The Truth About Meta's Andromeda Update (What You Need to Know)

0:00So, the Andromeda update is what everyone is talking about in the Facebook ads and the Instagram ad space right now. And I think we need to talk about it, too, because there is a lot of fear-mongering and a lot of misinformation flying around about Andromeda right now. So, in this video, I'm going to go through what Meta actually changed, the impact it's actually having on ad results and the changes that you need to make to your current ad setup to make the most of Andromeda, and to make sure you don't get left behind. So, to help explain what the Andromeda update actually is, I've got an article from Meta up about

0:27this. I'm going to run through various sections. Actually quite a complex technical doc, but I think we can break it down into sort of easy to understand uh terms and focus on the parts that actually impact us as advertisers and and why we need to update and change things in response to this update. The first thing I want to highlight though is the date. So this was posted by Meta on December 2nd, 2024. So almost a year

0:50ago as of the recording of this video. And I think that's quite interesting. And the reason why it sort of taken a while and it's only really become the big topic of conversation recently is because it took Meta a while to roll this out. So it took as Meta often does with with new features and new updates. Um some advertisers get it early, others it takes longer. They also introduce this to some campaign types before others. But now we're in a situation where the vast majority of advertisers

1:14are operating in an Andromeda um system. That's why it's become a big topic of conversation, particularly from those advertisers that haven't updated how they do things. They may have seen results drop off. They're still operating under the the old setup. And so, a lot of people talking about it. And like I said, we need to cover this, too, cuz there's a lot of misinformation

1:32and fearongering about this out there. So, there's got sort of three summary points here. I'm going to skip over those because we're going to cover those in more detail as we go through it. I actually want to start with these two paragraphs here. Right. So, we've got here, AI plays an important role in Meta's advertising system by leveraging the power of machine learning to predict which ads a person will find most interesting. And that sounds like a fairly standard run-of-the-mill sentence for an ad platform, right? Like, we're using AI and machine learning to try and put the [clears throat] right ads in front of the right people that they're

2:01more likely to engage with. Um, but that's really key to Andromeda because Andromeda is doing this in a much more sophisticated way to what has been used previously. And I really want to sort of focus in on this predict which ads a person will find most interesting angle. You'll see why that's important as it comes up. And they sort of justify why that's that's important. It says this helps people learn about a business or product they are interested in while helping an advertiser meet their objectives such as increasing brand awareness, acquiring new customers, and driving sales. Obviously, if your ads are put in front of people that are more

2:30likely to be interested in your products and services, you are more likely to make a sale, generate a lead, etc. Okay, this next paragraph is also very interesting and revealing. So, retrieval is the first step in our multi-stage ads recommendation system. And that's what Andromeda is. It is a much more sophisticated ads recommendation system that operates differently and has consequences for us advertisers. Why we need to update things and change some of our best practices. At this stage, it's tasked with selecting ads from tens of millions of ad candidates into a few thousand relevant ad candidates. In the following stage, larger more sophisticated ranking models predict

3:04people and advertiser value to determine the final set of ads to be shown to the person. So if you think about if you're targeting the whole of the US for example, how many other advertisers are targeting the whole of the US? Tens of millions if not more of potential ad candidates and that needs to get whittleled down from all those different ads, all those different advertisers because so many people are advertising nationwide broad targeting um or or a

3:27variation of to really narrow it down. Okay. We'll we'll remove 99 whatever percent of the various options and then from there we'll narrow down even further narrow down further to work out which ad is most likely to be most interesting to that user but also most likely to meet the objective that the advertiser wants. Right? So that's kind of an overview of what Andromeda is and and how it operates. But we need to focus in on a few elements here because

3:54those are those are really important. Okay. Okay. And the meta is kind of setting the scene with this next section talking about challenges of opportunities in this new era of advertiser automation with generative AI. Um, and basically they're highlighting two points. I won't go through all the detail here, but highlighting two points that because advertisers are using AI to produce ad creative, uh, not just outside of platforms with various AI tools, but also within platform to create lots of variations of ads. So if you think about using advantage plus creative where you meta may add text overlay to your images or animate them or may add music or may

4:31uh you there's all sorts of different enhancements within advantage plus creative. Well those are all different ad variations and you can have tons and tons of different ad variations from just the one sort of seed ad that you've created as an advertiser. So using that stuff as well as the external tools to produce more ad creative basically means there's a explosion in the number of ads in the number of ad candidates. So you can see one of the big challenges of the AI era of advertising is going to be instead of meta having to sort through millions of ad variations it's going to have to sort through hundreds of

