B (2:32)

For instance, it completely 180 how we approach creative testing, to be honest. So I guess if you haven't heard of Andromeda yet, you might be living under a rock. But I'll give you a quick recap. It's a new update to the algorithm that Meta rolled out over the past all of 2025. It was kind of like a gradual update and they have shifted how you approach targeting with your creative. So before it was a lot more dependent on the media buyer to make sure that they were targeting the right audience through things like interest stackers, targeting different genders and breakdowns in age and regions and things like that. But with Andromeda, it is your creative that does the targeting and leads that for you. So the biggest recommendation is being more open in your settings, which has its problems, and I want to talk about that for a little bit. But also being open, casting a wider net and trusting Meta's black box to do the hard heavy lifting for you because it can actually now read what is in your creative and hit the right people using AI, which is scary.

A (3:37)

Yeah. The metaphor that I've heard was that it's sort of like a concierge at the, you know, the hall of Alexander, the Library of Alexandria, and this AI concierge, instead of like having the decision Plinko down into different sort of segments and logic, it's just, you know, I just uses AI magic to find the right ad for this right person based on everything they know. How explicit do you need to be in how you break out your funnel, for instance, with your. With your campaigns these days? Or is it something like. Or Andromeda really is also able to know where people are at in their buying journey, I assume as well.

B (4:10)

That's actually a big problem that I see in the audit stage, where people are. They're making the mistakes of setting up their campaigns too wide, but then not setting up their audience segments properly. And if you don't have that set up properly in the background, you're not really telling Meta who is a net new audience, who's engaged and who's existing already. And some of the frequencies I'm seeing coming through where people are just like hammering the same ad over and over again to an already engaged audience. And not using meta for prospecting anymore is really alarming. So just a quick tip. Definitely ask your media buyer, if you're listening to this podcast, what your frequency is looking like for new customers, engaged and existing over the last 30 days. Have them break out those audience segments and if you're seeing anything above a one point something in new, you're doing something wrong. So media buying still matters. And it's not that media buying is getting easier, it's just changing when it comes to creative and how you set up your funnel. We are still setting up purchaser ad sets, middle of funnel ad sets, and we're making sure that we have the right creative within those. But we also go quite broad with Advantage plus and dump a mix in there and see what happens, knowing that Andromeda should be able to do what it needs to do. But I also want to talk about your first question, which was how we approach creative testing. Because then I just told you what Andromeda was and didn't answer your question.

A (5:30)

Can I? Yeah, so that's okay. There's so much to talk about. I want to preface this a little bit because on this podcast, I think it's six years we've been doing this podcast. Now when we started, we talked a lot about something we called the pilot test, which was a way of doing some of the psychological profiling, digging deep down into your, you know, your avatar's interests and their pain points and things like that. And basically using a prescribed system of running ads and then using CTR to essentially learn each how. How each part of the ad was working, and then forging together excellent ads based on all the best practices forged from these sort of CTR based tests in order to forge winners. And I feel like that was forever ago. You were here when we were using the pilot test, right?

B (6:14)

I was, yeah. I. That's what I was trained up on when I was a media buyer. So we would take one image and everything would be same down to like the font size, the image, the cta. The only thing that would change was the headline. And we would make five versions of that ad. That headline alone was the isolated variable that tested different angles. So what we were trying to prove out in that pilot test was what angle does our audience care about the most? And then we would take that angle and we'd build off of it and that would be pilot testing. In a nutshell, we can't do that anymore with Andromeda. Because now AI is looking at all of these statics and it's going boring. You're giving your audience the exact same thing. We're going to flag you for creative fatigue and now your CPAs are going to rise as a consequence. So it's insane. You can't pilot test the way you used to anymore. Instead, now you need a lot more Net New Creative because its AI can pick up on if you're making small tweaks like that. But that Net New Creative needs to align with strategic thinking. Like diversity isn't enough to win to it. You also need to know how to speak to the audience it's hitting its ads with.

A (7:23)

And you have to have holistic ideas. Whereas before, as you were describing the sort of holdout testing of finding these, all these very elements, Andromeda will not punish you for having that much redundancy in your ad account. And so you need to sort of put up like whole shots kind of thing of just like your best concept of what an audience might like about the product, kind of really thought out from top to bottom. Is that kind of right?

B (7:46)

Yeah, I think like you'll see it across LinkedIn, you'll hear it at Pilothouse a lot. It's like this new world of Persona testing. But the problem is people aren't approaching it right because it's really hard to do without a clear strategy. So it's one thing to have like what's a good brand to use. For an example, let's use a jewelry brand. So just generic. You're selling jewelry and you've identified that your target audience is a working professional where they're in the office every day and you've like drilled down the pain points and things like that. It's not enough just to be like, oh, we need B roll of someone at their desk working with our necklace and our earrings on anymore. You need to understand like what their pain point is, what their core attributes are. And then you need to link that to the brand now as the solution to those pain points and core attributes. So for example, if a working professional is maybe wealthier because she's an in office, has a high paying job, you're not so concerned about messaging like bundling and saving. But maybe she has office attire that can't be styled in multiple ways, isn't like unique in nature. So like your jewelry is the answer to that of having like uniqueness in the office despite the dress code and talking about like how to level up your styling when you're in office and you're wearing your boring millennial gray suit. So it's just. It's crazy logic now where it's not enough to just be like, bundle and save or sweatproof jewelry. You need to actually know who your Personas are and then how to trigger them to win the click.

