
How to Scale Facebook Ads Using Artificial Intelligence
Scaling Facebook ads past a certain point can become extremely challenging. AI intelligence can be a meaningful catalyst for growth.
The Growth Paradox
At present, marketers are led to believe that identifying and targeting specific audiences on Facebook is a simple task. The acceptable best practice is starting off with very broad targeting, and being heavily reliant on Facebook in order to hone in on distinct audiences.
For example, Tim Burd, the proclaimed ‘Godfather’ of Facebook Ads, said at Affiliate World Conference: “... I like to do a minimum audience size 1 million. Ideally 2 million plus and even if it’s 70 million, that’s good, but 1-2 million minimum. I never go less than that.”
The classic theory is that Facebook knows its users and segments them into what we simply describe as ‘buckets’, each bucket possesses different attributes and a different intent to take action.
We allow Facebook to take a selection of users from different buckets and optimize them over time towards our desired results.
By relying on audiences who have taken a valuable desired action, marketers can essentially grow their target audience and user base.
And indeed, this process of identifying high-value, high-intent customers, as well as replicating lookalike audiences and then scaling Facebook ads based on these insights, works very well. However - and for this reason we have gathered here today - at some point the plateau flattens. Customer acquisition costs go up and ROAS goes down. Broad audiences and lookalike replication becomes less effective.
The question is:
How to break through this plateau and identify new, highly-specific audiences?
At this stage, the trigger finger response of most marketers is to increase ad spend. But this doesn’t solve the problem of increased ad spend: while it may enable you to reach more audiences, your customer acquisition costs will still be on the rise. So, the real answer lies not in increasing spend - but rather in targeting more specific segments with the right messaging and user intent.
In other words, at this point, you need to restrategize. Marketing is a dynamic ecosystem in which both campaigns and audience segments continuously peak and ebb like the tide. This means that part of your ongoing strategy needs to be gaining new ground across the board -
- new audiences
- new messaging
- new offers
You can't expect the same tactics to create value ad infinitum - otherwise, we'd all be out of a job.
So how do you meaningfully scale Facebook ad campaigns?
In order to break new ground, your best bet is to work off of segments that have performed well for you in the past, and then go boldly where you have never gone before by testing how they perform after changing specific attributes. This can be done on many levels: by targeting the same segment in other geos, by altering your offer or messaging, by adding life events, etc. Here are some examples:
Refine audience segments - A good example of this is targeting customers with a specific interest such as fitness & wellness.

When you combine this with a life event, say, new parents, then you are able to find a highly refined audience segment, i.e. mothers trying to get back into shape after having a child.

Tweak audience location - In this instance, you can take a well-performing ad set and simply change the geographic location you are targeting. Facebook actually allows you to include up to 25 individual countries as well as 250 specific cities in a single ad group. Use this flexibility to test and make an educated guess on where your next target audience may be.

Diversify your offering - This is a tactic taken out of the eCommerce playbook and can be very effective. For example, if you are currently targeting people interested in buying treadmills, you may want to consider trying to sell them a related item such as running shoes, a water bottle or even sportswear.
Basically, what you’re actually doing is testing and validating new secondary audiences - a great tactic to use once you have maxed out on lookalikes. It involves identifying your best audience segment based on transaction value, lifetime value, engagement and order size, and changing one or two audience characteristics. It is in this way that you can create an infinite number of potential target audiences who are ‘very similar’ to the audience you want to replicate.
The tactics outlined above can bring immense value, and help you scale Facebook ads past the point where you are currently stuck. The major issue, however, is that discovering, testing and validating new audiences or campaigns is hard. Since in essence you are working on instinct, it takes a lot of resources both in terms of time and money to validate your hunches. You need to define the new segment, create the campaign, run it, optimize it, and wait for results. This is a long, iterative process, that has the potential to suck the will to live from the marrow of your bones.
Scaling Facebook Ads Using AI
The crux of the matter is of course that there is an almost endless number of attribute combinations that can make up your high potential audiences. That’s why it’s no surprise that AI and ML solutions, that can deal much better than humans with vast amounts of data and data combinations, have been brought on the scene. AI-driven audience discovery solutions can quickly and efficiently assess multiple variants of your audiences to predict which ones will give you the highest performance. In this sense, using AI audience intelligence to scale Facebook ads is both low-cost, and time-efficient.
At Novarize, we use AI and ML to pinpoint high-potential audiences, along with all their relevant data, for every brand - and even for every product category. By taking into consideration the brand, the products, and the competitive landscape, Novarize scans the open internet to surface specific audiences with a high probability to convert. The tool’s audience analysis and discovery engines work constantly to find relevant audiences as they evolve, so that not only does your targeting stop being a guessing game - it is also always timely and in sync with your audiences’ current development.
To enable optimal engagement, Novarize gives you a deep analysis of each segment, including:
- Interests
- Income levels
- Media consumption
- Social media activity
- Brand affinity
- User intent
- Sentiment
These insights can be put into practice immediately. So, when you optimize your Facebook ads for data-driven user intent, sentiment and brand affinity, for example, you are able to acquire new audiences by appealing to consumer’s most current wants and desires. This is a great solution for marketing practitioners who do not want to sink their entire marketing budget on audience testing.
Our goal at Novarize is to empower marketers to make an impact on behalf of their brand by getting to their most valuable audiences quickly and efficiently. By minimizing the resources usually geared at discovery/testing/validating cycles, Novarize helps marketers make data-informed decisions that actually increase ROAS and generates growth.
Try implementing Novarize’s artificial intelligence capabilities so that you can scale your Facebook ads quicker and smarter!
Topics: