This is the full transcript of Facebook Ads Are Falling Apart (The Andromeda Update), published on YouTube by Dr. Matt Shiver. Every paragraph carries the moment it was spoken, so you can click any line to jump straight to that point in the video, search the whole thing for a word, or copy it out.
0:00Facebook made a massive algorithm update in July. And I'll tell you what, it has been wrecking havoc with business owners who advertise on Meta, including Facebook and Instagram. And in this video, I'm going to break down exactly what the change was and how to combat this so you're not left behind. All right, so let's go ahead and get started. I'm going to share my screen real quick. Okay, so the update was on Meta Andromeda. That is the AI algorithm that Meta uses to place the right ads in front of the right people. Now, this is a post from Engineering at Meta. This happened on December 2nd, 2024, but it
0:38didn't fully unleash or like give it to all advertisers until July. That's why I saw some wacky things in July. I saw cost per click go through the roof on some accounts. I saw it drop like crazy for others. I saw cost per lead do the same. We had some people change strategies completely from DM ads to uh book a call funnels and vice versa because the results just were kind of wacky. Okay, now this engineering one I'm going to I'm going to summarize this. This is kind of more about the AI algorithm and the updates to it. Uh it talks about it partner with Nvidia and
1:10uh other AI developments. Honestly, I read this and it wasn't that helpful. Okay, so I found a uh another one that I think is more userfriendly. So, I'll link this down below, but at a high level, like what Andromeda is doing is Facebook realized that their issue was that they had a hard time with ad retrieval. And that's essentially means they had so many ads. They had so many advertisers. They had a hard time putting the right ad in front of the right person. So, what we used to do is we used to make generic hooks. And the generic hook would be like, "Hey, if you're um a busy mom trying to lose 20
1:44pounds, here's the ad, right?" So the generic hooks worked because Facebook had a hard time with retrieval. So those bigger hooks that called out a larger amount of people actually worked better. But now what we're seeing is Facebook wants more creative diversity. So it's actually those big hooks that appeal to a lot of people aren't getting the uh cost per click, aren't getting the cost per lead, aren't getting the cost per sale, and aren't scaling like they used to. So let me let me break down this article down here. I think this one was was written very well. And what I'm going to do is I'm going to break down
2:13this article and then I'm going to share with you guys on my iPad exactly what we're doing from a campaign structure, number of adsets, number of ads, all of that at the end. So if if you're a nerdy media buyer like me, uh stick around until the end. Okay, so again, this will be linked in the description below. The biggest thing, like I mentioned before, is the problem was retrieval. Okay,
2:34Andromeda is the new algorithm, right? And it takes millions of ads, narrows them down to a few thousand candidates, and it does in a way that Metac claims is faster, smarter, and more personal than ever before. And again, like I mentioned, creative diversity or diversification is your new ad strategy. Okay? So, creative diversity does not mean just pumping out near duplicates with minor tweaks. It's about creating a range of ads that differ in meaningful ways. So, previously, as an advertiser, I would have, you know, uh, one I'd make a video ad. I'd have one body and I'd have three different hooks and then boom, there's my three ads. Well, Meta
3:10is saying, "Hey, we don't just want that. We want more diversity." So, what we're seeing and what what I've heard a lot of other uh people in the space seeing as well is that having the ad all the way through being unique and different actually performs better than having just three different hooks. The
3:26same thing applies to static image ads. If you've got a white background static ad and you have a black background static ad, that's just a kind of minor near duplicates or minor tweaks. Again, this creative uh strategy and diversification that Andromeda is telling us to use, it does not want us just to do that. Okay? It wants us to make more. Okay? And again, I'll share
3:46with you guys how to do that at the end. But creative diversification shows up in a few areas, right? Concepts and angles. So, this is like what is the ad about? What's the pain point? what's the what's the hook angle that's bringing people in? It wants diversity in the formats. So, it doesn't just want videos. It wants vid short videos, long horizontal style videos, static images, and carousels. And then personas speaking to different customer types. Um, I'm going to talk about this in in when we get down to here. This is going to actually
4:16I would ignore this for the time being. If you have different personas, we're going to actually need to break those out into separate adsets or campaigns. I'll talk about that again at the end, but I think this kind of actually confused it a little bit. So, I'd say the biggest thing here is that it wants more concepts and it wants more formats so it can find the right ad for the right person because now it has that capability to do that. Okay. So, what this means now is we need to stop chasing that perfect combo of the ad copy and creative and feeding just one video ad and having one hook and one
4:48headline that that day and age is over. Okay? I'm telling you right now, if you just want you don't don't want to use the algorithm and you just want to try to be smarter than the algorithm, you were not going to win. Meta makes billions of dollars. I think I believe billions, probably tens of billions, I don't know, but they make a lot of money. Okay? And their algorithm has so much data points on you and on all of its users. Do you really think that you're going to beat it? Do you really want to sit behind the dashboard and just pump, you know, put $10 into this
