Log in with Facebook
Log in with Facebook

Have your Facebook Lookalike audiences stopped performing post iOS 14? Here’s why.

Apple's ATT (App Tracking Transparency) has launched with iOS 14.5 and it's changing how Facebook ads work!
Keep up-to-date on any and all new changes. Stay ahead of the curve with Socioh’s pro-tips and tricks.

Sign up to get relevant updates about iOS14 changes and Facebook advertising delivered to your inbox.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Pre-iOS, things just worked. Facebook had rich intent data and no matter what your input, it always brought in sales.

That party is over. 

One of the casualties of iOS14 has been the performance of lookalike (LAL) audiences.

The problem is that the most commonly used tool for syncing your past buyers with Meta (so you can make LALs from this list) is usually your email marketing tool — Klaviyo being the most common one. 

Why is this a problem? Read on to find out.

So what exactly is the problem with using your email marketing tool to sync audiences?

We focus on Klaviyo here but this issue exists with all email providers and CSV file exports of customer lists. 

There are 2 ways in which Klaviyo can share seed audiences with Meta:

  1. Klaviyo's native integration with Meta
  2. Uploading a CSV file to Meta with predicted CLTV

Let’s look at each in detail:

Klaviyo's native integration 

The native integration automatically syncs a list of email addresses from Klaviyo to Meta. This is usually a list of all your past purchasers. So, let’s say Jack has just purchased a product from your store. His information will be added to the seed audience you are sending to Meta in the next sync, always keeping the list up-to-date. 

The problem with this is that Klaviyo's native integration with Meta doesn't send any value field for your past purchasers. Here’s a screenshot from their documentation:

Source:https://help.klaviyo.com/hc/en-us/articles/360015054631-How-to-Create-a-Value-Based-Lookalike-Audience-with-Klaviyo-CLV

As a result of this, Meta treats every customer in these lists in exactly the same manner — it treats each customer as though they have the same value. So it doesn't matter that Jack bought a T-shirt worth $30 while Jenny bought 2 dresses worth $500. They will both be given equal importance by Meta.

This is simply not right if you want effective Lookalike modeling. Meta strongly recommends that a value field be included next to each customer in your list.

CSV export with predicted CLTV

Most email clients offer this - the ability to export your buyer or subscriber segments into a CSV file. Klaviyo offers to include a predicted CLTV (Customer Lifetime Value) of each customer as well in this .csv file. 

If you are using Klaviyo to make custom audiences on Meta, this should be your preferred method of doing so. 

Based on this value, Meta would know how much weightage or importance to give to each customer when creating the Lookalike audience. Any lookalike audience made from this custom audience will be a 'Value-based' lookalike audience. Here’s the screenshot:

Source:https://help.klaviyo.com/hc/en-us/articles/360015054631-How-to-Create-a-Value-Based-Lookalike-Audience-with-Klaviyo-CLV

Now, Meta strongly recommends that you always make lookalike audiences from a value-based source whenever possible. LALs made from your CLV segment list will be value based.



See https://www.facebook.com/business/help/917879191754763 for details.

However, there is another disadvantage when you manually export your customer list with CLTV values.

New buyers are not sent to Meta automatically.

These segments or lists do NOT sync with Facebook. So if a new user is added to the customer segment, it’s not automatically pushed to Facebook.

NOTE: If the seed audience is refreshed (new records are added), meta is forced to refresh all LALs made from it every 72 hours. This keeps the LALs fresh (potential new buyers are added regularly) and the performance remains high for a longer period of time. 

Without the refresh, LALs fatigue quickly and performance tends to drop within weeks.

Is there a solution?

Ideally, every Lookalike audience you create should be based on RFM (Recency, Frequency, Monetary value) analysis to score each customer.

This gives Facebook a much better picture of the value of each customer as compared to attaching a customer’s lifetime purchase value in the value field.

We recommend making lookalike audiences using Socioh. Not only do we assign each customer a value when sending your list to Facebook, we also keep all the customer segments fresh and synced. This means new orders are automatically added in and the values constantly refreshed and sent to Facebook to keep your lookalikes fresh.

This solves both the shortcomings of using Klaviyo (or other email marketing tools) to create your seed audiences.

Have questions on how our audiences tool works? Drop us a comment or write to us!

Sign up to get relevant updates about iOS14 changes and Facebook advertising delivered to your inbox.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.