5:02millions probably billions of different ad variations um as that sort of grows in an exponential way. So that's that's the first big challenge that Andromeda is designed to to cope with, but also to allow us to benefit from. And then the second big challenge here is tight latency constraints. Basically, Meta's got to do this really, really fast because if you're midscroll or you're swiping through stories or whatever it is, that next ad wants to have been sorted through all the various different ad candidates and worked out just the right one so that it can be put in front of them straight away. There's no sort of loading, buffering, etc. And you also

5:33can take advantage of changing user interest and focus and attention really quickly. This has the potential to get really complicated. I'm going to try and keep it simple and bring it back to the parts that matter for us advertisers. You know, we're not developers, right? We're we're we're advertisers. We want to get better results. So talking about deep neural networks and using the super GraceHopper super chip and deliver superior performance. But it's this first sentence again in this paragraph that I think is interesting. So Andromeda improves performance of meta ad system by delivering more personalized ads to viewers and maximizing return on ad spend for advertisers. That's the name of the game

6:05here. If we can have a more personalized ad delivery system, we get better results. If you think about this is is what Andromeda is trying to do is be a much much more sophisticated version of what we used to try and do manually as Facebook and Instagram advertisers many many years ago when we try and be ultra specific with our targeting and match the messaging in our ads to the target audience because that's been a known marketing tactic that works since forever. Right? If you just talk to everyone, you talk to no one. If you talk to people very specifically and it resonates with them and you call out

6:34their pain points, they're more likely to listen to it, to take action, to believe that your product or service will work for them. And that's what this is looking to do. It's looking to do this at a massive scale and get things really personalized for the individual to uh help improve performance. So, let's move on to the next section. So, we've got Advantage Plus automates budget allocation, audience targeting, and bid adjustments, streamlining, campaign management, and boosting performance through more ads in the system for different audiences. So, we've seen the effect of that for a number of years, right? As Meta has been able to automate uh things like

7:03targeting uh budget allocation, bid adjustments automatically, we have gotten better results. It's why more and more advertisers have switched to Advantage Plus campaigns or switch to open targeting that takes advantage of some of this stuff. That's what that first part's talking about. For example, when advertisers who did not previously use Advantage Plus Creative turned on its AIdriven targeting features, they experienced a 22% increase in return on ad spend from our ads. We estimate the business using image generation are seeing a 7% increase in conversions. So, this next part is kind of crazy. Even at this early stage, more than a million advertisers use our generative AI tools

7:39to create more than 15 million ads in a month. Now, like I said, this article was released almost a year ago. I guarantee that number has increased massively and more and more people are using the tools like I said within meta also externally but this is really focusing on the stuff within meta to produce way more ads way more ad variations they're using advantage plus creative um and taking advantage of that. So this effect is only going to continue to increase uh particularly as well as advantage plus creative and those enhancements that you see become more and more sophisticated. And then after that it gets really complicated

8:13with all sorts of diagrams and more talk on neural networks and more talk on hierarchical indexing and stuff that honestly is too complicated for most advertisers certainly too complicated for me gets very technical and just not stuff that I think we really need to worry about. Um that's more developer world than it is advertiser world. So if I had to sort of summarize Andromeda and a good way of thinking about it, I think it's like Andromeda is like Meta building out its own Netflix recommendation system. So instead of everyone just being shown the same films or TV shows to watch within comedy, uh Netflix is going to show different

8:48options first to different people based on what it thinks they are most likely to be interested and watch based on previous viewing habits and things like that. That's that's what Meta is doing. So previously of course the system before Andromeda still worked out these are the best types of ads to put in front of these people [snorts] but it was done as sort of more of a group level. So okay we've grouped these people as interested in ABC. Therefore they meet tick the boxes and meet the requirements to be shown ads from advertisers XY Z. Whereas now with Andromeda we're getting much more specific to the individual. So, it's

9:20more sophisticated, more personalized. Uh, like I said, just like a Netflix sort of comedy playlist recommendation system where not everyone sees the same thing. Some people are shown different things um to others, even if they are both of those people are categorized as being interested in comedy. So, in a second, we need to talk about why Meta created Andromeda and the changes that we need to make as advertisers to our Facebook and Instagram ad strategy to get the best results now that Andromeda um is here. But first, I want to quickly let you know about my mentorship program, which gives you all the support you need to get the best possible meta

