A/B Testing Website Load Time Formula
Learn how website load-time differences and observed conversion rates are used to estimate a faster variant's projected conversion and revenue impact.
This calculator compares a control page with a test variant. It measures the speed and conversion-rate differences, then projects the observed conversion change across all monthly visitors to estimate a full-rollout revenue effect.
- 100% Free
- No Sign-Up Required
- Private & Secure
- Mobile Friendly
Projected Monthly Revenue Impact
Where:
Find the conversion-rate difference in percentage points, apply it to all monthly visitors, and multiply the resulting conversion change by the value of one conversion.
Variables Explained
| Variable | What It Means | Unit |
|---|---|---|
| monthlyVisitors - Monthly page visitors | Estimated number of visitors to the tested page in one month. | visitors |
| controlLoadTime - Control load time | Average load time of the current control page. | seconds |
| variantLoadTime - Variant load time | Average load time of the test variant. | seconds |
| controlConversionRate - Control conversion rate | Percentage of control visitors who convert. | percent |
| variantConversionRate - Variant conversion rate | Percentage of variant visitors who convert. | percent |
| averageConversionValue - Average value per conversion | Revenue or selected business value associated with one conversion. | currency |
| variantTrafficShare - Variant traffic share | Percentage of test traffic allocated to the variant. | percent |
Step-by-Step Calculation
Calculate load time saved
A positive result means the variant loads faster than the control.
loadTimeSaved = controlLoadTime - variantLoadTime
Calculate the speed improvement
This expresses the time saved as a percentage of the control load time.
loadTimeImprovement = (loadTimeSaved / controlLoadTime) * 100
Find the conversion-rate change
This is the absolute difference in percentage points between the two observed rates.
conversionRateChange = variantConversionRate - controlConversionRate
Calculate relative conversion uplift
This shows the conversion change relative to the control rate.
conversionUplift = (conversionRateChange / controlConversionRate) * 100
Estimate variant test exposure
This estimates how many monthly visitors were assigned to the variant at the selected allocation.
variantTestVisitors = monthlyVisitors * (variantTrafficShare / 100)
Project incremental conversions after rollout
The projection applies the observed percentage-point conversion difference to all monthly visitors.
incrementalMonthlyConversions = monthlyVisitors * (conversionRateChange / 100)
Project monthly revenue impact
The estimated additional conversions are multiplied by the average value of each conversion.
projectedMonthlyRevenueImpact = incrementalMonthlyConversions * averageConversionValue
Example: a page variant loading one second faster
Load time saved
3.5 - 2.5
1.0 second
Load time improvement
(1.0 / 3.5) * 100
28.6%
Conversion-rate change
3.3% - 3.0%
0.3 percentage points
Relative conversion uplift
(0.3 / 3.0) * 100
10.0%
Variant test visitors
50,000 * (50 / 100)
25,000 visitors
Projected incremental conversions
50,000 * (0.3 / 100)
150 conversions
Projected monthly revenue impact
150 * 75
$11,250
Final Result
The variant saves 1.0 second and shows a 10.0% relative conversion uplift, projecting 150 additional monthly conversions and $11,250 in monthly revenue impact after a full rollout.
Assumptions
- ✓Control and variant conversion rates represent comparable audiences, periods, and measurement methods.
- ✓The observed conversion-rate difference continues when the variant is shown to all monthly visitors.
- ✓Average value per conversion remains unchanged after rollout.
- ✓Average load time is a useful summary despite differences by device, location, browser, and connection quality.
Limitations
- !The calculation does not test statistical significance or confidence in the observed conversion difference.
- !A conversion change may reflect content, targeting, seasonality, tracking, or other test differences rather than load time alone.
- !A monthly projection may differ from actual results when visitor volume or conversion value changes.
- !Average page load time may conceal performance problems affecting a particular visitor segment.
Common Mistakes to Avoid
Entering a relative uplift, such as 10%, as the variant conversion rate instead of entering the measured rate, such as 3.3%.
Treating a change from 3.0% to 3.3% as a 0.3% relative increase; it is 0.3 percentage points and a 10% relative uplift.
Using only variant traffic to calculate the full-rollout projection; the rollout estimate uses all monthly visitors.
Assuming a positive observed difference proves that speed caused the result.
Using order value when gross profit, lead value, or another conversion value better matches the test objective.
Related Formulas
Frequently Asked Questions
How is website load time improvement calculated?
Subtract the variant load time from the control load time, then divide the time saved by the control load time and multiply by 100.
What is the difference between conversion-rate change and conversion uplift?
Conversion-rate change is measured in percentage points. Uplift expresses that change relative to the control conversion rate.
How are projected incremental conversions calculated?
The calculator multiplies monthly visitors by the conversion-rate difference expressed as a decimal. A 0.3-point increase equals 0.003.
Why does variant traffic share not change the full-rollout revenue projection?
Traffic share estimates variant exposure during the test. The rollout projection assumes that 100% of monthly visitors will see the variant.
Can the formula show a negative impact?
Yes. If the variant conversion rate is lower than the control rate, the projected conversion and revenue impact will be negative.
Ready to calculate your result?
Use the calculator to get instant results with your own inputs.