How The Facebook Ad Algorithm Actually Works (And How To Beat It) transcript

Most people running Facebook ads think that the algorithm is working against them. Ever since Andromeda, they're going down an endless cycle, constantly recreating ads, building new structures, restarting, restarting, restarting, and ultimately seeing a decline in their return on ad spend. And after helping over a hundred million-dollar businesses run Facebook ads across my agency, I've noticed that there is one key factor that is killing ad accounts.

Across almost every account that I audit, people are working against the algorithm instead of working with it. So, by the end of this video, you'll understand how the Andromeda algorithm is actually better than the old algorithm. You'll also understand the two things that the algorithm is optimizing for, and you're probably not going to expect one of them. You'll fully understand how avatar targeting is the new targeting, and you'll walk away knowing exactly how the learning phase works so you stop the endless cycle of pausing and starting your ads again and again.

Every time the algorithm changes and everybody freaks out is the most opportune time to take advantage. When they zig, we want to zag. Okay, everybody complains about the Andromeda algorithm. I want to show you a few examples of brands that we've scaled tremendously in this era. And I'm going to compare this against the hardest period to compare against, which is Black Friday. So, in this account right here, this is an enormous account.

We scaled this brand from previously driving $800,000 in revenue over a 6-month period to driving $8.9 million in revenue. Like, let that sink in. 6-month period, up 987%. Or take this next example. We're up 126% in total ad spend. We increased ad spend $1.9 million since the algorithm dropped. All the while, return on ad spend has stayed flat. We were able to more than double this business while keeping return on ad spend dead flat.

And if those examples were way too big for you, then take a brand like this where we've only spent $30,000 over the course of a full 6-month period. Year-over-year, this brand is up 31% in return on ad spend all the way up to a whopping 24.4 return on ad spend. And the first thing you need to understand if you effectively want to use this algorithm to your advantage is what it actually optimizes for. So, inside of our ads manager at the campaign level, there's a setting called campaign bid strategy.

This is typically selected at highest volume or value, and some people use cost caps or cost per result goals which aim for efficiency metrics instead. Now, we're going to focus on highest volume and value. The highest volume and value bid strategy focuses on two things, and it literally tells you those things. It focuses on volume, and it focuses on value. Now, what is the volume and what is the value? We'll get to in a second.

But as a core principle, you need to understand that the algorithm cares about just two factors. First, can it drive scale, aka volume? And second, can it do that as efficiently as possible to drive you as much value back for how much you're spending? Now, quick PSA, this is why you often see ads in your account that spend a whole lot of money that don't actually return. It's because the algorithm thinks that that ad is scalable.

So, it thinks that there's something there. It's checking the box for volume. What we ideally want is ads that both scale in volume and of course scale with efficiency, meaning they drive both volume and value. Those are true winners. So, when we move to the ad set level, our performance goal is going to have a direct impact on the algorithm as well. Now, for our performance goal, we're basically looking at two different options.

First, maximize number of conversions, and second, maximize value of conversions. Now, I'm not going to go through these in depth right now, but what we need to understand is that these impact the value side of the campaign. Let me explain. At the campaign level, when we're selecting highest volume or value, this second half, this value here, is matching whatever we put in this performance goal. So, the value that we are caring about is either number of conversions or value of conversions.

Now, value of conversions typically means more revenue. Number of conversions means we want as many of those conversions, as many people to buy as possible. So, for most brands, you're going to be running maximize number of conversions, but if you're running value of conversions, know that that is just impacting the second half of the campaign level setting. After that, we are choosing our conversion event, which of course has a major impact on the algorithm.

See, the algorithm only optimizes for whatever conversion event you are choosing. So, if you have custom conversion events, it needs to learn, and then it's going to focus on driving that custom event. If you are targeting something like an initiate checkout or search or an add payment or add wish list, it will push to that active event. This is why in 99.99% of cases, we focus on the main primary purchase event. That's because not only does the algorithm and standard events have a ton of information from all the businesses it's ever worked on, but it also is focused on the one thing that you want, which is driving purchases.

So, when we put this all together, it's actually quite simple. You're choosing your campaign bid strategy, which focuses on two things, volume and value. You're then connecting your value to whatever we are choosing in our performance goal. So, that's either going to be number of conversions or value of conversions. And then again, all of this is connected to your conversion event, which in our case is going to be the purchase.

So, in a setup like this, the algorithm is just looking at two different things. First, it's looking at all of your ads and if they can drive volume, meaning can they spend and can they scale? And then it's particularly linking that volume to if they could drive the maximum number of conversions of the purchase event. Now that you understand the two things that your ads are actually optimizing for and that the algorithm is looking at, we now need to address how targeting works.

You've heard a thousand times over creative is the targeting, but to be honest, that is way too broad. It doesn't give you any direction when you're actually creating your ads. So, here's the framework that I like to use that you can copy. So, when people say creative is the targeting, what they really mean is that this is an avatar-based system. What that really just means in simple terms is that whomever you're talking to or calling out in the ad itself is who is going to be naturally targeted by that ad.