5:13ad and put $10 out of this ad and just go back and forth? That sounds exhausting. Do not do that. I'm telling you right now, instead what Facebook wants you to do is instead of trying to chase that combo, it wants you to load it up. It wants you to load the campaign up with multiple different ads, multiple different headlines, multiple different
5:30body texts so it can find the right one. Okay, targeting largely happens in the ad level. Now, I've been talking about this for a while now. Um, this is just the way that it's going almost even now. If you are in the campaign setup and you're in an adset level and you try to target uh like a very tight-knit group, it's going to go well beyond that. So, it's not very good nowadays to uh have a very tight audience. Facebook doesn't even let you do that. So, again, the ad itself does most of the targeting. The words you say in the ad and the body copy that you have in the ad are going
6:04to put this ad in front of the right people. Facebook wants us to keep using the Advantage Plus defaults. Okay, Advantage Plus campaigns are the cleanest fit. If you guys have been in the ads manager, you've seen there's the opportunity score. It is really pushing for Advantage Plus. Okay, Advantage Plus campaign budget, Advantage Plus targeting, advantage plus creative, all this stuff. Okay, we need to learn how to use this. Okay, there are certain still some settings that I have turned off, but I am working more and more on how do I give this algorithm more uh leeway. And the the only way that that's going to actually work for me and for
6:40for you and everyone else here is if we ha if we're optimizing for the right event. And I talk about pixel conditioning a lot, but we have to have a really healthy pixel that's only sending back qualified leads if we're going to follow this plan. If we're just going to optimize for leads or clicks or DMs, we're going to get a bunch of junk leads, okay? Because it's going to send us volume. So, we have to learn how do we filter out and create more friction in our ads and through our process to
7:05allow the advantage plus to work for us. Because if we're optimizing for leads or messages, we need to get quality leads coming in, not just total volume of leads. If we don't specify that, that's what we'll get. Okay. And number four is simplify your structure. Okay? There are very few reasons to over complicate campaign structure. And you you need a different campaign for each goal, but
7:25you may may not even need that anymore. You will rarely need multiple adsets for targeting and you'll rarely need multiple adsets. I think that's and you'll really I think they just did a duplicate here. Uh the heavy lifting happens at the ad level. Okay, so this is what I'm going to talk about at the end cuz I think this I spent so much time this past week trying to figure out what's the best campaign structure and I'm going to share it with you from what I found. You know what the crazy part of? I I had a call with my meta rep. She had no training on this. She's like,
7:53I've been learning myself. I've been trying to get other people to learn about it, but they're not they haven't made an announcement about this, nothing. So, if you're just, you know, talking to your meta rep or you have like a a business coach who's not in on this stuff or marketing coach and still teaching the old same old strategies, you're going to be left
8:09behind. All right, so keep going here. Now, ad limits. So, there was a time where Meta would only suggest six ads per ad set. Uh there were times where we only said, you know, one ad per ad set. Like back in the day, we would just force one ad in front of one audience to test it out. Nope. Uh, we then went from
8:25there to three and then it went to six. Now, now that recommendation has disappeared. I know people who have 50 ads in an adset and it's working. Okay. Now, this is where it gets complicated and that's what I'm going to show you at the end. It's going to be super cool. But, um, yeah. So, if you guys want to read the rest of this, let's go ahead and read it. But, I'm going to share with you guys now on my screen, which I'm very excited about, the campaign structure of how we're going to be testing this. Again, um, this is all new to us, too, right? This happened in
8:52July. We didn't know about it really until September. Not know about it. We we let things ride for the month of August. It's September. I made changes to it last week. So again, there's a lot of changes that are happening and as we learn more, we'll share more. Okay. So let's jump in. Okay. So previously I'm
9:11going to say the old way right here. Okay. So this is the the old way of running ads. We had a campaign. Okay. And then we had multiple ad sets inside. Okay. I'm going to call ad set one, add set two, add set three. And often times the old old way is we just had one ad inside. Add one, right? Add one, add one. So we would do this because we were trying to figure out what's the best
9:35audience to serve these ads in front of. So this was like kind of step one. Once we found the best ad set, then the step two would be a very similar structure. So I'll just put C for campaign. Then what we would do is we would do potentially even like you know yeah ad set one that was let's just say this was the winner ad set one same thing ads set one these parentheses just mean adset set one ad set one and then we would even launch just like individual ads so it was like add one add two add three and you know we were just trying to
10:07force a specific ad in front of a specific audience. Now, just to be clear, this old way was like 20 uh 2019 through like 2022ish way. And then in 2023, things started to change a little bit. They started to come out with some of this more advantage plus stuff. And we made some changes. And then here is kind of the the more traditional way that I see a lot of people run ads now. So oldish way, all right, the oldish way was we'd have a campaign, okay? And then inside of here we had ad set one, ad set two, ad set three. Okay, so these are just like
