> ## Documentation Index
> Fetch the complete documentation index at: https://docs.abtestly.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Quick start

> Get your first A/B test running in 15 minutes.

# Run your first test in 15 minutes

This guide takes you from signup to live experiment. No technical background required beyond editing your site's HTML.

## 1. Sign up

Go to [app.abtestly.com](https://app.abtestly.com) and create an account.

The Free tier includes:

* 1 site
* 1 active experiment
* 3,000 monthly tracked users (MTU)

Enough to validate the platform on a small page or a low-traffic site. For the current Starter / Pro / Business prices and MTU caps, see the [pricing page](https://abtestly.com/pricing).

## 2. Add your site

After signup, click **Add site** and enter your domain (e.g., `example.com`). You'll get a unique snippet code generated specifically for that site.

<Tip>
  Each site needs its own snippet — they're cryptographically tied to one domain. If you run `acme.com` AND `blog.acme.com`, add them as two separate sites and install the matching snippet on each.
</Tip>

## 3. Install the snippet

Copy the snippet code from your dashboard. Paste it into the `<head>` section of your HTML, **as the first script tag**, before any other JavaScript:

```html theme={null}
<head>
  <!-- ABTestly snippet — must be first script in head -->
  <script>!function(){/* your unique snippet */}();</script>

  <!-- Your other scripts come after -->
  <script src="https://www.googletagmanager.com/gtm.js?id=..."></script>
</head>
```

<Warning>
  The snippet must load **synchronously and before other scripts** to prevent flicker. Don't use `async` or `defer`. Don't load via Google Tag Manager. See [Snippet install](/snippet-install) for details.
</Warning>

## 4. Verify installation

Open your site in a fresh browser tab. Open DevTools (F12) → Console:

```javascript theme={null}
window.__ABTESTLY_CONFIG__
```

You should see an object with your site ID and config version number. If it returns `undefined`, the snippet didn't load — check the [snippet install guide](/snippet-install).

## 5. Create your first experiment

In the dashboard, navigate into your site and click **New experiment**. The pipeline editor walks you through six steps:

<Tip>
  Before you launch, you can check any variant on your live site with a
  [preview link](/preview) — it renders the variant for you (and works on
  drafts) without starting the test or recording any data.
</Tip>

<Steps>
  <Step title="Basics">
    Name, hypothesis, and the primary metric you'll measure.
  </Step>

  <Step title="Locations">
    Which URLs the test runs on. Pick a saved Location or define rules inline (e.g., URL starts with `/pricing`).
  </Step>

  <Step title="Audiences">
    Who's eligible. Filter by country, device, UTM, cookies, or any combination. Or leave blank to target everyone.
  </Step>

  <Step title="Variations">
    The original (control) and your variant(s). Each variant is a chunk of custom JavaScript and/or CSS that gets injected on matching pages.
  </Step>

  <Step title="Goals">
    Conversion events to track. Page visits, clicks, engaged sessions, or custom events fired via JavaScript.
  </Step>

  <Step title="Review & start">
    The pipeline shows green check marks once each step is complete. Click **Start experiment** when ready.
  </Step>
</Steps>

## 6. Start the test

Click **Start**. The new config propagates to all Cloudflare edge locations within \~60 seconds. Visitors start seeing variants based on their deterministic bucket assignment.

## 7. Watch results

The dashboard shows real-time exposures, conversion rates per variant, and statistical significance (95% / 99% confidence flags). Most tests need at least 5,000 exposures per variant to reach 95% confidence — plan your test runtime accordingly.

## Next steps

* [Snippet install reference](/snippet-install) — CMS-specific install guides
* [Preview links](/preview) — see a variant on your live site before launch, without affecting data
* [Targeting overview](/targeting-overview) — Locations, Audiences, conditions reference
* [Plans & pricing](/plans) — MTU explanation, grandfathering, cancellation