A (9:18)

One of the things that came up in the podcast that I liked was, was that was someone thinking about what would someone spend their money on instead if they weren't spending their money on, like, depending on the size of AOV you have. But it's like this whole sort of like, spectrum of questions. You can kind of ask around who your Persona is and then you can filter out different angles or strategies kind of from there. Do we still test different strategies? Is there still an aspect of like, I guess you don't. I guess you just put them in there and ones that win get picked.

B (9:48)

I would say going back to your point on, like, how do you win in comparison to the competition if you're running, like, only conversion focused campaigns and you're just screaming the same USPs that all of your competitors have? So let's go back to jewelry. And if you look at every jewelry brand, they're all saying waterproof, despite maybe it not being waterproof because there's a lot of, like, fast fashion jewelry out there. Right. But they're all screaming waterproof. That seems to be the big thing. Atelier is like a really big house brand that's running that. And I think a lot of competitors are following suit. When you only run conversion campaigns, your ad is going to be compared to all of those competitors. So you are definitely not going to win if you're screaming the same thing that all of your competitors are running. And that's why core differentiators are important, and that's why Persona building is important. So you need to worry about, like, building up your brand and resonating with the audience that you're trying to speak to. But you also need to worry about, like, what do we have as a brand in our messaging that if you lined us up with five competitors, they don't have and how can we put that front and center to make sure that we win? So it's not just Persona work. There's definitely, like, problem solution creative, there's differentiator creative, and, like, ways to approach it while doing your testing. But it's. I would start with understanding your audience before doing any of that work.

A (11:10)

And yeah, understanding your Persona can inform those other formats as well. Right. The very specific angle. And I think that's What Meta is looking for is. Is creative really to speak to people.

B (11:20)

One thing, too, is AI slop. It's like a term we like to use at Pilot House. I think because I'm in the agency world, my ads now like, that I get hit with because I'm like, in this bubble are make 500 ad creatives in 30 seconds. Like, fire your agency. Use this AI tool instead to win on Meta. You're not going to win if you use that. That is like, pure AI slop that doesn't have a strategy behind it. And what is going to happen is you're going to think, like, okay, one in 500 might be a winner, but you're now spending money without logic on 500 ads and you get one winner. What if instead you spend money on a strategy that can give you 30 really good ads? You have 15 winners come out of it, and with all of the winners and losers, you're able to draw insights on. I know exactly why these ones won and exactly why the other ones lost. So that's the big difference between brands that are like, maybe we just need the AI tool and brands that are doing, like, heavy building of who their segments are, who their Personas are, and how to effectively speak to them.

A (12:29)

I wonder if you could be a little bit more specific. What does actual creative testing look like for Pilothouse?

B (12:36)

So I think that's been the most frustrating thing for media buyers and for us to wrap our head around, because you build out a batch of creatives. If you were to look at it in levels, you have your Persona. And I've said Persona maybe 50 times on this podcast. But say you have, like, your working professional female Persona within that, you build out concepts for that Persona. So it's one concept that should have a hypothesis of, like, what you're trying to prove out. So it would be okay. We want to understand if this Persona cares about this certain thing within our product. We built a concept that aligns with that. So maybe for a working professional in the jewelry category, it would be one necklace, five suits. And that's just the point of, for office attire, you just need to buy one chain and then you get down to the ad level. And the new way to test out that concept is to take the messaging that you've developed at that concept level, but put it within maybe five uniquely different ad creatives. So one would be like a carousel that shows. Swipe it. Five different outfits. One would be a branded static of, like, a really strong necklace shot in the studio with a bold headline. Nice and clean and easy. The third could be an iPhone photo of, like, a girl taking a selfie in her cubicle. And the fourth could be a get ready with me creator video. And the fifth could be a, like, TikTok trend that touches on those points with, like, a text ball of a girl in her office. So you test it out. It's like two layers of the messaging is the first thing you're testing. Secondly, you're testing how to resonate with them, like, through the creative visually. And then you dump them all into the same campaign. But you're able to have findings, which is kind of cool.

A (14:18)

I love it. It also gives the opportunity to create moments in someone's day. Like, if you were just launching a bunch of static ads or a bunch of video ads or whatever, but as opposed to, like, you're launching across the ecosystem in all these different formats. They say, like to build a brand, you know, brand awareness. It's like two impressions in the morning, two impressions after lunch, two impressions in the evening kind of thing. That's the most, you know, basic way to say it. But by just going diverse with your ad formats, you have that bigger chance of reaching people at other moments in a relevant time.