9:51ad results, including multiple daily live calls with me and my team. Uh, unlimited one-to-one support, ad account audits, exclusive training material, like I said, all the stuff you need to get the best possible results. If that sounds interesting, there is a link in the video description. Go ahead and click on that and book a call with one of my team members and hopefully we get a chance to work together directly very

10:08soon. So, why did Meta create Andromeda? Why? Why? Why? Yeah, I've heard all sorts of ludicrous things said uh from people saying that Meta did this because they want to worsen results and hurt advertisers. Like that is absolutely not the case. Meta is highly incentivized for us advertisers because we are their customers to get better results because if we all stop getting any leads and sales from Meta tomorrow, we would all leave and Meta would stop generating revenue and the value of the company would crash. Like that's absolutely not what they want. They want us to get the the complete opposite. In fact, they are highly incentivized to get us

10:39advertisers the best results possible because then we keep paying them money and if they get us even better results, then we spend more money with them, we scale, they grow as a company. So, just wanted to sort of dispel that myth because I've seen that knocking around quite a bit. So, why create Andromeda? Well, the real key here is personalization. We've already talked about that quite a bit, but putting more relevant ads in front of people makes it more likely they will take the action that us advertisers want them to take,

11:02and they will also like it too, right? People don't mind advertising if it's for things they are interested in as much as they do if it's for things they are not interested in. Okay? So it helps improve performance for us advertisers and keeps the user happy. There also a big time component to Adromeda. So um trying to do it faster, deliver the right ad to the right person at the right time. And that time element also makes it more likely that you will get

11:25the result you want as an advertiser. Cuz if even sometimes in some industries if your ad is delivered to someone u who's the right person, but you got the wrong day. They've already bought the such and such thing that they were looking for and researching and you could have got an ad in there, but it

11:39didn't quite work it out fast enough. Then you could be missing out on return on ad spend and on sales because of that. Like Meta had personalization previously and they had systems to be able to do that relatively quickly, but people were grouped as opposed to being narrowed down to the specific individual. And yeah, that basically left rorowass on the table in Meta's opinion. And we can see that because when we tried to group people manually with our own targeting options and serve them ads, we didn't get as good results as we did when we went broad in most cases for most advertisers over the last

12:10few years. So that's a good example of how this system can work uh better by instead of grouping people making it more specific to the individual. And then of course Andromeda was to some extent a necessity. They had to update the system because with AI and automation creating way way way more ads than we've seen previously. Well those all need to be sorted through and organized. Okay. So for some advertisers

12:31that leads to an obvious question right? If Andromeda is supposed to improve performance why have my results gotten worse? And just to be clear not all advertisers results have gotten worse. There are many advertisers that are seeing similar results to what they had pre-Adromeda or have even seen an improvement in results, right? But there'll still be advertisers that have seen a drop off that be thinking, well, if this thing's more sophisticated and designed to get better return ad spend, why am I not seeing as good a results in in my ad account? I think there's a few things to factor in here. So, Andromeda has only really become the big topic of

13:00conversation in the Facebook and Instagram ad space fairly recently, and that could easily coincide with a seasonal decline. If your business typically performs better, say during the summer or struggles in Q4, that could easily explain it. It could be that CPMs cost per thousand impressions are always way up in Q4 because there's a lot more advertiser activity with Black Friday and and Christmas and that's when advertisers spend a lot of their budget. A lot of advertisers spend a lot of their budget. So, those things could easily be a factor. You could have changing market dynamics. Those are things that are kind of out of your control. The big thing that is within

13:36your control is have you updated your strategy and your way of doing things to factor in the Andromeda update because if you haven't that could easily be causing a big drop in performance. Andromeda is something new. There are differences and the best practices that go with it have also updated. Now what that means is that that's a great potential opportunity because if you had great results and came in, they dropped off. If you then adjust your system and your setup to try and take advantage of Andromeda, you may well see better results, not just than what you're getting right now if they have dropped off, but potentially even better than

14:14you did previously because Andromeda is a more sophisticated thing to do. And then also, if your competitors haven't adjusted and haven't um gone with the new system and made those changes, which a lot won't, then that's extra potential competitive advantage for you. Okay. Okay, so in a second we're going to run through what you need to change, but first I want to let you know about Hyros. Hyro is the best tracking and attribution software for meta ads. It's a software all the big players in this industry use, including myself. I use it to track results for my own ads as well as all my organic results as well. So if

