I'm sharing with you today a podcast interview I did with Damien Plutino, founder of the ad acquisition agency ERO Corp. Damien is someone I've worked with for years on our own client acquisition; he manages millions of euros in ad budgets for his clients, and he's constantly reverse-engineering the algorithms to find what works.
In this conversation, we talk about Andromeda; the major update to Meta's advertising algorithms that happened last September, and what it means for coaches, consultants, therapists, and expert entrepreneurs who use paid content to bring in clients. We cover how the old ad model worked, why it's no longer enough, what a "social funnel" is, how many videos you actually need to produce, the specific mistakes that waste your ad budget, and a hands-on creative strategy you can implement right away.
Below the video, I've written a detailed article that recaps the main lessons from our conversation. If you prefer reading to watching, you'll find everything you need below. And if you prefer watching, hit play!
Historically, the algorithm was somewhat capable of distributing ads, but it was mostly guessing. The whole principle of online advertising is monetizing ad space on social media platforms. To maximize their profit, platforms set up a bidding system.
Before Andromeda, those bids were somewhat blind. The algorithm looked at three things. First, the advertiser's reputation; your profile score, whether you consistently produced quality work or had ads rejected. Second, user behavior with your content; did people click, did they stay a bit longer on the page. And third, external signals like the performance of the page you sent people to, including bounce rates; people who clicked but immediately left.
This allowed Facebook, Instagram, and the same applies to Google's platforms, to get a rough idea of how things were working. But it was a rough idea. The role of the media buyer back then; that's someone like Damien, a person who buys the media placements and manages the ad account; was to maximize the performance of all three factors for a well-defined audience. Make sure the right person sees the right ad, clicks, and has the right behavior on the next page. It was somewhat technical, but when you did the work well, you got consistent performance.
More recently, particularly driven by TikTok's influence, platforms realized they could analyze behavior much more precisely and analyze the content itself. They could match people's interests with how long they spent consuming each piece of content.
There's an organic algorithm called GEM that made this possible. And it's partly to feed data to GEM that Meta acquired WhatsApp. WhatsApp is free, but if it's free, you're the product. All the information you share with your friends somewhat naively; Facebook, Instagram, and WhatsApp (the Meta group) analyze it to capitalize on understanding your interests.
So there's a kind of enormous matrix, a giant table with tons of data about you. Some estimate sixty thousand criteria. You have a score for how interested you are in a given topic, personal attributes, and things like that.
Where the magic of advertising comes in, especially with the advertising algorithms and AI; where we really made a leap forward last September; is that now, armed with this information that they were already using to show you new videos when you scroll through Reels on your feed, they can now do the same for advertising content, with an extraordinary level of precision about your interests.
Thanks to GEM and Andromeda, you can now show advertising content to each person who might be interested in a topic, but also at the right level of awareness. You can present the right content at the right moment and create what the Americans call a "social funnel"; a slippery slide that starts right on the platform, before even sending people to your sales pages.
This is a massive paradigm shift for advertising. Before, it was very pushy; "jump in and trust me." Now, because people have become more skeptical after seeing low-quality content and abusive practices, the approach is to give value first, not just push for the purchase decision. It allows you to build the relationship and more naturally inspire trust in an era of online skepticism.
When most people think about creating ads, they think of an image, a text, or a video that says: "Buy my product, buy my coaching, here's what's in it and why you should buy." This type of content corresponds to what the Americans call the bottom of funnel; the end of the process, when people already know they have a problem, already know a solution exists, and you just need to present your specific solution and its terms.
But the statistic (which comes from Chet Holmes' book The Ultimate Sales Machine) is that only about 3% of your potential audience is at the decision stage at any given time. These are the easiest to convince to buy, but they're also the most expensive when you're bidding on ad placements, because the platforms know they're valuable.
The audience you can educate; the famous "nurturing" principle; is much larger. Through marketing education, you can take someone from a painful problem they didn't know had a solution, to understanding the criteria for choosing solutions. On that journey, you can reach almost 30% of people. After that, some won't have the money, for some it won't be the right time; their cat died the day before, whatever it may be. But instead of concentrating on the 3% who are easiest to convert (and most expensive), you can reach people earlier.
And because you're reaching them earlier, they potentially have no preconceptions; or better yet, you can give them the right preconceptions about your product, the right criteria that will make them prefer your product or service. I think at the heart of Andromeda, Meta said: "OK, to improve the user experience, we'll allow advertisers to educate earlier and stop all being in the 'I have an amazing deal, you'll earn two thousand euros tonight by doing nothing, click here' mode."
A lot of YouTubers and colleagues in the space say: "You need to maximize volume, produce tons of creatives." But the real principle of this algorithm is to recreate the buyer's journey. And one of the best ways to create content without burning out is to build what Damien calls a persona; what I call an avatar; a virtual profile of the buyer where you express their pain, desires, wishes, and their ideal solution.
