New Customer Optimization in Meta

1. Why Optimize for New Customers in Meta?

Without a dedicated new customer optimization strategy, Meta optimizes for every purchase, regardless of whether it comes from a new or existing customer. As a result, a significant share of your advertising budget may be spent on existing customers—users who could often be re-engaged more cost-effectively through owned channels such as email, WhatsApp, SMS, or your CRM.

By optimizing for Mable's NewCustomerPurchase event instead, you provide Meta with a much clearer optimization signal. The algorithm is rewarded only when a new customer completes a purchase, allowing it to learn which users are most similar to your existing new customers. Over time, this shifts more of your advertising budget from existing customers toward acquiring new ones.

A common alternative is to exclude existing customers using Customer Lists or Custom Audiences. While this approach can be effective, it does not change the algorithm's underlying optimization objective.

As long as your campaigns continue optimizing for the standard Purchase event, Meta is rewarded for every purchase—whether it comes from a new or returning customer. Naturally, the algorithm continues to learn that existing customers are often the easiest users to convert.

If those existing customers are then excluded through audiences, your targeting starts working against the optimization signal. The algorithm is still trying to find people who resemble your existing customers, while being prevented from showing ads to them.

The most effective approach is to combine optimization for the NewCustomerPurchase event with excluding the Mable Active Customers Audience. This ensures that both signals reinforce each other. Meta is rewarded only for acquiring new customers while simultaneously receiving additional information about which users are already existing customers. As a result, the optimization signal and audience exclusion work together, typically delivering better performance than either strategy on its own.

⚠ Important:

Optimizing for NewCustomerPurchase does not mean Meta will generate only new customers. The algorithm looks for users who are most similar to your existing new customers. Because new and returning customers inevitably share certain characteristics, there will always be some overlap, meaning a portion of conversions will still come from existing customers.

Even so, we consistently see the following benefits in practice:

• Higher new customer rates

• Lower Customer Acquisition Costs (CAC)

• A larger share of ad spend allocated to new customer acquisition

• Better scalability for new customer growth

  ---

2. Prerequisites for New Customer Optimization

2.1 Review Your Meta Best Practice Setup

Before switching your campaigns to new customer optimization, make sure your Meta setup is configured according to best practices.

In particular, verify the following:

• Only the Mable Dataset is used for your campaigns.

• The dataset should ideally be connected to only one Catalog.

• The Product Identifier sent by Mable should match the Content IDs in your Catalog.

A properly configured setup ensures that Meta receives consistent optimization signals and correctly matches products to your catalog.

You can find the complete setup guide here ⬇:

Meta Best Practice Setup

2.2 Understand and Verify Mable Custom Events

When you connect a dataset to Mable, Mable automatically creates several custom events, including:

• NewCustomerPurchase

• ReturningCustomerPurchase

• NetRevenueAllCustomerPurchase

• AllCustomerPurchase

For each of these events, Mable also automatically creates a corresponding Custom Conversion.

2.2.1 Custom Events vs. Custom Conversions

⚠ Important:

Always use the actual Custom Event as your campaign optimization goal.

Example:

NewCustomerPurchase (Conversion API) ✅

NewCustomerPurchase (Custom Conversion) ❌

Custom Events are sent directly from Mable to Meta and provide the most accurate optimization signals for campaign delivery.

Custom Conversions, on the other hand, are primarily intended for reporting and performance analysis. They are calculated by Meta based on the underlying events and may occasionally differ slightly from the actual event count. For this reason, they should not be used as campaign optimization goals.

For more information on analyzing Mable Custom Events, see ⬇:

New Customer Reporting in Meta

2.2.2 Verify Your Custom Events

When new Mable events are first created, they may appear with a warning icon in the Meta Events Manager and need to be reviewed once before they can be used.

Only after this verification can the events be selected as campaign optimization goals.

If NewCustomerPurchase is not available as a conversion event, the most common reason is that the event has not yet been verified.

To review and approve the event:

Events Manager → Mable Dataset → Hover over the warning icon → Review

Once verified, the event will be available as an optimization goal within your Meta ad account.

