Did Andromeda Ruin Meta Ads Forever? (Let's talk...)
By Dara Denney
Summary
Topics Covered
- Andromeda Isn't New or Selective
- Gem Reads Customer Minds
- Lattice Unifies Models
- Creative Diversity Wins
- Merge Organic and Paid Teams
Full Transcript
Here's the thing about Andromeda. You're
probably thinking about it all wrong. I
see a lot of people blaming bad performance on the new Andromeda update.
And I'm going to keep it 100 with you. I
have a lot of opinions about this update that I think are going to actually make you uncomfortable. So, what we're going
you uncomfortable. So, what we're going to do today is I'm actually going to give you the clearest explanation on what these updates actually are. And
there are three in total. It's not just Andromeda. Then I'm going to explain why
Andromeda. Then I'm going to explain why these updates may have contributed to bad results, just not in the way that you think they are. In fact, I'm going to be calling you out a little bit. But
before we begin, my personal hot takes on these updates and why I think it's really important that you clearly understand them. Number one, these
understand them. Number one, these updates have rendered all targeting interests look alike, even prospecting versus retargeting pretty much useless.
You'll understand why I explain more.
And number two, it's also made your account structure far less important.
Kind of a big one, right? So, let's get into these updates. So, there are three new updates in total, right? There's
Andromeda, Gem, and Lattis. And let's go into the most infamous first. Andromeda.
Actually, let's first go into what it's not. It is not a new algorithm. And it's
not. It is not a new algorithm. And it's
also not new. It was actually launched last year in December of 2024. So, it is not responsible for your poor ads performance lately with an asterisk
here, right? Because I can explain why
here, right? Because I can explain why some of you might have worse performance now as a result of these changes. And
finally, Andromeda doesn't actually control which singular ad is shown to each individual user. Let's dive into this. So, Andromeda is part of the
this. So, Andromeda is part of the algorithm that actually decides which ads are selected for delivery. It
narrows down millions of ads from thousands of brands to a smaller pool that can be shown to each individual user. Simply put, Andromeda is the
user. Simply put, Andromeda is the retrieval process for gathering relevant ads for each user every time they sign on to Facebook or Instagram. That is it.
Andromeda doesn't even technically select the best individual ad at any given moment. That is actually another
given moment. That is actually another update's job. Now, the way that I like
update's job. Now, the way that I like to think about Andromeda is this. It's
kind of like a personal assistant that you've tasked to help it find you the best perfect white t-shirt. So, what
that personal assistant is going to do first, right, is based on how well they know you and what type of events and travel you have coming up, they are going to narrow down a selection of a
few options for you to choose. And that
process is all that Andromeda does, but just for ads. Now, before we dive into the gem and lattice updates, which are really crucial for fully understanding this new algorithm, I first want to give
a huge thanks to the sponsors of this video, Particle. They are longtime
video, Particle. They are longtime friends of ours because they play a really critical role in my own research process, which honestly is now doubly important with all these new updates.
With Particle, you can see real-time sales data from the top DTOC brands. And
I particularly love it to be able to zoom in on what type of products and bundles are actually selling today.
Something that I'll explain more in depth later in this video and why some of you might have had worse results after the gem and Andromeda updates. And
again, not to spoil the plot, but being able to increase your AOV so you can actually spend more to acquire your customers is going to be a crucial part
of your success moving forward in this new ecosystem. So, some of my favorite
new ecosystem. So, some of my favorite bundles that I've seen using Particle recently that have honestly really given me a few bursts of great creative inspiration for this have been Hexclad's
Ultimate Everything Collection, which honestly seemed to crush this Black Friday, Cyber Monday. And similarly,
Magic Spoon also has an everything bundle. So, when I saw these, actually,
bundle. So, when I saw these, actually, I immediately reached out to some of the clients that I was working with and I'm like, "Hey, do we have an everything bundle?" Cuz if not, this seems like a
bundle?" Cuz if not, this seems like a really easy thing that we could do to really increase our AOV. Now, the reason why Particle ties in really well with Andromeda and Gem is that you're going to be able to get a much better sense of
your competitor's revenue and sales by skew. And they also have centralized
skew. And they also have centralized access to tons of marketing assets from your competitors like social media ads and even emails. So, if you really think that Andromeda is the reason why your
sales tanked, I would actually take a look at your competitors and particle and you might actually see, hm, they're doing something a lot better when it
comes to their bundle strategy, their product strategy, and how they're positioning those products across their entire marketing ecosystem. Okay, let's
get back into these updates. Right. So
after Andromeda narrows down the pool of ads, that's when the other new AI models like Gem and Lattis come into the picture. So Gem is the generative ads
picture. So Gem is the generative ads recommendation model. And after
recommendation model. And after Andromeda selects this pool of ads, right? That's when Gem comes in. And Gem
right? That's when Gem comes in. And Gem
is essentially a superbrain. I actually
really loved the way that this Meta employee described it. He described it like a new superbrain that can read an entire library of book in seconds while
understanding and retaining every single detail. So, the way that this comes to
detail. So, the way that this comes to life in Gem is that it can essentially read the mind of your customers and figure out what they're going to want to
do next week, maybe even before they do.
