Leverage Machine Learning to improve your Facebook ads ROAS by 47%


In this article, you’ll learn how you can Leverage Machine Learning to improve your Facebook ads ROAS by 47%.

We’ve run the tests, analyzed the data, and the results are unanimous: even the most seasoned media buyer doesn’t stand a chance against machine learning.

The performance metrics are so skewed that we believe if your media buying strategy doesn’t include Meta’s AI tools, you’re wasting up to 50% of your ad dollars.

The best news is that scalability is possible when you transform your ad account and start utilizing every cent of your advertising dollars. 

No, we’re not talking about lifetime value scalability. We’re talking about lead products sold to new customers at profit scalability.

We’ve delivered some awe-inspiring results using the strategies we’re about to discuss. Here are a few highlights:

E-Commerce Facebook Ads Results
  • Reduced cost per sale by 75% within three weeks
  • Delivered returns as high as 26X – see above
  • Generated millions of dollars in revenue for our clients
  • Introduced thousands of customers to new companies and lead products

Let’s jump in, and we will show you how you can do the same for your business.

Facebook Vs. Apple: The great data war is happening right now.

Leverage Machine Learning to improve your Facebook ads ROAS by 47%

Facebook and Apple have waged an all-out war over consumer data. In the iOS 14 and 15 updates, marketers get radically reduced. Targeting criteria has been reduced, costs are up, and accurate tracking of user behavior is much more challenging. Combine this with the fact that the attribution window is now limited to 7 days, and the consumer can deny access to their data; performance marketers are on their heels.

Ultimately, this leaves marketers in the position to make decisions that are harder to justify, especially when you are looking at scaling an unorganized and bloated ad account. The alternative, you guessed it, Leverage Machine Learning to improve your Facebook ads ROAS.

We knew this change was coming, and naturally, Facebook responded by creating new tools and systems to get the most out of your ad dollars. While some of the tools we will talk about aren’t brand new, they’re continually being improved and are at the point where we advise our clients to run their accounts this way.

This guide will serve as your toolkit to fight the algorithm changes and maximize your Facebook results by getting the best data.

On average, we see a 47% improvement in ROAS (return on ad spend) across accounts when we implement our systems – and when it comes to Facebook ads, we believe this stack of tools is the most powerful.

(It’s important to mention that if you don’t have the Conversion API set up, none of this will be possible, so read this first if you haven’t done that yet.)

The Power 5 Stack For Facebook Advertising

Facebook Power 5 Stack For E-Commerce

Facebook’s most advanced set of Machine Learning and Account Optimization tools is the Power 5 Stack. 

In their words: 

“The days of manually hacking your way to ad success are no more. Top direct-response advertisers are now leveraging a specific set of automated ad tactics to unlock new phases for growth. We call these tactics the “Power 5,” and when used together, they can transform ad performance and scale across the Facebook Family of Apps.”

– Facebook

It’s a big deal. We believe that it may be the only way for you to properly scale and create sustainable ad campaigns – when done correctly, it allows you to push as many new customers to your business as possible.

Auto Advanced Matching

Auto Advanced Matching ensures that all available events from your conversion API and pixel are sent back to Facebook and match the customer data to the user.

Leverage Machine Learning to improve your Facebook ads ROAS by 47%

What happens is that additional information about your users is sent to your ad account, allowing Machine Learning to create a more accurate customer avatar. You’ll see improvements to your tracking because what Facebook does is match the data on your site to the information they have on their platform. 

Besides being able to register and attribute the conversions to the appropriate campaigns, you’ll also increase your custom audience sizes. When we track the users correctly – it’s easier to create custom audiences to remarket to them.

Most importantly, this will lower your cost per result by giving the algorithm the data it needs to show your ads to the most beneficial buyers in your audience.

We highly recommend you take Facebook’s suggestion and set up your Advanced Matching to be Automatic, you’ll get more data this way, and Facebook can take what it needs to continue to feed the Machine Learning.

Better data = improved scalability.

Account Simplification

Facebook Account Simplification

The job of a media buyer used to be managing tens or hundreds of campaigns, manually playing with the switches like an F18 pilot, adding more budget to performing campaigns, and rapidly shutting down campaigns that flopped.