2.3 Select the Correct Dataset

If your Meta ad account contains multiple datasets, always select the dataset connected to Mable when creating a campaign.

Only the Mable dataset includes the custom events generated by Mable, such as:

• NewCustomerPurchase

• ReturningCustomerPurchase

• AllCustomerPurchase

• NetRevenueAllCustomerPurchase

If a different dataset is selected, these events will not be available as optimization goals.

If NewCustomerPurchase does not appear as a selectable conversion event, first verify that the correct dataset has been selected.

2.4 Reporting on Custom Events

Mable Custom Events only become available as reporting columns in Meta Ads Manager after campaigns have been optimized for those events.

Until then, you can use the corresponding Custom Conversion for reporting purposes.

However, when it comes to campaign optimization, the recommendation remains the same:

Always optimize for the Custom Event—not the Custom Conversion.

For more information about reporting after switching to NewCustomerPurchase optimization, see ⬇:

New Customer Reporting in Meta

2.5 Create an Active Customers Audience

For new customer optimization, we also recommend creating a Mable Active Customers Audience.

By default, this audience includes all customers who have made a purchase within the last two years and is continuously synchronized with Meta.

It is used to exclude existing customers from your campaigns and complements optimization for the NewCustomerPurchase event (see Section 1).

You can enable the audience in Mable under:

Optimizations → Audiences

For detailed step-by-step instructions, see⬇:

How to Create a Custom Audience in Mable

Troubleshooting Audience Creation

If your audiences were not created or you receive an error message, one of the following is usually the cause.

1. Meta's Terms of Service for Custom Audiences Have Not Been Accepted

Before Custom Audiences can be created, Meta requires you to accept its Terms of Service for Custom Audiences.

If you haven't accepted the terms yet, you'll need to do so once before creating audiences. You can find the step-by-step instructions here:

How to Accept Meta Terms of Service in Mable

Once the terms have been accepted, you can retry creating the audience.

2. The Meta Ad Account Has Reached Its Audience Limit

Meta limits the number of Custom Audiences that can be created within a single ad account. If this limit has been reached, no additional audiences can be created until existing ones are deleted.

3. Setting Up New Customer Campaigns in Meta

Once all prerequisites have been met, the NewCustomerPurchase event should first have enough historical data before being used as a campaign optimization goal.

⚠ Important:

Before launching your first campaign optimized for NewCustomerPurchase, the event should have been sent from Mable to Meta for at least 7 days.

If you've only recently connected Mable, we recommend waiting a few days to allow Meta to build sufficient event history.

If your Mable integration has already been active for more than seven days and the event is being sent continuously, no additional preparation is required.

For the best results, configure the following two settings.

3.1 Select NewCustomerPurchase as the Optimization Goal

First, verify that the Mable Dataset is selected as the Conversion Source.

Only the Mable dataset contains the customer events generated by Mable, including:

• NewCustomerPurchase

• ReturningCustomerPurchase

• AllCustomerPurchase

If another dataset is selected, NewCustomerPurchase will not be available as a conversion event.

When creating the new ad set, select NewCustomerPurchase under Conversion Event.

⚠ Important:

Select the NewCustomerPurchase Custom Event, not NewCustomerPurchase (Custom Conversion).

This ensures that Meta is rewarded only for purchases made by new customers and gradually learns to find users who are most similar to your existing new customers.

If NewCustomerPurchase is not available as a conversion event, the most common reasons are:

• The wrong dataset has been selected.

• The event has not yet been verified in Meta Events Manager.

Instructions for verifying events are covered in Section 2 of this guide.

3.2 Exclude the Active Customers Audience

We also recommend excluding the Mable Active Customers Audience from your campaign targeting.

This audience includes all customers who have made a purchase within the last two years and is updated automatically.

Excluding this audience provides Meta with an additional signal indicating which users are already existing customers.

Combining:

• Optimization for NewCustomerPurchase

• Exclusion of the Active Customers Audience

typically provides the clearest signals for new customer acquisition.

