A/B Testing Service-Level Agreement Calculator Examples
Worked A/B testing SLA examples for balanced, uneven, and traffic-limited experiment scenarios.
These worked examples show how daily eligible traffic, allocation, sample requirements, and minimum run time affect A/B test SLA progress and the recommended total duration.
Balanced split meets the SLA
A 50/50 test receives 1,000 eligible visitors per day for 14 complete days.
Input Summary
Daily eligible visitors
1,000
Test duration
14 days
Variant allocation
50%
Required sample per group
5,000 visitors
Minimum duration
14 days
Calculation Breakdown
- 1Total visitors1,000 * 1414,000 visitors
- 2Visitors in each group14,000 * 50 / 1007,000 visitors each
- 3Sample completionmin(7,000 / 5,000 * 100, 100)100%
- 4Duration completionmin(14 / 14 * 100, 100)100%
Result Summary
Total visitors
14,000 visitors
A/B Testing Service-Level Agreement Calculator
The test is 100% complete against the stated SLA, with an estimated 7,000 visitors in each group.
Balanced split needs more traffic
A test has run for 14 days, but only 400 visitors per day are eligible.
Input Summary
Daily eligible visitors
400
Test duration
14 days
Variant allocation
50%
Required sample per group
5,000 visitors
Minimum duration
14 days
Calculation Breakdown
- 1Total visitors400 * 145,600 visitors
- 2Smaller group5,600 * 50 / 1002,800 visitors
- 3Sample completion2,800 / 5,000 * 10056%
- 4Recommended durationmax(14, ceil(5,000 / 200))25 days
Result Summary
Total visitors
5,600 visitors
A/B Testing Service-Level Agreement Calculator
The test is 56% complete because sample collection, not run time, is the limiting requirement.
Uneven allocation delays the SLA
A risk-limited variant receives 20% of 2,000 daily eligible visitors.
Input Summary
Daily eligible visitors
2,000
Test duration
14 days
Variant allocation
20%
Required sample per group
5,000 visitors
Minimum duration
14 days
Calculation Breakdown
- 1Total visitors2,000 * 1428,000 visitors
- 2Variant visitors28,000 * 20 / 1005,600 visitors
- 3Control visitors28,000 - 5,60022,400 visitors
- 4Recommended durationmax(14, ceil(5,000 / 400))14 days
Result Summary
Total visitors
28,000 visitors
A/B Testing Service-Level Agreement Calculator
The test meets the SLA at 14 days, although the variant group is much smaller than the control.
Minimum duration controls the plan
A high-traffic product test uses a 50/50 allocation and a 21-day minimum run time.
Input Summary
Daily eligible visitors
5,000
Test duration
7 days
Variant allocation
50%
Required sample per group
5,000 visitors
Minimum duration
21 days
Calculation Breakdown
- 1Visitors per group after 7 days5,000 * 7 * 50 / 10017,500 visitors
- 2Sample completionmin(17,500 / 5,000 * 100, 100)100%
- 3Duration completion7 / 21 * 10033.3%
- 4Recommended durationmax(21, ceil(5,000 / 2,500))21 days
Result Summary
Recommended duration
21 days
A/B Testing Service-Level Agreement Calculator
Overall SLA completion is about 33%, even though the required sample is already collected.
How to Read Your Results
SLA completion is the lower of sample-size completion and duration completion.
A 100% result means the entered minimum traffic and duration thresholds are met, not that a test result is conclusive.
Sample shortfall shows how many additional visitors the smaller group needs to reach its target.
Recommended total duration is an estimate based on average eligible traffic and the selected allocation.
An uneven split makes the smaller group the likely constraint on completion.
Assumptions & Important Notes
- Daily eligible traffic remains broadly similar to the entered average.
- Traffic is allocated as entered and visitors are counted once.
- The required sample is a per-group requirement.
- The test continues without material pauses or measurement changes.
Related Examples
Frequently Asked Questions
Can an A/B test be 100% sample complete but below the SLA?
Yes. If the minimum run time has not been reached, duration completion remains below 100% and controls the overall result.
What happens with a 30/70 traffic split?
The 30% group is the smaller group, so its daily traffic determines the estimated days needed to meet a shared per-group target.
Should I enter all site visitors?
Enter visitors eligible for the experiment, rather than all site traffic, if those figures differ.
Why does the recommended duration round up?
The estimate uses complete calendar days, so any partial day needed to reach the smaller-group target is rounded up.
Ready to calculate your own result?
Use the live calculator with your own inputs, timing, and preferences.