Online Path Team
15 September, 2026
Andromeda is part of Meta’s AI-powered advertising technology that helps determine which ads are most relevant to each person. Think of it as one part of the system working behind the scenes every time someone opens Facebook or Instagram. Meta has an enormous number of potential ads it could show that person. Andromeda helps narrow those options based on signals that suggest which ads could be most relevant. For businesses advertising on Meta, the technology itself isn’t really the important part.
What’s important is how it changes the way we build campaigns.
Traditionally, advertisers spent a lot of time telling Meta exactly who they wanted to reach. You might create different audiences based on age, interests, demographics and behaviours, then divide those audiences across multiple ad sets. The thinking was simple: the better we define the audience, the better the advertising will perform.
Meta’s AI is changing that.
Rather than relying as heavily on advertisers to decide exactly who should see an ad, Meta can use its own data and behavioural signals to help find the people most likely to respond. That doesn’t mean targeting is dead. Location, demographics, exclusions, customer lists and other targeting controls can still be important. But there is less reason to add restrictions simply because we can.
The focus is shifting from “How precisely can we define our audience?” to “What information can we give Meta to help it find our audience?”
And that’s where creative, data and campaign structure become particularly important.
So, what does that mean for the way you actually run your Meta Ads?
There are five key areas advertisers should be thinking about differently:
Let’s look at each one.
More targeting doesn’t necessarily mean better targeting.
Imagine an Australian home improvement business running separate ad sets for:
There may be significant overlap between those groups. Separating them also means separating the campaign’s budget and conversion data. A simpler campaign structure can give Meta more information in one place and more flexibility to determine where the best opportunities are. Instead of five small audiences, it may make sense to consolidate them into one or two broader audiences and allow Meta to learn which people are most likely to convert.
This doesn’t mean every campaign should use completely broad targeting.
A local Adelaide business still needs geographic targeting. A product aimed at a particular demographic may still need age or gender parameters. Existing customers may need to be included or excluded depending on the campaign.
The point is simple:
Don’t restrict Meta unless there’s a good reason to.
This is perhaps the biggest change businesses need to understand. Your ads aren’t just communicating with your audience anymore. They’re also helping Meta identify your audience. When someone stops scrolling to watch an ad, clicks it, saves it, engages with it or ultimately converts, they’re giving Meta another signal about the type of person who responds to that message. Meta can then use those signals to help determine who else may be interested.
Consider an outdoor furniture retailer.
Its potential customers might include someone who has just built a new home, someone renovating their backyard and someone planning an outdoor entertaining area. Previously, an advertiser might have tried to create three different targeting groups to reach them. Now, the creative itself can do more of that work.
One ad might focus on furnishing a new home. Another might show a backyard transformation. Another might focus on entertaining friends outside this summer. Different messages attract different people and create different signals for Meta.
So rather than only asking: “How do we target different audiences?”, we also need to ask: “How do we create advertising that speaks to different audiences?”
There’s an important distinction here. More ads don’t automatically mean better advertising. Creating six versions of the same graphic, changing the headline slightly and calling it “creative testing” isn’t necessarily giving Meta much new information.
What matters is creative diversity.
That means testing genuinely different ideas. For example, your ads could address different:
You can also test different formats, including video, static images, carousels and Reels. The objective isn’t simply to produce more content. It’s to find out which messages resonate with different potential customers. Don’t just test different designs. Test different ideas.
Creative is only one side of the equation. If you want Meta to find more people who are likely to become customers, it needs reliable information about what a successful outcome looks like. That’s where conversion tracking becomes critical. Your Meta Pixel, Conversions API and conversion events help send information back to Meta about what happens after someone interacts with an ad.
If Meta is optimising towards landing page views when what you really need is qualified enquiries, you’re potentially teaching the algorithm to find people who visit pages, rather than people who become customers. Likewise, if conversions aren’t being tracked accurately, Meta has less reliable information to learn from.
Before increasing your Meta Ads budget, make sure you’re confident that:
AI-powered advertising is only as useful as the information you give it. Poor data doesn’t become better because AI is analysing it.
Meta gives advertisers a lot of numbers – CPM. CPC. CTR. Reach. Impressions. Video views. Engagement – They’re all useful for understanding what’s happening inside a campaign but none of them necessarily tells you whether your advertising is working for your business. A campaign could have a fantastic click-through rate and generate very few sales. Another could have a higher cost per click but generate substantially better leads. Ultimately, the metrics that matter should connect back to your business objectives.
Depending on the campaign, that could mean:
The question shouldn’t simply be “Are people clicking our ads?”
It should be: “Are our ads generating the right business outcome at an acceptable cost?”
Consider a campaign with five separate audiences and two ads in each. That’s ten different combinations competing for budget and generating their own sets of data. For some advertisers, a better approach may be one or two broader audiences supported by several genuinely different creative concepts. Instead of dividing the budget across numerous audience segments, Meta gets more data within a consolidated campaign and more creative options to match with different people.
That doesn’t mean there is one “correct” Meta Ads structure. The right setup depends on your budget, objective, audience, offer, available creative and how much conversion data you’re generating.
That’s an important point.
Meta becoming more automated doesn’t mean campaign strategy has become automatic.
Across the Meta campaigns we manage, we’re increasingly focusing on simpler campaign structures supported by stronger creative testing and reliable conversion tracking. Rather than creating numerous small audiences simply because the targeting options exist, we’re looking at whether that segmentation actually improves performance. We’re also placing greater emphasis on testing different creative concepts rather than repeatedly making small variations of the same ad. The objective isn’t to hand everything over to Meta and hope its AI works. It’s to give the system better information to work with, while continuing to monitor the results and make strategic decisions based on what the data tells us.
You don’t need to rebuild every Meta campaign tomorrow.
But there are three areas worth reviewing.
Are you splitting audiences because your performance data says you should? or because that’s how you’ve always structured Meta campaigns? If multiple ad sets are targeting very similar people, there may be an opportunity to simplify the structure and consolidate your data.
Look beyond how many ads you have running. Are you testing genuinely different messages, customer problems, benefits and motivations? or are most of your ads variations of essentially the same idea? Your creative strategy increasingly plays a role in helping Meta understand who responds to your business.
Finally, make sure Meta is learning from the outcomes that actually matter. Check that your conversion tracking is accurate and that you’re optimising towards meaningful business outcomes, rather than simply the easiest actions for Meta to generate.
Meta is making more decisions about who sees an ad, where it’s shown and which creative is delivered. That can make campaign execution look simpler but it doesn’t make strategy less important. If anything, the role of the advertiser is shifting.
Instead of spending most of our time manually defining audiences, we need to make better decisions about campaign structure, creative strategy, conversion tracking and what we’re asking Meta to optimise towards. The advantage isn’t knowing every detail of how Andromeda works.
It’s knowing what signals to give Meta so its AI has the best possible chance of producing the result your business actually wants.
If you’re unsure whether your Meta campaigns are structured for the way Meta works today, Online Path can review your targeting, campaign structure, creative and conversion tracking to identify opportunities to improve performance.
📧 Book a Call to discuss your ad strategy
📊 Get a Free PPC Audit and find out where your campaigns could be working harder.
Contact us for an online marketing solution to build your brand and grow revenue.
Contact Us
1300 055 000