From that persona, you can identify three dimensions for creating content:
1. Advertising angles. Talk about the problem, share a case study, address a specific objection around your service's promise. If I'm targeting expert entrepreneurs; coaches, therapists, consultants; I might make one video about the therapist who only treats one patient at a time and wants to help more people. That's not the same video as one for someone who wants to create their first online course. Each angle needs its own content.
2. Creative diversity by format. Renew interest by presenting the same message in different formats. Depending on your industry, images, videos, or carousels work well. Ideally, you create one of each; so the same advertising angle gets expressed as a face-to-camera video, an image, and a carousel.
3. Sub-personas. If you want to go further and scale, you study not just the persona of the whole population, but you go deeper. For example, if you're selling a dating strategy for Tinder, you don't just say "Optimize your Tinder profile." You specifically address the shy guy, the loudmouth, and so on; with a video that speaks directly to their specific situation. That level of detail is what makes the content truly resonate and what gets it rated highly by the algorithm.
Damien gave a powerful example from a client who sells gates; a local Belgian business. Just by increasing creative diversity, they cut their cost per lead in half.
Before Andromeda, Meta would listen to your videos quickly, transcribe them, and barely look at the visuals. You could make roughly the same face-to-camera video over and over. Now, if you show an ugly gate, then a beautiful gate, then the founder presenting the factory, then a showroom tour with different finishes; Andromeda considers each of those as different content. And based on the actual script, it knows where to place each video in the matrix. It might show the showroom tour more to women, the factory tour more to men (Damien's words; he apologized for the cliche). Maybe older people want to see the finished gate.
The point is: by varying the visual aspects, you reach the right people with the right content, even though they're all homeowners who need a gate. Damien calls this approach a matrix; a table with five columns and five rows, and you make sure to fill every cell so that every sub-profile can see what would make them take the next step.
Here's what happens when you don't understand this. Damien has seen clients send in a batch of eight videos. One video spends 80% of the budget, the second gets 15%, the third gets 5%, and the other five get literally zero; not even one cent. The algorithm looked at the first three and said: "You sent me the same thing eight times. I'm done."
Back to the drawing board.
So if you're planning to sit down with your phone and film fifteen videos in a row; "Hi coaches!", then "Hi therapists!", then "Hi consultants!"; those videos, visually, from the platform's perspective, are the same video twelve times. The platform won't show multiple versions to the same person because it thinks they've already seen it. You need to make them visually distinct; different locations, different outfits, different formats (interview, face-to-camera, walking in the street, whiteboard, facility tour).
There's a stat going around (Damien heard a different number at Alex Hormozi's event) about ten creatives in testing for every 10,000 euros of monthly budget. But what Damien's team observes is that this isn't quite accurate once you have your first complete matrix.
The real numbers:
If you can't do 25 right away, start with 10, then do batches of 5 regularly, iterating on the content to fill out the matrix. For budgets under 5,000 euros per month, staying around 10 creatives with a single persona and four or five advertising angles is reasonable and acceptable.
The only advice Damien can give is: grab your iPhone and a good wireless microphone. That's it. Whenever you have a background that's a bit different, whenever there's wallpaper that appeals to you, make a video; it gives you a different look.
Try to create B-roll; illustration footage. Film when you're in meetings with clients, record your Zoom calls like we were doing. Any occasion is good for injecting creative diversity into your content.
One of Damien's clients told him that a casual recommendation to buy an Osmo Pocket 3 was a game-changer. It's a gimbal camera with a wireless mic, it's stabilized, it has dynamic tracking that follows you in the frame, and you can use it as a tripod, a webcam, or on-the-go. For about $500, it does everything, and it pays for itself quickly.
The last point is to keep editing as simple as possible, at least in the short term while you're getting data. Don't go into elaborate shoots with professional video teams or spend nights perfecting the editing of your first creative. Solutions like OneTake let you rough-cut quickly and not spend hours in post-production. It's more important that the first ten videos go out than that they're perfect.
As a concrete example: when we launched the Cercle des Experts a Haute Valeur, I followed Damien's advice and filmed all the videos with my phone in a single day. One in the living room with my laptop connected to the TV showing the free tool. One on the terrace facing the sea one way, one on the terrace facing the other way (visually, two different things; same terrace). One down at the marina, one by the beach. Within a five-minute walk, I had several videos with different looks. Some had wind noise, some had background chatter from other people. But I didn't spend time editing; I put them through OneTake, got the edited videos with clean sound, and published all of them the same day.
With Andromeda, content now accounts for roughly 50% of your ad performance, up from maybe 20-25% before. But that means there's still 50% of the work to do. Once people click, do they bounce because your capture page isn't clear? Do they sign up? Is it the right profile?
The algorithm continues to loop between everything: it shows the content, people engage, they click, they visit the page, they sign up. Then it reinjects those signals, targeting a more precise pocket of prospects: does the message still work? Do people still click? And so on. It's an iterative loop that's essentially endless.