10:47multiple different adsets inside. And what we would do, we found worked really well was we had the campaign budget set at the campaign level. So if I put like, you know, $1,000 a day in here and I said, "Hey, I'm optimizing this for calls." What I would used to do in this old way, oldish way was um each of these adsets here, right, were like proven audiences and I I typically would launch like a new concept. Okay, a new concept in each of these. Okay, so each of these new concepts could be like, hey, I'm going to do a B-roll ad. All right, so it's a B-roll ad and that's just let's
11:23just say that's ad set one, right? This is to a proven audience and this would be B roll one and then there would be B roll two and then there would be B roll three right and let's say ads set two was a drawing video right so I'm using my iPad drawing one drawing two and then drawing three right and then you know same thing for this this could be like talking head right so I'm just say talking head one talking head two talking head three so this was how I used to run ads all the way up until this point. This was I'm going to say
11:55oldish way. This is yeah, let's just say up until before Andromeda. So, let's just go 20 202 23 to now as in before. Okay, my handwriting is beautiful. All right, so this is what we were we're running. And this was super simple because all I had to do was just every time I had a new concept or a new idea, I would just launch a new ad set with that concept inside. And if Facebook liked it, it would spin on it. If it wouldn't, it wouldn't. Well, all that has changed now because as we just saw out here, the issue that Facebook was having was figuring out which of these
12:25ads to be served to what person. Now with Andromeda, this is all changing. So now the new way, okay, so this is the new way to run ads is this. Okay, we have campaign. Okay, and let's say this this is for a uh let's just say it's a VSSL campaign. So this is like, you know, you're running ads to a free training so people would learn about what it looks like to work with you, your process, and you sell them at the end, right? to VSSL. Okay, VSSL. And then we do this one ad set. And then we
12:54have 10 to 50 ads in that one ads set. And this ad set is broad. Okay, broad. And it really wants you to use Advantage Plus. Now, again, we're we're working on on doing this. If you guys have have been in there, they've actually added value rules to this recently. So, what I've been doing, I've been using Advantage Plus audiences, but I've added a value rule. Okay, I've added a value rule for this adset. So there's it's
13:20broad targeting. There's no interests. There is a minimum age set at the age of 25. 25 age. And then what I've done here is I put a value rule on this. And the value rule is to decrease bid by 90% for I'm just going to put it this way for old people. Okay? cuz I found that people I I don't know I I run you know learn how to run Facebook ads but there is people 55 65 plus who are engaging on my ads and I was like you know I want to give Facebook you know disadvantage plus audience but I don't want any of these
13:54old people to opt into my stuff. So what I'm doing here is I just added a value rule and uh you can decrease the bid by 90% for old people or whatever age range you want. You can do the same thing for gender, you can do the same thing for location, etc. But this is what I'm currently testing right now to um to combat that. You know, I could probably still just go in and turn off, you know, go like limit further reach of ads on the adset level. And then when I do that, I could just set age range and turn off the suggestion. But I'm giving
14:24this a shot now. Just wanted to share that. But this is the new way. And we're just going to launch ads on a consistent basis inside of here. Now, the problem with this is the learning phase. Okay? It's something that we used to be worried about, right? because every time I add a new ad into my ads set, it re-enters the learning phase. And if the learning phase decreases performance or it makes it reset the whole thing and you have a winning campaign, that could be a problem because what I would do with this with this concept here is I would just do new concepts. I would do
14:55this like two to three times per week. Yeah. Pro more like most of the time two, right? So I got a new new set of ads I worked on that morning or you know I was in Canva, made some ads. Go to the ad set, launch them. Go on the ads set, launch them. But the problem with that, this worked fine because the learning
15:10phase happens at the ad set level. Okay. But if we're launching ads inside of a new ad, existing ads set, it's going to reset the phase. Where over here, if I just launched an ad in a new, like I launched a new ad set, sure the learning phase would start, but it wouldn't affect the older campaigns. Now, that's not the case. So, when I launch the ad set, it resets the learning phase. So, what I'm going to do, at least this is the plan, is instead of dropping the new creatives two to three times per week, I'm going to drop creative one to two times per month, right? One to two times
15:47per month, drop 10 to 50 ads inside adset. Okay? And then once I drop them and I publish them, right? So, if I drop them and publish them, let's just say I've got 50 ads at the start of the month that are that are live. 50 start and then every day I check that campaign and what we're looking for are ads that are just not spending. Okay, so ads that are not spending means Facebook doesn't like them. So, you know, let's just say if I if I started with 50, this looks like so uh I started with 50 to start at the end of week one maybe I've got 40