For example, if I call out older men or younger women, those folks are more likely to be targeted in the actual audience. Combining a avatar with a specific angle, usually known as a problem and solution, creates what I like to think of as a concept. When you have an avatar plus an angle together, you have a concept. That concept does the targeting for you by pushing problems and solutions to problems to a specific audience.

So, this is the official Andromeda released graphic. And what we really just need to understand from this is two simple layers. First, you upload your ad creative over here on the left-hand side. That ad creative then goes into a personalized ad retrieval system. Long story short, your ads get matched to the most appropriate candidate that is going to do two things. First, they need to be called out in the avatar just like we have here.

And second, that candidate needs to have a history or at least a likelihood of actually acting on your conversion event based on the millions of signals that Facebook has. Now, we understand the core of the algorithm, but what actually happens when you launch your ads? How do you get out of that cycle of constantly uploading ads, pausing ads, uploading ads, pausing ads, what I like to call the death loop. Basically, what happens and what most people are terrified of is the learning phase.

Now, you just need to understand how this works and this will pretty much eliminate all of your fears of the learning phase. It will allow you to scale faster, more aggressively, and you won't be afraid of pulling back when it's the right time. The learning phase is a mechanism built into the delivery system of meta and as you can see on my screen right here, we have dozens of ad sets that are learning limited. This isn't the worst thing.

It doesn't mean they can't perform. It doesn't mean they can't spend. It doesn't mean they're screwed in any way, shape, or form. All it means is that they are not driving as much steady performance as everything else. Now, historically, the learning phase has been a measure to basically tell the advertiser not to pause it down. Facebook is not sure that it's going to drive a consistent result. Basically, when you get to the active stage of delivery, your results tend to be more consistent.

When you're in the learning or in learning limited, you still need to wait an adequate amount of time to get to active until stability occurs. This doesn't mean that those ads are good, bad, or anything in between. So, let me show you what happens when you launch a new ad. This is our typical marketing funnel. So, at the bottom here, this is where the purchase occurs. Once you make it past this very bottom, the money is actually transacted.

Now, what this basically means, there's a couple stages to every funnel. There's the bottom of the funnel, there's the middle funnel, and there's the top of funnel. So, we're just going to put top of funnel, middle funnel, and bottom of funnel and then this is where the money is made here on the bottom. So, typically in the bottom of funnel, you are product, problem, solution all aware. Meaning, you understand the brand, right?

You are brand aware. In the middle of the funnel, you are typically product aware. So, think of it like this. If you're in the middle of the funnel, you know that you need a a pair of sneakers. You just don't know what brand you're going to buy yet. That would be the brand aware section. And the top of the funnel is typically where you're not product aware, you're not brand aware, you might be problem aware. Like if you have back pain, you might know you have back pain, but you might not be aware that there's products and brands to solve that problem.

So here, you are simply problem aware. Now at each stage of the funnel as you work down, you're more likely to buy because you're problem solution aware, product aware, and brand aware. Meaning you are very, very close to the purchase. Now what happens naturally as you spend more money on Facebook is your first dollars go to the very bottom of the funnel. So we'll just use this circle here to represent the very first dollar that you spent.

It's going to be really, really close to the bottom of the funnel. And as you continue to spend more money, you're going to start to work your way up the funnel higher and higher. This is typically why advertisers hit a wall. Now, this isn't necessarily a bad thing. Most people think of this as a very, very bad thing because they see their return on ad spend going down. But what you're actually doing is you're capturing huge chunks of the market earlier and earlier.

So instead of framing this as, "Oh, this is always a negative, always a negative, always a negative." Think of this as you're finding new customers who are never going to hear about your brand otherwise. Let's say you get to this stage where you're generally hitting true top of funnel, right? These people aren't even aware there's a product to solve their problem. This person here represented by this dot is going to take literally more time to convert than this person right here.

So your learning phase or the fluctuations you see when you scale your ads are often just having to do with how many impressions you're driving at the top of the funnel. The next impression or the next dollar you spend is higher in the funnel than the previous in most cases. And because of that, those impressions just take more time to convert. Right? This last dot compared to this dot right here, this dot's journey is generally going to be shorter, while this dot is going to be longer.

And because of that, we need to deploy patience when we are in that learning phase. Also, we need to understand that increasing our budgets by just 10% all that's going to do is add one or two extra dots here. If we increase massively, we are going to add a lot more dots. Yes, maybe at the tippy top of the funnel, but if we just deploy some patience. Literally, if we just wait, then this group here might eventually convert if we are patient enough.

By the way, if this is clicking right now, I just wanted to give you a heads-up that I have shared this in my school community months ago. So, if you want all the earliest access to the models, the frameworks, and even our internal SOPs, you can go to school.com/facebook. There's always a link down in the description. Or if you just want someone to handle your ads, you want to be hands-off so you can work on different parts of your business to actually grow, then you can go to the moonliters.com/apply.

Again, there's always a link down in the description. That is all for today's video. I hope you got a ton of value out of this. I love doing whiteboard sessions like this. So, if you did, please let me know in the comments below. And I'm going to put on my screen right here my exact campaign structure that we're using for 2026, 100% for free if you'd like to go there as well. That's all for today. Thanks as always for watching.