goneBanner and any overlay removed
Consent appliedThe CMP’s consent state matches the choice
Scripts ranEvery script that the choice lets run is complete
Main threadLong-task blocking time after the click
Download sizeThe first load. Unpacked is the size after decompression. That is the code that the browser must parse.
TransferredBytes over the network during the first load700 KB+110 KB vs no CMP621 KB+31 KB vs no CMP675 KB+85 KB vs no CMP688 KB+99 KB vs no CMP711 KB+122 KB vs no CMP743 KB+153 KB vs no CMP801 KB+211 KB vs no CMP754 KB+164 KB vs no CMP843 KB+253 KB vs no CMP1.12 MB+532 KB vs no CMP
UnpackedThe same bytes after decompression1.31 MB+461 KB vs no CMP944 KB+97 KB vs no CMP1.19 MB+348 KB vs no CMP1.19 MB+346 KB vs no CMP1.30 MB+458 KB vs no CMP1.44 MB+592 KB vs no CMP1.70 MB+855 KB vs no CMP1.56 MB+711 KB vs no CMP1.85 MB+1.00 MB vs no CMP8.05 MB+7.20 MB vs no CMP
DeferredBytes transferred after the first load settled
RequestsRequests during the first load17+10 vs no CMP16+9 vs no CMP11+4 vs no CMP9+2 vs no CMP11+4 vs no CMP9+2 vs no CMP19+12 vs no CMP24+17 vs no CMP13+6 vs no CMP16+9 vs no CMP
Third-partyRequests to other hosts during the first load094242121769
Returning visitorsA visitor who accepted earlier loads the page again. We expect the banner to stay hidden.
Banner hiddenThe banner stayed hidden on the return visit
Consent knownTime from navigation until the CMP restores the stored choice
Consent checksNo tracker or cookie before the visitor chooses, and the choice holds afterwards. A check we could not run is not a failure.
ChecksLoad, Accept, Reject, Save, withdrawal and keyboard checks4 of 4 passed3 of 4 passed1 could not run1 of 4 passed2 could not run, 1 failed3 of 4 passed1 could not run4 of 4 passed4 of 4 passed4 of 4 passed4 of 4 passed4 of 4 passed4 of 4 passed

A value without a “vs no CMP” line had no measurable change from the same page without a consent manager.