16:20left, right? And then the next week goes by, then I'm down to 30 left. And then maybe I'm down to 20 left of like the ones that are spinning and are performing decently. Right? So then when I'm working on my drops, I'm going to look at these 20 and say, what's the common theme? Is there anything that I think I could do better? Right? And then are there new angles that I could also hit with this? So I'm going to just put like first would be, you know, creative refresh. And this is again finding
16:48winners and make more. Okay, make more. And when I say make more, just like we talked about earlier, like it wants creative diversification in formats. So if I have a winning video, can I turn that into a carousel? Can I take that same same winning creative video and make it into a static? So I'm going to do that and I might even just take the same script and film in a different area. So the first thing I'm going to do is if there's 20 left, probably about like let's just say I'll probably make creative refresh, I'll redo about 10 to 20 that are creative refresh already
17:16proven angles, right? And then the other part would be maybe 10 I'm just say 10 let's say 10 new angles right 10 new angles two 10 different types of ads that could hit on different pain points um and test those out. All right so again that month would drop and then I would have 50 ads to start and I would just keep doing that. So it would go 50 down throughout the month to you know 20 to 30 and then again re replenish what I I've cropped off. Okay. So, hope that makes sense. Um, that's what I'm going to be doing. And have you guys heard
17:50what I was talking about there of like I'm going to cut ads when they don't get any spend. So, because the way this is set up with one advantage plus and 10 to 50 ads in here, notice there is no top off ofunnel. So, no topofunnel slash middle ofunnel slashbottomofunnel split. Okay, that doesn't exist. So with this new strategy, what we're seeing is that uh inside of the ads set, the ads work together, right? So you might have a top offfunnel ad, right? Then you also might have a middle ofel ad and then you
18:25also might have a bottom off ofunnel ad. Again, this is creative diversity, right? So, like if we look at this setup, we put all the ads that we have in here because with this new algorithm update, what what we're noticing is these ads here, they work together. Okay? They work together where previously when I had, you know, this ad set one, ad set 2, ad set three, and they were all launched together. They were all independent. Sure, I'm sure Facebook shared some of that data back and forth, but not that much. Right? So now when we have this set up here and we're launching all these new ads, you
19:00might say like, "Holy cow, that's a lot." We can actually have ads inside of here that are more top offunnel, right? Top offunnel would be things like your personal story, right? Your personal story. Hey, here's how I got into this, right? I used to be 20 pounds overweight. I tried this thing. Blah blah blah, right? That we could even have a value, you know, a value ad with a soft CTA, right? And then middle of funnel is going to be more direct with, you know, click this link, get this thing, more direct with CTA for lead magnet. And then bottom of funnel would be more testimonials. Okay? There's
19:37there's a lot of other ways we could do this, but let's say they're all driving still to the same funnel. I'm not going to get into uh you might there might be an opportunity or an argument here where we'd have different funnels inside of these campaigns as long as they're optimized for the same thing. I'm not testing that yet, but that might be a
19:51play here. Okay. Again, pretty crazy. But I I would still drive like again, all this would be to like a VSSL funnel, right? So like top of funnel, personal story, longer, talk about your experience, then drive to VSSL value ad. Hey, here's three ways to do this thing. If you want to see more, click this link. Middle of funnel, hey, if you're struggling to get X, Y, and Z and you want this, go click this video and watch this thing. More around the pain and problem. bottom of funnel, you know, here's how I help so and so do X, Y, and Z. Go watch my VSSL, right? So, they're
20:25all, as of right now, my recommendation is that we're all driving to the same uh lead magnet. Again, after a little bit of time, that may change. Okay, so this is the campaign structure that we're going to be testing. One campaign, one ad set, very broad, and just keep launching new creatives in here. And again, we're going to do that a frequency of one to two times per month to uh prevent the ads set from going
20:48into the learning phase too many times. Okay? You can have as many ads as you want in this. Again, the biggest thing here is creative diversity. Okay? We're not doing this super controlled way anymore. It just doesn't work. This way can still work. Okay? This can still work. I'll be honest and transparent. This one could absolutely work. It does help you organize things better, but there's probably going to be a performance decrease because these adsets here are not working together
21:13like they are in our new uh our new way. All right, so that's what I got in this update, guys. I hope you guys enjoyed this video. If you guys like this and you want more updates on Facebook ads, you want to learn how to run them, all the things, go ahead and smash that subscribe button. I create daily content just like this to help you grow and scale your business online. All right,
21:33peace.
No line in this video contains that word.
Where these words come from. This is the caption track YouTube holds for this video, written automatically by YouTube rather than by the creator. We read it, tidied the line breaks and laid it out so it can be read. The plain text version is at https://viewrankai.com/tools/youtube-transcript/Yz0z1K1aysw.txt.
All rights in this video belong to Dr. Matt Shiver. Watch it on YouTube. If this is your video and you would rather this page did not exist, tell us and we will remove it.