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

# Retargeting & Campaign Strategies

> Set up retargeting, lookalike audiences, and testing strategies in AdStellar.

This guide covers how to set up retargeting campaigns, build lookalike audiences, plan your budgets, and structure tests for the best results.

***

## Retargeting Campaigns

Retargeting shows ads to people who have already interacted with your business: website visitors, email subscribers, video viewers and more. In AdStellar, you retarget by launching into an ad set that uses a Custom Audience.

**1. Create Custom Audiences in Meta**

Go to Meta Ads Manager > Audiences and create custom audiences (website visitors, email lists, video viewers, etc.).

**2. Create an ad set with that audience**

On the Launch page, click **+ New** in the **Ad sets** card. Open the **Custom audiences** row and choose your audience. See [Choose or create ad sets](https://docs.adstellar.ai/guides/launch/ad-sets.md).

**3. Launch your ads into it**

Load your creatives, fill in the ad setup and launch. The ads are shown to people in your custom audience.

> [!TIP]
>
> Create a separate ad set for each retargeting segment (e.g., "Website Visitors - 7 days", "Add to Cart - 30 days", "Email List") so you can compare them. Click **Add another ad set** in the pop-up to copy the one you are on and change only the audience.

***

## Lookalike Audiences

Lookalike audiences find new people who are similar to your best customers. They're one of the most effective ways to scale cold prospecting on Meta.

**1. Create a source audience**

In Meta Ads Manager, create a Custom Audience from your best converters (e.g., purchase customers, high-value leads).

**2. Create a Lookalike Audience**

In Meta Ads Manager > Audiences, create a Lookalike Audience from your source. Choose a percentage (1%–10%) — lower percentages are more similar to your source.

**3. Create an ad set with the lookalike**

On the Launch page, click **+ New** in the **Ad sets** card. Open the **Custom audiences** row and choose your Lookalike Audience.

**4. Launch your ads into it**

Load your creatives, fill in the ad setup and launch.

> [!TIP]
>
> Start with a **1% lookalike** for the most similar audience, then expand to 2–5% as you scale. Use a separate ad set for each percentage so you can compare performance in [AI Insights](https://docs.adstellar.ai/guides/insights/ai-insights.md).

***

## Budget Recommendations for Testing

| Scenario | Recommended Budget |
| - | - |
| **Minimum for Leads** | \$5/day per ad set |
| **Minimum for Sales** | \$10/day per ad set |
| **Recommended for testing** | \$20–50/day per ad set |
| **CBO campaigns** | \$100+/day total (Meta auto-allocates to top performers) |

* When each ad set has its own daily budget, your combined daily budget is the sum of all the ad sets you launch into. Plan your total budget accordingly. An ad set with a lifetime budget spends that amount over its whole schedule, not per day.
* Let campaigns run **3–7 days** (or until \~50 conversions per ad set) before making decisions.
* With CBO at \$100+/day, Meta will automatically shift budget to your top-performing ad sets.

***

## Testing Strategies

### Creative Testing

Test **3–5 creatives** at a time. Load them on the Launch page and launch them into the same ad set. After 7 days, check [AI Insights](https://docs.adstellar.ai/guides/insights/ai-insights.md) to see which creative wins.

### Audience Testing

Create **3–4 ad sets** with different targeting approaches (interests vs. lookalikes vs. broad) and launch the same ads into all of them. [AI Insights](https://docs.adstellar.ai/guides/insights/ai-insights.md) will rank each audience by your result metric.

### Copy Testing

Add up to five primary texts and five headlines in the **Ad setup** card. They are stored inside one ad and Meta picks which to show, so the number of ads stays the same. This applies to ads with one file: an ad with several sizes uses only the first primary text and the first headline. To compare two texts as separate ads, load the creative twice and give each ad its own text. See [Change one ad or many](https://docs.adstellar.ai/guides/launch/edit-ads.md).

***

## Scaling Winners

Once you identify what is winning through [AI Insights](https://docs.adstellar.ai/guides/insights/ai-insights.md):

1. **Relaunch winners** — Launch the winning creative and copy again from the [Launch page](https://docs.adstellar.ai/guides/launch/overview.md), into a new or a duplicated ad set.
2. **Increase budget** — Raise the budget on your winning audience/creative combination (increase by 20–30% at a time to avoid resetting the learning phase).
3. **Build lookalikes** — Create Lookalike Audiences based on your converting audience segment and launch them in new campaigns.
4. **Let the Agent scale for you** — Ask the [Agent](https://docs.adstellar.ai/agent/launch-campaigns.md) to build a campaign from your top creatives.

> [!WARNING]
>
> Avoid changing campaigns during the learning phase (first \~50 conversions). Edits to budget, targeting, or creatives reset the learning phase and delay optimization.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.