So, Andromeda creates a pool of ads for a user and Gem narrows that down even further as sort of the super brain power. And then that's where Lattis
power. And then that's where Lattis comes in, which is essentially a massive library of all of the AI models of the algorithm. And this is where GEM
algorithm. And this is where GEM actually lives, right? It's a part of Lattis, which is significant because in the past, Meta has actually had all of these different models separated, which
is why it really benefited us to have interest targeting, lookalike targeting, and segmenting things out between prospecting and retargeting even. But
now with Lattice, the exact right ad can be matched up to the right user. So it
is actually Lattice that selects that final ad, not Andromeda or Gem. And
let's go back to my metaphor, right?
Let's bring this full circle. Andromeda
is like the personal assistant who will narrow down a pool of white t-shirts that you might want to buy, all from different brands. And Gem is the
different brands. And Gem is the superbrain that selects an even smaller pull of white t-shirts based on even more detailed insights on you, your behavior, and the advertiser goals. And
it knows what types of fabrics you like.
You know, cotton versus linen, for instance. And lattice is the system that
instance. And lattice is the system that brings everything together. It is the one that selects the final t-shirts based on all data and outputs from Gem and other models. So to put it really
plainly, Andromeda and Gem are simply the systems that Meta created so that they could sort through millions of different ad combinations for each individual user. Candidly, it broke that
individual user. Candidly, it broke that into steps so that Lattis could make that final ad selection. Okay. So what
can advertisers do if results are down?
How has this affected advertisers in the last few months? Because I have heard from a lot of people, Andromeda killed my results. And that's not exactly true,
my results. And that's not exactly true, but there is a reason why your results could have gone down as a result of it.
And to be honest, it's really simple.
Your competitors are likely taking better advantage over everything that Andromeda has to offer. They're likely
out bidding you. So, they have a bigger budget to acquire a customer, and they're a lot more strategic on their creative strategy, aka they probably have more better ads than you do.
Nothing more and nothing less in my opinion. And think about it. If your
opinion. And think about it. If your
competitors doubled down on their creative strategy and really focused on their creative operations and making sure they could get more better creatives out the door and you just kept
the status quo, yeah, Andromeda and Gem are much more likely to select your competitor's ads versus your own if you just kept things business as usual.
Because Andromeda and Gem really does reward brands who are supplying great creative volume, but more importantly, creative diversity because it makes
their jobs a lot easier to select relevant and engaging content, which again, I don't want to say that volume's the answer. Volume is not exactly the
the answer. Volume is not exactly the answer. It's much more about your
answer. It's much more about your creative efficiency and your velocity.
How quickly can you get excellent creatives out the door and into the ad account? So, I always tell brands this.
account? So, I always tell brands this.
This is what you can do. Number one, you need to know your core personas. I
actually really like to call them audience segments. What are the
audience segments. What are the different audience segments that you can develop creative for? Motion actually
has a really great intended audience tag that I really love to be able to like bring this to life a little bit more, especially if you think you have only one. And for those segments, you need to
one. And for those segments, you need to develop a diverse range of creative formats. And when you find certain
formats. And when you find certain formats are top performers, that's when you really need to start doubling down on different messaging angles and even different awareness levels so you can
give the algorithm as much manpower as possible. And for those of you that are
possible. And for those of you that are overwhelmed, I absolutely hear you. This
suddenly becomes a question of not how to make the best ad possible, but it's more about your creative operations and how quickly you can act on those high quality creative ideas. So, my final
thoughts for this, a really interesting wrinkle to all of this is this trend or desire for more brands to put out top offunnel content, which candidly is just
great organic content. The gap between organic and paid content in my eyes is really starting to close as we move into 2026. And for brands that really
2026. And for brands that really prioritize unique organic content, I think they're going to be in the best position to pump out those new creative angles and [music] ideas for a wide
range of personas. I've even seen a few operators say that they were going to be combining their paid social and their organic social teams into one, which I think is a sign of things to come. And
that's all. I will see you guys next week. Let me know if you want a video on
week. Let me know if you want a video on how to develop your creative strategy more for Andromeda. I think that's going to be really important.
Loading video analysis...