Reducing CPA by 41% for a D2C Apparel brand by eliminating signal dilution in Meta ads

Industry: Fashion and apparel
Business model: Direct-to-consumer
Monthly Meta ad spend: More than $80,000
Markets: India and GCC
Platforms: Facebook and Instagram
Primary goal: Reduce cost per purchase and increase purchase volume

Note: “Clothora” is a placeholder name used to protect the client’s identity under an NDA.

Overview

Clothora is a mid-sized direct-to-consumer apparel brand focused on gender-neutral comfortwear.

The brand was running several sales campaigns across different markets, audience segments, and product categories. Its team had already tested multiple creatives and targeting strategies, but performance had remained flat for approximately 45 days.

The account was experiencing three consistent problems:

• Cost per purchase was increasing
• Purchase volume was unpredictable
• The team could not clearly identify which campaigns, audiences, and products were driving profitable growth

The objective of the audit was not to increase spending. It was to improve the quality of the signals being sent to Meta’s advertising system.

The challenge

Clothora’s Meta account had become overly complex.

Several campaigns were targeting similar audiences, promoting overlapping products, and optimizing toward the same conversion goal. Instead of helping Meta learn faster, the structure was dividing data and budget across campaigns that were competing with one another.

This created a signal dilution problem.

What we found

Three sales campaigns were targeting overlapping interest groups and lookalike audiences.

The account was also running both Advantage+ Shopping campaigns and manual catalog sales campaigns using nearly identical product sets.

In addition:

• The same products appeared in multiple campaigns
• Audience exclusions were limited
• Conversion events were not prioritized correctly
• Campaign-level UTM structures were inconsistent
• Some creatives generated clicks and engagement but did not contribute meaningfully to purchases

As a result, Meta was receiving conflicting information about which audiences, products, and creatives should receive more budget.

Why signal dilution was hurting performance

Meta’s delivery system depends on consistent conversion signals.

When several campaigns target similar users and promote the same products, each campaign receives only a portion of the available data. This can slow learning and make it harder for the system to identify patterns associated with high-value customers.

In Clothora’s account, the campaign structure was causing:

• Internal audience competition
• Product-level cannibalization
• Fragmented conversion data
• Inconsistent budget allocation
• Weak attribution between creatives and purchases
• Unstable learning across campaigns

The account did not need more campaigns. It needed clearer campaign ownership.

The diagnostic process

We completed a full campaign, audience, catalog, and measurement audit.

Campaign and audience review

We reviewed campaign breakdown reports to understand how spend, reach, frequency, and purchases were distributed.

We also evaluated audience overlap across:

• Interest-based audiences
• Lookalike audiences
• Retargeting pools
• Page and Instagram engagers
• Video-view audiences
• Existing customer exclusions

The review showed that multiple ad sets were attempting to reach similar users with similar products.

Pixel and events review

We reviewed Meta Events Manager to confirm how conversion events were being received and prioritized.

The audit focused on:

• Purchase event quality
• Event duplication
• Browser and server event matching
• Domain and event prioritization
• Attribution consistency
• Lower-funnel event configuration

Purchase data was available, but the account structure was giving too much importance to weaker signals such as ViewContent and AddToCart.

Catalog overlap analysis

We created a Google Sheets and Apps Script workflow to map products across campaigns.

The tool used catalog exports to:

• Pull product IDs
• Map products to campaigns
• Flag duplicate product promotion
• Identify campaigns competing for the same SKUs
• Connect creative IDs with conversion activity

The analysis showed that 47% of the product catalog was being promoted by more than one campaign.

This meant that nearly half of the catalog had no clear campaign ownership.

The strategy

The restructuring plan focused on three areas:

1. Catalog organization
2. Campaign consolidation
3. Signal quality improvement

Step 1: Organize the product catalog

The first step was to give every product a clear role.

Products were divided into three groups.

Best sellers

These products had consistent purchase history, stronger conversion rates, and enough data to support automated delivery.

New arrivals

These products required controlled prospecting and creative testing before being introduced into broader automated campaigns.

Low-intent products

These products generated browsing activity but had weaker purchase rates or limited historical demand.

