Skip to main content
Free Calculator

A/B Test Duration Calculator

Estimate how many days a fixed-horizon experiment needs, then compare the available sample and smallest detectable effect week by week. Plan conversion rates or continuous averages from eligible daily traffic and exposure.

Live plan

7 days

62,468 total observations required

Traffic used for duration

Use eligible experiment observations, then set how much of that traffic will enter this two-arm test.

Allocation and statistical settings

Defaults are suitable for a standard fixed-horizon plan. Open these when your traffic, allocation, or pre-registered hypothesis differs.

Fixed-horizon plan

7 days

62,468 total observations required

Total sample

62,468

Control

31,234

Variant

31,234

Duration assumes 10,000 eligible observations/day at 100% exposure. Edit Traffic used for duration. Target 5.50% · alpha 5% · power 80% · two-sided. Switch to Analyze to carry this plan forward.

Detectable effect by test length

The smallest effect this fixed-horizon plan can detect as cumulative eligible traffic grows. It is not the lift you should expect.

Based on 10,000 eligible observations/day at 100.00% exposure.

  • Week 1

    Selected MDE feasible by this week
    Total observations
    70,000
    Control
    35,000
    Variant
    35,000
    Detectable effect
    9.44% relative0.48 pp
    Target
    5.472%

  • Week 2

    Total observations
    140,000
    Control
    70,000
    Variant
    70,000
    Detectable effect
    6.63% relative0.34 pp
    Target
    5.331%

  • Week 3

    Total observations
    210,000
    Control
    105,000
    Variant
    105,000
    Detectable effect
    5.40% relative0.27 pp
    Target
    5.270%

  • Week 4

    Total observations
    280,000
    Control
    140,000
    Variant
    140,000
    Detectable effect
    4.67% relative0.24 pp
    Target
    5.233%

  • Week 5

    Total observations
    350,000
    Control
    175,000
    Variant
    175,000
    Detectable effect
    4.17% relative0.21 pp
    Target
    5.208%

  • Week 6

    Total observations
    420,000
    Control
    210,000
    Variant
    210,000
    Detectable effect
    3.81% relative0.20 pp
    Target
    5.190%

  • Week 7

    Total observations
    490,000
    Control
    245,000
    Variant
    245,000
    Detectable effect
    3.52% relative0.18 pp
    Target
    5.176%

  • Week 8

    Total observations
    560,000
    Control
    280,000
    Variant
    280,000
    Detectable effect
    3.29% relative0.17 pp
    Target
    5.164%

Preserve this plan before launch.

Carry the sample horizon, method, alpha, power, allocation, direction, stopping rule, and practical threshold into a private pre-test record.

Save this plan in GrowthLayer · free account

Share links store inputs in the URL fragment, which is not sent to the server. Anyone with the link can read those inputs.

The calendar estimate is ready. Preserve the assumptions behind it.

A calculator answers one question and forgets it. GrowthLayer keeps each test your team saves — the hypothesis, the numbers, and the decision — so future planning can start from a reviewable record.

  • Structured history of saved experiments
  • Import results from CSV in one step
  • Program summaries across saved tests
  • Free to start — no card required
Start freeSee what's included

Methodology

This calculator estimates calendar duration for a two-arm, fixed-horizon A/B test. It first calculates the required control and variant observations. Binary conversion metrics use a large-sample two-proportion plan based on baseline rate, relative or absolute MDE, alpha, power, test direction, and allocation. Continuous averages use the baseline mean, historical standard deviation, MDE, alpha, power, direction, and allocation. Required total observations are divided by eligible observations per day multiplied by experiment exposure. The duration is rounded up to a whole day. Eligible traffic means the randomization units that can actually enter the experiment after targeting and exclusion rules—not every session or page view. The estimate assumes traffic, eligibility, allocation, metric definitions, and the business environment remain reasonably stable. Complete business cycles and review holidays, campaigns, outages, novelty effects, and instrumentation changes separately. Those calendar risks cannot be solved by a sample-size formula. This is not a sequential stopping boundary. Do not repeatedly inspect an ordinary fixed-horizon p-value and stop at the first favorable result. Multiple variants require an explicit multiple-comparison plan outside this focused two-arm calculator.

Sources and limitations

Frequently Asked Questions

How long should an A/B test run?
Run until both arms reach the sample required by the declared baseline, minimum detectable effect, alpha, power, direction, and allocation. The calculator converts that sample into days using eligible daily observations and the percentage of traffic exposed to the experiment. Calendar rules alone cannot establish adequate power.
How does this A/B test duration calculator work?
It first creates a fixed-horizon sample-size plan, then divides the total required observations by eligible observations per day multiplied by experiment exposure. The result is rounded up to a whole day. It assumes traffic stays reasonably stable and the same eligibility rules apply throughout the test.
What counts as eligible observations per day?
Use people or randomization units that can actually enter the experiment—not total site sessions. Exclude traffic outside the targeted audience before entering the daily number. If only 50% of eligible traffic will be exposed, enter 50% separately rather than halving the daily estimate by hand.
Should every test run for at least two weeks?
Completing weekly cycles can reduce day-of-week distortion, but two weeks is not a universal sample rule. A high-traffic test may reach its planned sample sooner; a low-traffic test may need much longer. Treat weekly cycles and the powered sample as two separate checks.
Why does minimum detectable effect change duration so much?
Smaller effects are harder to distinguish from noise, so detecting them requires more observations and usually more days. Set MDE to the smallest effect that would change the business decision, not the smallest lift the team hopes to see.
Should I use a one-sided or two-sided duration plan?
Two-sided is the recommended default because a variant can help or harm. Use an increase-only one-sided plan only when that rule is declared before launch and an effect in the opposite direction would not be called significant.
Can I stop when significance appears before the estimated end date?
Not with an ordinary fixed-horizon design. Repeatedly checking and stopping on the first favorable p-value changes the false-positive behavior. Reach the declared sample or use a sequential method designed for repeated looks.
What if traffic changes after the test starts?
The sample target does not change merely because traffic changes, but the calendar estimate does. Recalculate the duration with a conservative eligible-traffic estimate, and investigate large allocation changes or sample ratio mismatch before interpreting the result.

Related Calculators

Unified A/B Test CalculatorA/B Test Sample Size CalculatorStatistical Significance CalculatorMinimum Detectable Effect Calculator

Choosing a method or double-checking another tool? Compare 11 public A/B test calculators and download the evidence matrix.

Updated for 2026. Built by GrowthLayer — built for evidence-aware experimentation teams.