Content performance has become primary, but if your capture page doesn't capture, your campaign will run out of steam even if your creatives are good.
What Damien's team sees even more recently, from analyzing data and reverse-engineering the algorithm, is that the algorithm loves it when you have different pages for different sub-personas.
For clients with larger budgets, they create a different page per sub-persona, a campaign with a dedicated ad set and creatives using the matrix, all landing on a page that matches. This further maximizes performance.
But this is for people spending tens of thousands of euros per month. For smaller budgets, the advice is even more important: try to target one sub-persona from the start. That way, you get good ad performance, and as the Americans say, you become the biggest fish in the bowl.
The "YouTuber advice" that says "just produce tons of stuff and see what happens" is clearly the biggest mistake, especially when it comes to creative diversity. As we said, you need between 10 and 25 creatives fairly quickly, but it's better to make even just 5 that are very different for each stage of the funnel than to make 10 that look the same.
If you're making 10 videos with the camera in the same spot, same background, same t-shirt; small details matter. I mentioned filming in three or four locations. Change your t-shirt. Maybe change your hairstyle; we did something pretty catchy with my haircut once, and that's exactly the kind of thing Andromeda loves.
If you have several personas, segment, but not too much. If you've identified four sub-personas, only work on one. If your budget is limited; say 2,000 euros; that allows you to work one persona correctly. Don't subdivide and try to create 25 creatives across four personas.
Damien gave a specific example: a client who worked with government employees. By narrowing from all government employees to just teachers, they gained almost 30% in performance. They didn't do teachers, police officers, administrative staff, and civil servants separately. They did one, well. And they concentrated the budget.
If I go back to my old dating coaching business, with a small starting budget, I wouldn't target "love coach, solve love problems." I'd specifically say: "I'm talking to you if you're a recently divorced forty-something, getting back on the dating market, and you don't know anything about apps." A much more targeted audience. Why? Because it's better to reach 10,000 very similar people with an extremely precise message than to reach 10 each of a thousand different profiles with a super vague message.
This is something we struggle to do because of FOMO; the fear of missing out; which pushes us to say "I'll talk to a few more people, maybe I'll get lucky." But it's almost the opposite you need to do. The more specific you are from the start of your message, the more problems you solve; you can structure your pitch, structure your product, solve delivery problems. It's a win-win on every front.
As Damien put it: 50% of performance is your content, but there's the performance of the funnel and even the tracking; the data flowing back to the ad platform. Don't forget what happens after the click. The real formula is: simple structure + rich creatives + converting pages, all properly connected to the ad platform.
One statistic that shocked Damien: he's been working with one client for 15 months. In those 15 months, CPMs (cost per thousand impressions) have nearly quadrupled for that client's placement. But with the optimization work they've done; the creative diversity, the CRO improvements, the continuous iterations; they're almost at the same cost most of the time, and some weeks they've even been at more than 20% improvement.
That means they've managed to contain a potential 4x price increase. If the client had done nothing, or kept doing exactly what they were doing before last summer, the impact would have been devastating in terms of margins and costs.
This is also a reminder to diversify your channels. Meta is becoming more dominant, and if they can multiply costs by four, that's a lot of power. Build your own channels; email lists, WhatsApp groups, Telegram channels; so you can stay in contact with your audience without paying every time they see your content.
Damien made a point I really liked: advertising is a faucet. If your faucet leaks, not much water reaches the vase at the end. Treat your leads like a rare resource; because they are, and you paid for them.
The clients who do well over years are the ones who have respect for this resource. They don't just acquire; they cultivate the relationship and optimize revenue per subscriber. That's a startup metric we call LTV (lifetime value), and it's not used enough in the information product world. The entrepreneurs who succeed maximize it; not at their clients' expense, but by offering more value to their clients and growing their revenue in return.
By retaining and maximizing that revenue, you build stability. You can get through periods like Black Friday more smoothly, just rounding the corners, and have a year that's stable and well-planned.
If you're spending at least 2,000 euros per month on ads and feel like your costs are rising, your creatives are getting stale, or your ROAS (return on ad spend) is stagnating, Damien offers a free Creative Diversity Audit.
It's about 30 minutes together, after which he analyzes your creative diversity, audits your funnel, looks at what your competitors are doing for positioning, and gives you a concrete recommendation and action plan; no strings attached.
Reservez votre audit gratuit ici (page en francais)
Book your free audit here (English)
Throughout this interview, Damien mentioned multiple times how important it is to produce diverse creatives quickly and not get bogged down in editing. That's exactly what OneTake does; it handles the entire editing process automatically: noise removal, silence cutting, captions, transitions, and even a unique background track. You film with your phone, drop the footage into OneTake, and get a professional video back in minutes.
If you need to produce 10 to 25 creatives to fill your matrix, OneTake is the fastest way to go from raw footage to finished ad content without spending nights in a video editor.