14:43someone books a call for our done for you services, for example, I can see exactly where that person came from with a lot of accuracy. That allows me to make better decisions for my marketing and my business. High rose has been able to track that from our ad spend in this campaign we've generated 96,000 but £58,000 of that was not reported by meta. So if I was just looking at my data within my meta ad account I might come to the conclusion that this campaign is doing okay but in reality

15:09this campaign is doing a lot lot better. Now this is a relatively extreme example because of the nature of our business. So most of our products services have recurring billing elements. Meta might be able to see what we make initially from our ad spend on that initial transaction, but not those subsequent payments. That's why it it shows the importance here of being able to accurately see what we actually get from our ad spend in order to be able to make

15:31better decisions and optimize correctly. So, if you want to improve the performance of your ad campaigns without having to touch your ads, go ahead and sign up to Hyros. There is a link in the video description. Okay, so what do we need to change now that Andromeda is it? How do how do we get the most out of the new system as advertisers? How do we update our strategy? All that sort of stuff. I've got some recommendations for you. And this increased personalization is is at the core of this because that really creates a big opportunity for us advertisers in what we can do. And I'll

15:56I I'll explain what I recommend you do and talk through the logic of it as well. So my biggest recommendation is an increase in volume of ad creative. More ads is better. Meta used to make recommendations around don't have more than six ads in an adset at any one time. Now we're regularly running adsets with 20 ads or more in there. If you think about what Meta is able to do, they're able to work out of those 20 ads, let's say, this one is right for this person and this one is right for this person. So that increased level of of specificity when it comes to

16:29personalization means that more ad creatives gives Meta more flexibility to put the right ad in front of the right people. Then within that increased ad volume, ad creative volume, you also want to try and increase creative diversity. So not just different ad variations, right? not just changing the image on a background, but different ads with different angles that focus in on different things and that different people might be interested in that might grab their attention uh more than others. Right? So, if you think about creating lots of different ads, you want to try and hit different things. And these will vary a little bit depending on what it is you're selling, but I've

17:03got a little list I'm going to run through. So, you could create different ads that try and hit people at different stages of the awareness ladder. So, the awareness ladder in marketing, you're probably familiar with it. You start people that are sort of not even problem aware, completely unaware. And then ads that are targeted, those people will often be like, "Did you know that such and such can be a real problem?" And then you go through the various other stages. And then you've got sort of problem aware, so they're aware of the

17:26problem but not aware of the solution. So then you have ads that explain the solution, but you don't need to explain the problem in the first place because they already know. And then you have ads to people that are solution aware, where you don't de you don't explain the solution more broadly, but you talk more specifically about your product or

17:39service um to get them to to buy that. You can create ads thinking, okay, does that apply to my industry? Could I create some ads that are explaining the problem in the first place, create other ads for people who already know the problem but don't know the solution. And and then all the way down that that's a good way of thinking about, okay, how do I create ads for those different different options? You could create ads that touch on different pain points, right? Maybe your product or service solves different issues. Maybe your competitors are more expensive and more timeconuming. So you could have one ad that really focused on the fact that

18:09you're cheaper, another ad that really focused on the fact that um people can save time, lots of different pain points, lots of different uh industries. Have a think about those. And then you can also flip that, right? And different desires. Does someone want to save time? Does someone want to save money? Does someone want to buy your product or service because they're going to look really good in front of their friends when they've bought it? Does someone want it because it's going to be relaxing? Does someone want it because it's going to be really exciting? All sorts of different desires that you can think of. Not all will apply to all

18:34products or services, but some will apply to some. And you can look to use ones that might work and different angles because different people might be interested in different things. And that's what a drummer is able to do. Able to go, this person wants to uh mitigate issues. They're booking a fasttrack queue at an airport, for example, right? So they don't have to stand through all the cues going through security. One person might be booking

18:54that because they want to save time. Another person might be booking that because they're an anxious traveler and they really don't want the risk of missing their flight stood in a queue for security. Another person might buy it because they want to feel swish that they're just walking straight through to the front of the queue. Those are three different desires for the same thing and different people will be interested in different ones. Like a business person is probably more going to be interested in the time saving. An anxious traveler and frequent traveler is going to be more interested in not missing their flight etc etc. Okay. And then different

19:20avatars within your target audience. So there are likely to be within your overall target audience different groups of people and you can speak to them specifically. Imagine advertising a gym, right? You're going to have people different ages. They're going to be different fitness levels. They're going to be coming from a different place. Are they trying to lose weight? Are they trying to gain muscle? Are they trying to get in better shape to compete in

