Facebook just quietly released the biggest change to its ad platform since the app install ad: Segment Asset Customization.
This is currently an alpha product only released to a handful of top advertisers and only available for video view objective campaigns, but in 2019 it will be rolled out more broadly to all advertisers and campaign types.
It will ultimately transform the way media buyers place ads on Facebook.
But in order to understand the groundbreaking nature of this new feature, it’s imperative to first understand the current state of ad personalization on Facebook.
The Infinite Segmentation Engine
Facebook offers advertisers the ability to slice an audience into any combination of demographic, location, life event, interest and behavior.
Need to reach 81,000 liberal women in their late 20s who are currently in long distance relationships and have cats?
How about 39,000 parents who had a child within the last 12 months who like yoga and heavy metal and primarily use an AOL email address?
Given that Facebook allows an unprecedented degree of audience segmentation, it also allows an unprecedented degree of ad personalization.
How to Personalize on Facebook
It’s no secret that ads will receive more engagement when they are personalized to their audience, but Facebook currently forces advertisers to make sacrifices to achieve effective personalization.
Imagine you want to personalize your ad creative so that women and men see different ads.
Currently the only way to achieve this is to split your audience and budget into two ad sets, creating one for men and one for women.
You could then assign gender-specific ads to each ad set.
In order to understand why this campaign structure is a sacrifice — and the specific disadvantages of creating separate ad sets across every personalization dimension — we must first understand the fundamental changes to Facebook’s ad serving algorithm over the last two years.
Action Propensity Segmentation
In the early days, the savviest advertisers would slice and dice Facebook audiences and ad creative in an attempt to figure out which niche audience would respond best to which niche ad creative. This was an enormously effective strategy and the Facebook auction would reward hyper-personalization with significantly lower costs.
In Q3 2016, Facebook launched a new ad product called App Event Optimization and the pendulum started shifting toward a new strategy.
With the advent of app event optimization Facebook increasingly began segmenting their audience in a new way: action propensity.
Within every audience, no matter how niche, there is a certain subset that has:
- low propensity to take any action from ads
- propensity to watch videos but not click
- propensity to click but not take additional action
- propensity to install but not take additional action
- propensity to take in-app action but not monetize
- propensity to monetize
Advertisers can’t directly select these action propensities through audience targeting. Instead, they must rely on event bid optimization to train Facebook to serve their ads to the most relevant subset of their targeted audience.
- low propensity to take any action from ads = brand awareness objective
- propensity to watch videos but not click = video view objective
- propensity to click but not take additional action = click objective
- propensity to install but not take additional action = install objective
- propensity to take in-app action but not monetize = app event objective
- propensity to monetize = value objective
In short, action propensity segmentation has enabled Facebook to demand the highest possible advertiser bid for each impression on its platform.
For example, a brand advertiser who is attempting to get the most video views for the lowest cost won’t be willing to pay a premium to reach users who monetize if that increases their cost per video view.
On the other hand, a performance marketer who only cares about maximizing directly attributable ROI would happily pay a huge premium to reach users who monetize as long as that drives higher ROI.
And indeed, this increasingly sophisticated action propensity segmentation is reflected in Facebook’s ARPU — up a staggering 34% for Q2 and 30% for Q3 comparing 2018 to 2017 for US and Canadian users.
The Shift from Static to Dynamic Audiences
Using app event optimization and action propensity segmentation, advertisers began seeing more success by targeting broader audiences across fewer ad sets, which allows Facebook’s increasingly powerful algorithms more room to identify the pockets of the audience with the highest propensity to achieve the advertiser’s objectives at the lowest cost possible.
Audiences are now fluid, as initial audience targeting is merely the starting point of a dynamic audience that evolves based on your optimization event (e.g. video views, clicks, installs, purchases etc) and the actual event conversion data flowing through.
When you divide your audience and budget across multiple ad sets, you may gain the ability to personalize ad creative, but you create silos which constrain Facebook’s algorithm from properly optimizing across your entire budget and audience to get you the best results for the lowest cost.
The power of Facebook’s app event optimization algorithm increasingly guarantees that micro-targeted ad sets with personalized creative almost always perform worse than more broadly targeted ad sets with more generic creative despite the higher engagement rate of personalized creative.
The holy grail of Facebook advertising is being able to personalize your ad creative based on a given audience variable without being forced to create separate ad sets across every audience variable. This would allow advertisers to simultaneously harness the power of personalization and Facebook’s algorithm.
And it is finally possible with Segment Asset Customization.
Segment Asset Customization
Using Segment Asset Customization, you can create a single ad set for your entire audience and then associate different ad creative with up to two of the following audience variables: (1) age, (2) gender, (3) location, (4) interests & behavior.
This means you can now personalize ad creative while still allowing Facebook’s algorithm free reign to optimize your ad serving to the subset of your audience likely to drive the most results for the lowest cost.
In order to succeed on Facebook in 2019, advertisers will need to shift their strategy to structure ad sets with broad audiences and deep-funnel event bidding while leveraging Segment Asset Customization to personalize their ad creative to the major psychographic and demographic subsets of that broader audience.
Feel free to reach out to me at firstname.lastname@example.org if you’d like to learn more about performance marketing on Facebook and Instagram
Flatiron Collective is a leading growth accelerator for consumer brands. We’ve powered Fanduel, Wag, Hinge, Vroom, Serena & Lily, Raise, Artsy & more.
* $425M+ in managed media spend
* $1.4B+ in venture capital raised
* 4x Webby award winner
* 3 exits in 2018