> For the complete documentation index, see [llms.txt](https://docs.humblytics.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.humblytics.com/split-testing-overview/using-the-humblytics-a-b-sample-size-calculator.md).

# Using the A/B Test Planner

Find out how many visitors your test needs and how long it will take, before you build anything.

The planner tells you how many visitors each version needs before you can trust the result, and how many days that will take with your traffic.

Go to **Optimize → Experiments** and click **Plan a Test**. The same calculation also appears in the **Plan** section of every new experiment.

<mark style="background-color:pink;">ADD GIF HERE</mark>

## 1. Pick a test goal

**Simple Engagement Test** (clicks, scrolls, time on page), **Conversion Test** (form submissions, purchases), **Revenue Optimization** (AOV, LTV impact), or **Multi-Step Funnel**. This fills in sensible starting values.

## 2. Enter your numbers

| Field                             | What it means                                                                                                        | Where to find it                                     |
| --------------------------------- | -------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------- |
| **Baseline Conversion Rate (%)**  | How often visitors convert today.                                                                                    | Your page's current rate in **Observe → Analytics**. |
| **Minimum Detectable Effect (%)** | The smallest improvement worth finding, **relative** to your baseline. 20% on a 5% baseline means detecting 5% → 6%. | Your call. Smaller effects need far more visitors.   |
| **Statistical Significance**      | How sure you want to be. **90% (Exploratory)**, **95% (Standard)**, **97% (Revenue)**, or **99% (High confidence)**. | 95% is the usual choice.                             |
| **Statistical Power**             | The chance of catching a real improvement if there is one.                                                           | Leave at 80%.                                        |
| **Daily Visitors**                | Visitors per day to the page you'll test.                                                                            | **Observe → Analytics → Pages**.                     |
| **Number of Variants**            | How many versions, including the control.                                                                            | Usually 2 (A/B). More versions need more visitors.   |

## 3. Read the results

* **Sample size per variant:** visitors each version needs.
* **Total:** visitors needed across all versions.
* **Recommended duration** and **Review results on:** how long the test will take, and the date to check back.

A **Low Traffic Warning** means the test could take months. The fix is usually to test a busier page, or test a bigger change so the effect is easier to spot.

Click **Create Test with These Settings** to start a new experiment with these values filled in, or **Export CSV** to share the plan.

## Example

| Input                     | Value         |
| ------------------------- | ------------- |
| Baseline conversion rate  | 5%            |
| Minimum detectable effect | 20% (5% → 6%) |
| Significance / power      | 95% / 80%     |
| Daily visitors            | 1,000         |

**Result:** 7,665 visitors per variant, 15,330 in total, about **2 weeks**.

Halve the traffic and the test takes twice as long. Halve the MDE and it needs roughly four times the visitors.

Next: [Deciding How Long to Run an A/B Test](/split-testing-overview/deciding-how-long-to-run-an-a-b-test.md).