Custom labels were added to the product feed so that each campaign could include or exclude products at the SKU level.

This prevented the same item from being promoted across multiple campaigns without a clear strategic reason.

Step 2: Rebuild the campaign structure

Overlapping campaigns were paused and replaced with a simpler structure.

Advantage+ Shopping campaign

The Advantage+ campaign was limited to proven best sellers.

This gave Meta a stronger product set and more concentrated purchase data.

Middle-funnel campaign

A separate campaign was created for users who had already interacted with the brand.

This included:

• Video viewers
• Instagram and Facebook engagers
• Website visitors
• Product viewers
• Cart users who had not purchased

Fresh user-generated content was introduced to move these audiences closer to purchase.

Manual prospecting campaign

A manual prospecting campaign was created specifically for new arrivals.

This allowed the team to test new products and creative concepts without disrupting the performance of the best-seller campaign.

Audience and product exclusions

Exclusion rules were added to reduce repetition and internal competition.

The rules helped prevent:

• Existing customers from re-entering acquisition campaigns
• Retargeting audiences from competing with prospecting campaigns
• The same products from appearing in multiple campaigns
• Repeated creative exposure across funnel stages

Step 3: Strengthen conversion signals

The next step was to improve the quality and consistency of the data sent to Meta.

Prioritize purchase events

Purchase was established as the primary optimization event.

Lower-value events such as ViewContent and AddToCart were still tracked, but they were not allowed to influence campaign structure more than completed purchases.

Improve pixel and server-side tracking

The Meta Pixel configuration was reviewed and updated with server-side tagging support.

This helped improve:

• Event coverage
• Event matching
• Purchase attribution
• Signal reliability
• Measurement across browsers and devices

Standardize campaign tracking

A unified UTM structure was applied across campaigns.

This made it easier to compare results across:

• Meta Ads Manager
• Website analytics
• Looker Studio dashboards
• Offline purchase data
• Campaign-level reports

Consistent campaign naming and UTM parameters also reduced reporting discrepancies.

Results after 21 days

The new structure produced measurable improvements without increasing the media budget.

What changed

The performance improvement did not come from a single creative or audience.

It came from reducing unnecessary competition inside the account.

By separating products, consolidating campaigns, and strengthening conversion signals, Meta received a clearer view of:

• Which products were most likely to convert
• Which audiences belonged to each funnel stage
• Which creatives generated purchases rather than clicks
• Which campaigns should receive more budget

This allowed the delivery system to make better optimization decisions.

Key learnings

More campaigns do not always improve performance

Running several campaigns with similar audiences and products can divide data and slow down learning.

A simpler structure often produces stronger signals.

Every product needs clear ownership

The same SKU should not appear in multiple campaigns unless there is a clear funnel or testing strategy.

Product-level discipline is important in Meta Ads, just as it is in Google Performance Max and other automated campaign types.

Clicks are not the strongest signal

A creative can generate a high click-through rate without generating profitable purchases.

Campaign decisions should prioritize purchase quality, revenue, and customer value rather than engagement metrics alone.

Automated and manual campaigns need separation

Advantage+ and manual campaigns can work together, but they should not compete for the same audiences and products.

Each campaign should have a defined purpose.

Strong signals improve machine learning

Clear product ownership, accurate conversion tracking, and concentrated purchase data give Meta better information to optimize delivery.

Better signals can lead to lower acquisition costs without additional spending.

Tools used

The audit and restructuring process used:

• Meta Ads Manager
• Meta Events Manager
• Meta audience overlap and campaign inspection tools
• Google Sheets
• Google Apps Script
• Product catalog exports
• Looker Studio
• UTM tracking
• Slack
• Notion

Final takeaway

Scaling does not always require a larger budget.

In many cases, the fastest performance improvement comes from removing unnecessary complexity.

For Clothora, the key growth lever was not adding more campaigns, audiences, or creatives. It was creating a cleaner system with clear ownership across products, audiences, campaigns, and conversion signals.

Within 21 days, the brand reduced its cost per purchase by 41.2%, increased purchase volume by 64.8%, and improved return on ad spend from 2.3x to 3.8x, with no increase in advertising spend.