It doesn’t work that way anymore.

The best way to optimize your ad account is by having no more than four campaigns and no more than five ad sets per campaign.

The campaigns we use (and recommend you use) are:

Prospecting: Top of funnel

  • Broad Audience 
  • Interest-Based Targeting
  • 2-5% Lookalike Audiences

Engagement: Middle of the Funnel (MOFU)

  • Reactivation of engaged brand audience
  • Facebook and Instagram followers and post engagers
  • Website Visitors
  • 1% Lookalike Audiences

Retargeting: Bottom of the funnel (BOFU)

  • Cart Abandonment


  • A place to test new creative content to see what works

(If you want to know more about how we set up our campaigns, click here to read our article discussing it in further detail)

By keeping everything simple and clean, you allow Facebook’s Machine Learning to auto-optimize for you 24 hours a day, in real-time. This sounded too good to be true on our end, so we tested it extensively over the last 12 months. The results are unanimous; the simplified account structure is a superior ad-buying structure.

Once you have your four campaigns created – the next step is to let Facebook determine what Ad Sets within the campaign deserve the appropriate percentage of ad spend.

Campaign Budget Optimization

Campaign Budget Optimization is the process in which Facebook determines at the campaign level which ad sets to allocate the budget. This automation takes the guesswork out of your hands so that you don’t have to assume which ad set will perform the best. 

Facebook’s AI will determine which ad sets get the optimal percentage of the budget and when to spend the budget so that your ads are delivered with the best opportunity to convert.

Facebook Campaign Budget Optimiztation

One of the most significant benefits of Campaign Budget Optimization (CBO) is that your ads are always learning, but they don’t get sent into learning mode. It used to be that anytime you made significant changes to your creative content, your ads were pushed back into learning mode, which increased costs and hurt deliverability. With CBO, adding new creative content to an ad doesn’t cause them to go back into learning mode. CBO saves you time and money.

Automatic Placements

When it comes to advertising, the key to profitable acquisition is all about putting the right message, in front of the right people, at the right time.

Media buyers used to make separate ad sets for each placement and then determine which were performing and spend accordingly. This system would often result in costs increasing dramatically or performance coming to a sudden halt.

By using automatic placements, you’re using Artificial Intelligence to determine the proper placement for the advertisement. It will analyze consumer behavior against your ads to decide whether to present your offer to your prospects on Facebook, Instagram, Messenger, and the Audience Network.

This will reduce your costs by finding the most cost-effective way to generate the right traffic to your site.

Leverage Machine Learning to improve your Facebook ads ROAS by 47%

Dynamic Ads 

Note that if you stumbled across this page and are not an e-commerce store or a business with a catalog, this section is not for you as you won’t be able to run DPAs.

By implementing the Power 5 Stack, the objective is to learn as much about your customer as possible and then place ads in front of them based on their actions on your website.

Dynamic Product Ads or DPAs may be the pinnacle of this system. They’re dynamic advertisements that are generated based on your website visitors’ behavior on your site. The products they showed interest in will be advertised back to them across Facebook’s network of sites.

When your friend says they think Facebook is spying on them, they are referring to this. And they convert like crazy.

The way it works is, when your customer visits your website, they’re tracked – and then what we do is send the data of their visit back to Facebook. Then strategic ads are placed before them for the exact product they just looked at. 

Want to see your sales volume go through the roof? Offer your DPAs at a slight discount.

Now, Leverage Machine Learning to improve your Facebook ads ROAS by 47% – and grow your business.

You now have the tools to implement Machine Learning in your Facebook Ads Account properly. Following these steps will improve your ad costs and help you create scalable, evergreen campaigns.

Want our help maximizing and managing your Facebook ad account? 

Don’t hesitate to reach out to our team. You can do so on our contact page or email us directly at: [email protected].

If you found value from this article and want to learn more about advertising your e-commerce business, here are a few more articles that will help you out.

Subscribe to our Newsletter

Want the best digital insights delivered right to your inbox? Then enter your email below. No spam. No BS. Just marketing tactics to help you grow.

We build, execute and optimize advertising systems that are seen by millions around the world.




Our Best Advice