Control CPP  and Lift the ROAS 37% by retaining cookies by 1PD and adopting First-Party Data Ops

Customer: CookD
Industry: E-Commerce

In the ditch Shopify series, we have got another beautiful honest customer story who had a very interesting challenge which Shopify just cannot handle. I mean Shopify CAPI.

The Customer Story

Cookd had products across various categories and price ranges, from high to low. However, the Customer Acquisition Cost (CPP) and the customer purchase value remained almost the same, which puzzled the business.

For instance, a campaign analysis revealed an intriguing fact: acquiring a customer with a purchase value of $300 costs $270 in CPP.

Goal: The main goal was to bring in more new high value customers for the brand using Meta as the primary acquisition channel. 

Secondary goal: This meant optimizing the CPP and increasing the number of purchases as well.

Meta AI leverages machine learning to predict the potential Return on Ad Spend (ROAS) an individual might generate. By analyzing user data, these algorithms estimate potential ROAS and place higher bids on your most valuable customers, thereby maximizing the value of your campaign conversions.

This highlights the importance of event optimization.

When you optimize for purchase events, you get more purchases. When you optimize for high Average Order Value (AOV) purchases, you achieve that. When you optimize for category-based purchases, you see those results.

Note: Meta recommends optimizing campaigns to maximize the value of conversions, which can lead to a 100% increase in incremental revenue ROAS compared to focusing solely on conversions.

Currently, when a user visits the site after 7 days, their cookies are deleted, which often results in GA4 reporting a high number of new users due to the 7-day cookie expiration.

To address this, we have implemented first-party domain tracking for Cookd. This allows us to retain cookies indefinitely and establish precise attribution for conversions.

Instead of optimizing the campaign solely for default purchase events, we create multiple custom conversion events based on Average Order Value (AOV) to train the algorithm and maximize conversion value.

The AOV-based conversion events are as follows:

  • High AOV purchase (e.g., above $1000)
  • Mid AOV purchase (e.g., between $500 and $1000)
  • Low AOV purchase (e.g., below $500)

These custom events allow us to refine the algorithm and focus on securing higher AOV transactions.

This is the ideal EMQ score for custom conversion events when we want to optimize for the campaigns.

We started by optimizing for high AOV purchase events in one campaign and gradually applied this strategy to others. The focus shifted from solely the ROAS metric to also considering CPP and overall incremental ROAS.

The results column for the campaign displays the number of conversions for the specific optimized conversion event.

For instance, if we optimize for high AOV events, Facebook results will show the number of high AOV orders only. However, our focus is on overall conversions, which include high, mid, and low AOV orders, along with the CPP for each AOV purchase event and the overall conversions

As we have enabled 1P domain tracking for Cookd, we have browser cookies saved in the server forever i.e. all fbclid and gclid are collected and stored forever. This helps in attribution and will significantly lift the ROAS.

Cumulatively, we see a significant impact by;

  1. 35% increase in ROAS
  2. 45% decrease in the CPP.

Get started with CustomerLabs CDP

Schedule a 1-1 Demo


Unified data to boost ecommerce growth


Engage your customers across the funnel with a unified martech stack


Increase product metrics with a unified martech stack


Scale your customers quickly with the right data