A/B Testing Core Web Vital (Monthly) Calculator Examples
Worked examples showing how control and variation Core Web Vitals pass rates translate into estimated monthly good-visit impact.
These examples show how eligible traffic, test allocation, and good Core Web Vitals rates affect the blended test result and the estimated impact of a full variation rollout.
Balanced test with a positive variation result
A 50/50 test compares a 65% control good CWV rate with a 72% variation rate across 100,000 monthly eligible sessions.
Input Summary
Monthly eligible sessions
100,000
Variation allocation
50%
Control good CWV rate
65%
Variation good CWV rate
72%
Calculation Breakdown
- 1Control sessions100,000 × 50%50,000
- 2Variation sessions100,000 × 50%50,000
- 3Blended good visits(50,000 × 65%) + (50,000 × 72%)68,500
- 4Full-rollout uplift100,000 × (72% − 65%)7,000 additional good visits per month
Result Summary
Full-rollout uplift
7,000 additional good visits per month
A/B Testing Core Web Vital (Monthly) Calculator
The blended test good CWV rate is 68.5%, and a full variation rollout estimates 7,000 additional monthly good visits.
Uneven traffic allocation during an early rollout
A site sends 20% of 240,000 eligible monthly sessions to a variation with a 76% good CWV rate, while control is at 70%.
Input Summary
Monthly eligible sessions
240,000
Variation allocation
20%
Control good CWV rate
70%
Variation good CWV rate
76%
Calculation Breakdown
- 1Group session counts240,000 × 80%; 240,000 × 20%192,000 control and 48,000 variation sessions
- 2Good visits during test(192,000 × 70%) + (48,000 × 76%)170,880
- 3Blended test rate170,880 ÷ 240,000 × 10071.2%
- 4Full-rollout uplift240,000 × (76% − 70%)14,400 additional good visits per month
Result Summary
Full-rollout uplift
14,400 additional good visits per month
A/B Testing Core Web Vital (Monthly) Calculator
The limited-traffic test has a 71.2% blended good CWV rate, while full rollout estimates 14,400 additional good visits monthly.
Variation with lower good CWV performance
A product page variation is tested on 75% of 80,000 eligible monthly sessions and records a lower good CWV rate than control.
Input Summary
Monthly eligible sessions
80,000
Variation allocation
75%
Control good CWV rate
82%
Variation good CWV rate
78%
Calculation Breakdown
- 1Group session counts80,000 × 25%; 80,000 × 75%20,000 control and 60,000 variation sessions
- 2Blended good visits(20,000 × 82%) + (60,000 × 78%)63,200
- 3Blended test rate63,200 ÷ 80,000 × 10079.0%
- 4Full-rollout change80,000 × (78% − 82%)-3,200 good visits per month
Result Summary
Full-rollout change
-3,200 good visits per month
A/B Testing Core Web Vital (Monthly) Calculator
A full variation rollout estimates 3,200 fewer monthly good CWV visits than control.
How to Read Your Results
The monthly additional good-visits result compares full variation rollout with full control rollout.
A positive result indicates more estimated good CWV visits with the variation; a negative result indicates fewer.
The blended good CWV rate describes the current test allocation, not the result of giving the variation all traffic.
Compare rates in percentage points for the clearest view of absolute performance differences.
Use consistent segments and definitions before comparing control and variation rates.
Assumptions & Important Notes
- All examples use the same Core Web Vitals assessment criteria for control and variation.
- Monthly eligible session counts are assumed to remain stable after a rollout.
- The examples are arithmetic illustrations and do not assess experiment validity or significance.
Related Examples
Frequently Asked Questions
Why can the blended test rate differ from the variation rate?
The blended rate includes both groups and weights each rate by its traffic share. It will generally sit closer to the group receiving more traffic.
Can a small rate change produce a large monthly impact?
Yes. Even a small percentage-point difference can translate to many sessions when the eligible monthly audience is large.
Should I use sessions or page views?
Use the measurement unit that matches the good CWV rates, and use that same unit consistently for both groups and the monthly volume.
Can I calculate device segments separately?
Yes. Separate mobile, desktop, country, or page-type calculations can provide a clearer estimate when those segments have materially different rates.
Ready to calculate your own result?
Use the live calculator with your own inputs, timing, and preferences.