> For the complete documentation index, see [llms.txt](https://support.fibr.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://support.fibr.ai/hypothesis-and-experimentation/how-to-launch-new-experiments-on-fibr-ai.md).

# How to launch new experiments on Fibr AI?

{% embed url="<https://youtu.be/sciBrsX6OwI?si=emOaJyb1fVZRX_cB>" %}

### Steps to create experimentation campaign on Fibr AI?

Ready to launch your first experiment? You'll go from URL to live test in just a few minutes. Here's how to generate hypotheses and set up your campaign.

#### Step 1: Start Your Hypothesis

**Option 1: If you're a first-time user:**\
Navigate to 'Experimentation' in your left sidebar, then click 'Get Started'

<figure><img src="/files/C2zsJKZY7pA6tMqWxiDS" alt="" width="563"><figcaption></figcaption></figure>

**Option 2: If you've created hypotheses before:**\
Navigate to 'Experimentation' in your left sidebar, then click 'Add New Page'

<figure><img src="/files/CVGyZIWosXSxOi5WADHo" alt="" width="563"><figcaption></figcaption></figure>

#### Step 2: Enter Your Page URL

Paste your website URL into the Hypothesis module. Fibr will automatically fetch your page structure, content, and visual elements to understand exactly what your visitors see.

<figure><img src="/files/i2TFrJiOSgoKiACHEfVu" alt="" width="563"><figcaption></figcaption></figure>

{% hint style="info" %}
Fibr AI generates hypotheses for the specific URL you enter, so make sure you're using the exact page you want to test.
{% endhint %}

#### Step 3: Choose Your Starting Point

After entering your URL, you'll see two options for creating hypotheses:

| Type                                  | When to Use                                                                                  | What Fibr Does                                                                                                   |
| ------------------------------------- | -------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------- |
| **Generate Hypothesis** (Recommended) | You want Fibr's AI to analyze your page and surface optimization opportunities automatically | Fibr scans your layout, copy, and metrics, then generates data-backed hypotheses tailored to your business goals |
| **I Already Have My Hypothesis**      | You've already identified what you want to test from previous analysis or insights           | Fibr skips generation and takes you directly to variant setup and tracking configuration                         |

<figure><img src="/files/5nXOAGogAvwCxlfBGpZM" alt="" width="563"><figcaption></figcaption></figure>

{% hint style="info" %}
If you already know what you want to test, Option 2 gets you to experiment setup faster.
{% endhint %}

#### Step 4: Fill the campaign details      &#x20;

Give your hypothesis a clear, descriptive name and choose the goal you want to optimize (like Lead Submitted, Add to Cart, or Start Free Trial).

<figure><img src="/files/ieEHQFrADZkOJg7Us4ma" alt="" width="563"><figcaption></figcaption></figure>

> **We Recommend Connecting Google Analytics (GA4)**
>
> Toggle on "Connect Google Analytics" to unlock more accurate, data-driven hypotheses. When connected, Fibr will:
>
> * Analyze real user behavior (scroll depth, time on page, bounce rate, and more)
> * Leverage audience insights to make recommendations that match your actual traffic patterns

{% content-ref url="/pages/hWzm3Lb4r9XX0TAAQXQb" %}
[Integrate Google Analytics (GA4)](/integrations/integrate-google-analytics-ga4.md)
{% endcontent-ref %}

#### Step 5: Add Page Context

Before generating hypotheses, Fibr needs a bit of context about your page and audience to make its suggestions spot-on.

* **Review Key Benefits:** Fibr automatically detects and pre-fills your main value propositions. Check that they're accurate.
* **Define Your Target Audience:** Select the audience segment this page is designed for.
* **Add Common Customer Objections:** Help Fibr optimize your content by sharing what concerns or hesitations your customers typically have.

<figure><img src="/files/dOLQsyW8SdCrDYpej6Sn" alt="" width="563"><figcaption></figcaption></figure>

Once you've reviewed or adjusted these inputs, click **Generate Hypothesis**.

#### Step 6: Review Your Hypotheses

Fibr will analyze your page, highlight what's working and what's not, and present you with the top data-backed hypotheses.

Each hypothesis includes:

* What element to change (text, visuals, or layout)
* Why it matters for your conversion goals
* The expected impact on performance

<figure><img src="/files/t3wPKqIKcOvdiVWAcydO" alt="" width="563"><figcaption></figcaption></figure>

You're ready to convert any of these directly into an experiment.

#### Step 7: Select Hypotheses to Test

Click **Test Hypothesis** to choose which hypotheses you want to run. Fibr lets you select your top picks based on potential impact and your testing priorities.

<figure><img src="/files/Jw6pjopK4quMbt0GIvZ2" alt="" width="563"><figcaption></figcaption></figure>

Once you've confirmed your selections, click **Create Experiment**.

#### Step 8: Review and Publish Your Campaign

Fibr automatically creates an experiment with each hypothesis as a separate variant. Traffic is split uniformly across all variants by default, but you can adjust the allocation based on your needs.

<figure><img src="/files/W3zooPiIEEgCl57bNnNd" alt="" width="563"><figcaption></figcaption></figure>

**Important:** Make sure your total traffic allocation always adds up to 100%.

Next, add your **Primary Metric** to track. This is how you'll measure which variant wins (e.g., conversion rate, click-through rate, or form submissions).

Review everything one last time, then click **Publish** to make your campaign live.

**Congratulations!** Your experiment is now running. Fibr will track performance and let you know when you have statistically significant results.