B (14:47)

Yeah, you also have a chance of being memorable. Like, if you're really trying to drill home the fact that, like, we are for the working professional, odds are they're not remembering. No one remembers your first static ad that you serve them, no matter how good it is. But if you hit them with the same message in like, a fun, engaging way in four different visuals, guaranteed, the likelihood of them remembering your brand is going to, like, tenfold. So I think that is the. That's the North Star there.

A (15:14)

So just sort of practically for people listening who think they might have Personas in their ads, but maybe aren't sure what goes into the creating of a Persona, how much of it is what the brand tells us versus what the data tells us. How do you actually do Persona creation for brands?

B (15:28)

It's a lot of research. We'd like to start with something that we call like the four Cs, which is taking a step back outside of just what the brand is saying across their website, who they are, but actually looking at culturally what's happening in their vertical. So if it's jewelry looking at, okay, maybe you're a fast fashion brand and you want to understand how people perceive fast fashion brands. Do they actually care about sustainability or do they not? Are they looking for dupe culture or are they looking for that quick trend that's going to die in a month. So just like, what cultural shifts are shaping consumer behavior? The next one would be the consumer itself. So where do they fit, like, in the economy? Are they spending money on jewelry right now? How they shop, where they shop for jewelry? Are they online? Are they secondhand? The category would be learning about broad jewelry market as a whole. So just where are we trended to go? Is this a vertical that's seeing consistent growth? And should we be growing within it? And then lastly, the. What would be the other one? Company. So what is our company saying in comparison to these three? Are we, like, completely missing the mark based off of, like, where we sit within the jewelry landscape, or are we doing the right things and building a brand around these changes? And so we look there, and then we start to do, like, deeper research, understand where we fit within the market and build Personas forward after that.

A (16:56)

I'd like to propose a fifth c. Let me know what you think. Competition. If. Especially if you're in a. If you have a brand that is, you know, if there are brands that are like you a little bit, go see, you don't have to steal the angles. Their angles might not work for you, but if you go to the ad library for your competition, you'll be able to see the angles that are kind of working for them if they're doing this right.

B (17:15)

Oh, absolutely. Or I remember working on a. There's a fertility kit brand, but worked with them for about a year. And one of the biggest things, It's a crazy space for, like, female ads for. Especially in your 30s. I don't know. You definitely wouldn't see this, Eric. But, like, are you pregnant yet? And, like, can my algorithm chill out for a second?

A (17:39)

Yeah, yeah, just chill.

B (17:41)

It's, like, screaming at you. But with us, the biggest unlock was like, okay, all of these guys are going stats forward. Like, 83% of women do this. Or, like, here's what you need to know about this hormone. We were like, how can we stand out in this, like, mass competition where this is extremely oversaturated. We went on symptoms where we talked about, like, hormone imbalances. And instead of saying something like hormone imbalance, which is boring, we started talking about fatigue, where we started talking about, like, growing tired of, like, waiting while you're trying to conceive and pain points and symptoms and put those front and center in the headline. And then we started to win. So I think competition in their ad libraries are really important because they're not always telling of what's working, but I think they give you a really good opportunity of how to make yourself different and better 100%.

A (18:32)

I want to pick back up on something that you mentioned. Just sort of like the way that people listening to this podcast, if they're maybe even not doing the ad buying themselves, they can go talk to their ad buyer and ask them some questions to help help understand whether they're really, you know, have groked this new Andromeda landscape.

B (18:48)

Well, I think the biggest one for the media buyer is in audits. I'm seeing a lot of accounts where they're just starting to go broad, but they're not setting up their audience segments properly and capping like how much they're spending into engaged and existing audiences. And it's resulting in like two audits back to back for well known brands that I didn't audit for one had over 100 touch points over 30 days in their engaged audiences. Which there's no world where someone needs to see an ad 100 times over 30 days and another one was 40. And the fact that I saw it twice makes me think that a lot of media buyers are just going, okay, everyone on LinkedIn is telling me go broad and I'm going broad and my life is easy now. You still need to be do like fine tuning and checking those things to make sure that meta's still filling your funnel and atop a funnel play for you. And then the question that I would ask your creative strategist would be show me your Persona work. Show me like the last batch of ads that you made and how we expect them to speak effectively to our target audience and our target Personas that we know are strong customers and have a strong ltv. And if you look at the last batch and there's not a clear through line, then your creative strategy is missing.

A (20:01)

Well, let's end it on that note. I think that was super valuable. If people have questions about creative strategy. What's your email? Would you mind just sharing your email address? If anyone out there wants to email Abby about creative creative strategy?

B (20:12)

Yeah, it's abbyilothouse. Co. I love talking about this, so if anyone wants to chat about it, I'm all ears.

A (20:20)

Amazing. Look forward to the next time we chat. And I look forward to our next DTC Dines adventure.

B (20:25)

Yeah, we'll see you there.

A (20:33)

Thanks for listening to today's episode. If you're not getting the D2C newsletter, you can subscribe for free at directtoconsumer.co. and if you want to learn more about Pilothouse's all killer. No filler services. Take off to Pilothouse Co. I'm Eric Dick and this has been the DTC podcast. We'll see you next time.