Facebook Value-Based Lookalike Audience VS Standard Lookalike Audience + How To Create Them | Ditly

Dedicated. Data-driven. Digital Marketing.

Lookalike audiences are one of Facebook’s most powerful audience targeting tools, which go far beyond the simple method of finding a niche audience and targeting it. Lookalike audiences manage to find the niche in people rather than the people in the niche.


A little while ago, Facebook rolled out a different kind of Lookalike audience, a Value-Based Lookalike audience and it has since begun to push this type of Lookalike as the default option.


So to explain what a Value-Based Lookalike audience is, let’s go back to a time when they didn’t exist and advertisers only had the choice of a basic Lookalike audience. Advertisers, like myself, would create lookalike audiences based on our existing customer databases, in order to find new customers who were very similar to our existing customers. That’s all well and good, and it often produced great results. Of course as advertisers, we wanted more!


When you have a large customer database you can usually split the customers into 4 varying segments of profitability. They are usually something along the lines of this:


  • High value customers; these customers tend to return very frequently and consistently to buy more from your business and usually spend a great deal of money.
  • Semi-regular customers; these customers return fairly frequently, spending a consistent amount each time and produce a steady stable revenue stream.
  • Lapsed customers; these customers were once either high value or semi-regular customers but have since not purchased in a while and therefore have lapsed.
  • One-time purchasers; these customers will purchase once from you and never return again, this type of customer is usually driven by a deal or offer such as Black Friday and has no intention to be a regular customer.


As advertisers, we always want to drive new customers that fall into the high value or semi-regular segments, as these have the strongest Life Time Value (LTV) for the business. So, instead of creating lookalike audiences from all customers we would extract the high value and semi-regular segments of our database only and create lookalikes of those people. Essentially telling Facebook that these are the type of people that we like, and we want more of them, so go find them.


What we were doing is exactly what Value-Based Lookalike audiences now do but automatically with AI! A Value-Based Lookalike audience is built from your customer database seed audience along with a value alongside each customer’s identifier. 


So for example, you will upload your existing customers email address and phone number alongside the amount they have spent with you since they have been a customer. Facebook’s AI tools will then harvest this data and begin to build out segments within it’s Lookalike audience which recognise what a High value customer looks like compared to what a One-time purchaser looks like.


Once you release ads to this Lookalike audience, Facebook will actively choose to bid higher for the people it deems are a close lookalike to your high value customers and will bid lower for those that look like a one-time purchaser. This gives Facebook the flexibility and information to generate a greater return on your ad spend.


One thing to note when building these value-based lookalikes is to not try and trick Facebook by uploading only your high value customers. Doing so will only limit the Lookalike audience’s ability to reach people and generate sales. This is because Facebook’s ad platform is an auction and at times your budget & bid may not be high enough to always bid for the potential high value customers, if you upload all customer data, then Facebook still has the option to bid lower to go after your one-time purchasers… therefore still generating revenue.


Over the years I’ve tested multiple value based lookalike audiences against standard lookalike audiences, using the same seed customer audience and ads. The CPM’s are often higher on a value based lookalike, almost 100% of the time they out-perform the standard lookalike audience on ROAS. Therefore, I would recommend only ever running value based lookalike audiences, and this is especially true if your business sells a range of products or services with a significant difference in price.