Scale a winning video by testing 10 headlines, then new countries and devices

Yev Marusenko… side of the social platform, and then you're inside the app. There's like desktop performance, but then you're in the mobile app, so then there's performance within that. So, there's two ways I'm measuring the performance. There's kind of like the obvious one, and it's how it fits into profitability of the ad and profitability of the company, meaning that like what are your goals, what LTV do we need, what cost per purchase that do we need.

Yev MarusenkoAnd this is where targeting comes in. The TikTok is like working better for like US targeting, and then Facebook is working better for global targeting. And then it's a topic like iOS versus Android. So, these ones on US, and we have different goals on trying to reach $50 cost per purchase versus $100 cost per purchase versus $200 cost per purchase. I kind of said three different numbers because we have different goals on, one, part of the ad, we want to make sure we get enough data into like the onboarding flow. So, if we are able to reach $200 cost per purchase, which is way more lenient than if we're like being more aggressive on we need to reach profitability where some of these ads are like $100 cost per purchase.

Yev MarusenkoAnd depending if we're looking to the LTV, so we have to like know your churn numbers really well on one part of the ads is kind of like a break-even approach, and part of it is like pure profit. So, you have to know that one of them you're building up, that way it's it's profitable and scalable, and the other one is we're optimizing the onboarding flow. So, I'm kind of like sending in traffic and knowing that we're getting 1% conversion rate …

HostSo, to you, scaling means you take this winning ad and then you make variations of it sounds like a few different levels. Like audiences, the copy of the headline, and then landing page variations? Is that accurate?

Yev MarusenkoOn the audience, uh making sure that it's clear on device and countries, less on the student or interest, because the algorithm is really good. They find that, so it's more like broader in terms of the audience of the reach. It's more about how it fits into the funnel inside the app. So, that's on the audience side. Like kind of different countries, cuz different countries have different LTV goals, different MRR goals, and kind of things that the ad is trying to uh accomplish. So, it's making sure that it's different countries because then this fits into are they going into a free trial or not. So, this is on the ad side.

HostSo, what are we looking at here? Is this the original video or is this like a horizontal scaling?

Yev Marusenko… so then it's scaling on that first part on like different targeting. And then what happens is inside the app on different flows. So, I'm kind of like scaling, that way, I have different cost per purchase metrics, scaling it on the ad platform side, knowing that it's different targeting which diversifies the risk, meaning that to scale it even more, there's differences that happen inside of the app depending how we have the different paywall.

Yev MarusenkoOnce there's multiple ads going, this is like different countries, different targeting, it's like the same creative, but then like changing like the intro or the outro. This is where like on the creative side, there's like some variations just to like not fatigue the audience, but making sure that there's enough different ads. And then there's hundreds of different ads at the ad group level.

HostHow do you horizontally scale that across different audiences and demographics?

HostIs there any outlier stat that's really interesting here?

Yev MarusenkoActually, this is a better example here where I take this approach. So, this was headline testing where it's the same creative, so it's a constant variable. So, it's the same video ad, same targeting, but I'm just testing just one factor, in this case it's creative, and seeing um just the differences in the performance. Then I turned off a bunch, so then these ones have a lot more ad spend than the ones that didn't work. But it's this approach where it's like look at these 10 ads. My earlier example where I'm like even if you're spending just like a few dollars per creative, you're going to start seeing these outliers where it goes from 1 click to 10 click or a purchase comes in and some like never get a purchase and some start getting two, three purchases.

Yev MarusenkoAnd then I revisited weekly to make sure that that pattern holds. So, it's doing this approach, but for each of your hypotheses, making sure that it's here it's just the headline, everything else is the same. And the same thing like with the TikTok examples where it's all kinds of different targeting, different countries, but it's the same thing. It's that identical creative, and then they start taking different flows, and then it's repeating …