19:41some sort of sport or athletic event? Are they doing injury prevention or injury rehab? You could create different ads for all of those different use cases and different avatars. Like older people are more likely to be interested in X. Therefore, we can create ads that are more suitable for them. And again, another good way of thinking through um the different ads, right? So, in general, I get this question all the time now. How many ads should I be looking to put in my ad campaigns? And my my advice here is as many highquality ads as you can produce. Don't necessarily run them all at once, particularly if you're operating with a

20:18small budget. Meta still needs to be able to test the ads and see which can perform best. And that's still a process that's going to occur. But the vast majority of people I speak to, they are constrained by how much good stuff they can produce, not by what Meta can handle nowadays with Andromeda. That's a big big change. That wasn't the case even a few years ago. What I see a lot of advertisers doing is they hear, "Oh, I should have 20 ads in an adset." And then they just go ahead and create 20 ads every week. But the only way they can create 20 ads every week is if

20:46they're basically variations on a theme or, you know, they're just swapping out something really small or they're just not very good. Don't do that. Don't create volume for the sake of volume. Uh well, actually, no, you do want to create volume for the sake of volume because that helps. But only do so if the ads you're producing are high quality and if the ads you're producing have different angles, hit different

21:07pain points. Talk to different desires. Talk to different people at stages of awareness. Uh talk to different avatars within your audience and then allow Meta to be really smart and work out, okay, this person needs to see this ad. This person needs to see this ad. And another part that we haven't talked about or what wasn't expressly said within that document, but I've talked about because it's obviously being featured within the new system, and I've mentioned this on other videos, is meta being able to add sequence. And if you have more creative diversity, the meta can add sequence more effectively. What do I mean by that? Some people within your target

21:40market are basically ready to buy. There's no point putting a an awareness style ad, a top offunnel ad in front of that person. You might as well just go for it. Go for the sale. go for the the lead inquiry, whatever it is you want people to do, right? With a more direct bottom ofunnel style ad. But some people do need those top offunnel style ads. So again, by Andromeda being really smart, being able to personalize if you put ads that would previously have hit different

22:05uh stages of the salesunnel in together. Meta is smart enough to work that out, right? Some people going to need to put a top of funnel ad in, then a middle of funnel, then a bottom funnel. Other people straight for the bottom of funnel. If you put them all in together, Meta can work that out for you as opposed to separating out into different campaigns or different adsets and then forcing the issue by saying, "No, only advertise to these people." This is really quite a difficult thing to get your head around if you are if you're someone who's been advertising on Meta for a long time and you come from a

22:31multi-step salesfunnel process, tons of retargeting, manually set up retargeting. Um, this is very different. Requires, you know, you to trust Meta a lot more. But I've talked over the past few months about different campaign structures and about including different ads that hit different um stages of the funnel, hit different pain points, using hook variations to create lots of different ad creative relatively quickly and easily. And I consistently see on those videos people coming back with, "Wow, I changed over to the new setup and my results have gotten so much better. I've I must have seen that comment a hundred times or a variation of it hundred times um in the last month

23:07or so." And that's because yes, lots of advertisers have experienced big drop offs from Andromeda because they did not update their system, their strategy, their setup to the new setup. And if you do, if you play by the new rules as it were, if you give basically Andromeda and Meta's machine learning system what it wants, it can work for you as best it possibly can and help you get um significantly better results. Okay. And one of the big changes that we need to make to facilitate that to facilitate all this extra ad creative uh and being able to get the best results and allowing method to do its thing is a

23:45change in campaign structure. And I have seen some nutty campaign structures recommended by people in this space to help um take advantage of Andromeda or combat Andromeda. And it's quite clear whenever I see these crazy campaign structures being talked about, especially in relation to Andromeda that these people don't really run many live ad campaigns because they're completely impractical. They certainly don't have 1,200 plus ad accounts um in their business manager like we do. And we're running thousands of meta ad campaigns at any one time. We very much see what works in the real world, not just in theory. And actually with Andromeda, you

24:20want a very simple campaign structure. Probably not just the one campaign, uh but you want a very simple campaign structure. And I go through all of that in detail in this video here. And I think this is a logical next step to go ahead and watch this after you've seen this to see exactly how we're doing things. This is very much done with Andromeda in mind. It's what Meta wants us to do. And once you switch over, you are likely to see significant improvement in results. So go ahead and give this a watch