For a complete step-by-step walkthrough—including selecting the Mable Dataset, choosing the NewCustomerPurchase event, and excluding the Active Customers Audience—see:

How to Set Up a Meta Campaign for New Customer Optimization

4. When Should You Optimize for NewCustomerPurchase?

For most brands, optimizing for NewCustomerPurchase is the recommended approach.

Most businesses use paid advertising primarily to acquire new customers, while existing customers can be re-engaged more cost-effectively through owned channels. As long as that is your strategy, NewCustomerPurchase is generally the better optimization goal.

There are, however, a few exceptions.

4.1 Businesses That Naturally Acquire Almost Only New Customers

If your business already consists almost entirely of first-time buyers—for example, because you sell high-ticket products, subscription products with little or no repeat purchasing, or products with very long repurchase cycles—the additional impact of optimizing for NewCustomerPurchase is typically smaller.

While using NewCustomerPurchase won't hurt performance, in these edge cases it often performs very similarly to the standard Purchase event.

4.2 Low Conversion Volume

For very small ad accounts with limited conversion volume, it may be better to continue optimizing for Purchase while simply excluding the Active Customers Audience.

Once conversion volume has increased, you can switch to NewCustomerPurchase. This is often the most practical intermediate step for smaller accounts.

Meta can continue optimizing based on the much larger pool of purchase events while existing customers are still excluded from campaign delivery.

As a general guideline, Meta performs best when each ad set generates sufficient conversion volume. A commonly used benchmark is around 50 conversions per ad set per week, or roughly 7 conversions per day.

Even if an account generates a healthy number of new customer purchases overall, the volume may still be too low if conversions are spread across many ad sets.

Example:

• 39 new customer purchases per day

• Distributed across 16 ad sets

This results in fewer than 3 new customer conversions per ad set per day on average.

In this scenario, Meta may struggle to learn from the NewCustomerPurchase event, causing ad sets to remain in the learning phase.

In these situations, the more stable approach is to:

• Continue optimizing for Purchase

• Exclude the Active Customers Audience

This allows Meta to learn from the larger volume of purchase conversions while preventing a significant share of existing customers from seeing your ads.

Once each ad set generates enough NewCustomerPurchase conversions, you can switch to the more specific optimization event.

5. Best Practice for Switching to NewCustomerPurchase

We generally recommend not moving your entire budget to NewCustomerPurchase immediately.

Instead:

• Duplicate your existing ad sets.

• Configure the new ad sets to optimize for NewCustomerPurchase.

• Start with a smaller budget.

If performance improves over time:

• Gradually increase the budget.

• Gradually reduce spending on the original ad sets.

This approach keeps your existing setup stable while giving Meta time to learn the new optimization signal.

Although it is possible to update existing ad sets directly, doing so typically results in greater performance fluctuations.

6. Learning Phase and Performance Expectations

After switching to NewCustomerPurchase, Meta first needs to learn how new customers differ from your historical customer base.

For many accounts, this learning phase lasts:

• Around one week

• Sometimes 2–4 weeks

In some cases, it may take even longer. For very large accounts with extensive historical data, we've seen the transition take up to five weeks before performance fully stabilized.

During this initial learning phase, ad sets optimized for NewCustomerPurchase may experience:

• Higher Customer Acquisition Costs (CAC)

• Fluctuating performance

• Lower efficiency

This is expected.

The algorithm is learning a new conversion pattern. The more data and budget available, the faster this learning process typically progresses.

7. Reporting After the Switch

After switching to NewCustomerPurchase, the most relevant KPIs in Meta Ads Manager change.

Instead of evaluating campaign performance based solely on total purchases, you should also assess results from a new customer acquisition perspective.

Key metrics include:

• NewCustomerPurchases

• Customer Acquisition Cost (CAC)

• NC ROAS

• New Customer Rate (NC%)

It's important to understand that some of Meta's default KPIs change in meaning after the switch.

Metrics such as Results or Cost per Result may initially look different than before, even if your actual campaign performance has not declined.

To evaluate performance accurately, we recommend updating your reporting columns and tracking new customer KPIs separately.

For more information, see ⬇:

New Customer Reporting in Meta