About Marcus: LinkedIn

ABOUT ROCKETSHIP HQ: Website | LinkedIn  | Twitter | YouTube


KEY HIGHLIGHTS

🗒 Setting up automated placement-level reporting to track ad spending on Meta platforms effectively is crucial for optimizing campaigns.

👍 Different creative types directly influence ad placements, with static creatives performing well on Facebook feeds and UGC excelling on Instagram Reels.

🛑 Avoiding restrictive ad campaigns for placements helps prevent higher CPMs due to Meta’s algorithmic penalties for limiting targeting.

🧢 Meta’s algorithm tends to skew toward younger audiences in subscription apps, even though older audiences often deliver higher conversion rates.

💵 Scaling ad accounts requires diversifying creatives and placements rather than relying solely on top-performing ads.

💲 Web2App funnels offer flexibility, enabling quicker iterations and experimentation with pricing structures compared to in-app funnels.

🌐 Web-based advertising is gaining momentum as brands bypass App Store fees and optimize web traffic for efficient user acquisition.

FULL TRANSCRIPT BELOW

SHAMANTH RAO:

 I’m excited to welcome Marcus Burke to the Mobile User Acquisition Show. Marcus, welcome to the show.

MARCUS BURKE: 

Glad to be here, Shamanth.

SHAMANTH RAO: 

We’ve talked about having you for a long time. I’m glad the stars aligned, and I’m excited to share a lot of very unique perspectives that you bring, which I’ve admired and appreciated for a very long time now.

I also like and appreciate that, on the surface, you’d think, “Oh man, that’s just a black box. You just kind of set it and forget it.” But, from what you talk about, it’s not quite as simplistic. Yes, some elements are a black box, but there are absolutely things you can and should control, especially given the unique nuances of subscription apps.

So, to get started, a big part of the black box thesis is not to worry about placements. To a certain extent, yes, you should have automatic placements, but you have a unique take and perspective. Tell us about how you recommend thinking about placements.

MARCUS BURKE: 

In the end, just like you said, Meta is moving in the direction Google has been running for a few years now, ever since they launched UAC. You have very little control over how you target or where you target.

There’s also minimal reporting on their end, and gaining control is difficult because their network is vast. It’s hard to know, “What do my specific placements look like?” or “In which context am I shown?” But on Meta, we’re still in a good position where the reporting is very strong.

Despite Meta recommending that you let the machine control everything, we can still know whether our money is being spent on Instagram, on Reels, in the feed, or the video feed. There are, I think, 25 placements by now. I don’t know; I don’t track it anymore.

But the idea is that I’ve seen, even pre-ATT, a vastly different performance depending on where my traffic was coming from. When I worked at Blinkist, we had distinct strategies for the Facebook feed versus the Instagram feed. We had different content formats for that.

We ran two different ad accounts targeting both of these placements. And a lot of that still holds. It’s just that visibility has suffered with ATT. Since traffic comes in as an aggregate, no one understands, and it’s harder to know how these placements perform on the backend.

As a first step, I urge everyone to understand where your money goes on Meta, so set up placement-level reporting. You can always use breakdowns for that. They tell you on every creative where your money went. You can automate that by going through the Meta Ads reporting.

You can pull CSVs and update them, put them into a spreadsheet, or hook it up through something like Supermetrics or Funnel.io to have an automated placement-level report. This way, you’ll see what percentage of your money is spent on Instagram, Facebook feed, or other significant placements and how that shifts over time. Whenever a shift happens there, you can expect something will happen in the app as well, regarding trial start rates or conversion.

SHAMANTH RAO:

A hundred percent. And as you said, even though it is a black box and you can’t always directly control it, you want to know where that money is going because, as you said, it leads to a performance shift. If I might ask, once you know where the money is going, how easy is it to influence that shift?

MARCUS BURKE: 

It’s not easy. What you don’t want to do is set up different campaigns for different placements. As soon as you restrict your placements, CPMs shoot up because you’re limiting the algorithm, and that gives you a penalty because you’re more targeted than other advertisers.

You want to try to influence that through creativity, as you always do. Creative equals targeting, which also equals targeting in terms of which placements you’re on, which doesn’t make it easy. As a first step, you need to figure out, on a creative level, which placements your creative is spending on.

Then, as soon as you go into evaluating your creative tests on a placement level, you’ll quickly see common patterns. For example, static creatives spend more money on feeds and Facebook, while UGC creatives are always big on Reels, mainly Instagram Reels.

By creative type alone, you already have some control over targeting. If you want to push more into Instagram Reels because they work well for you, you should do more UGC. Most advertisers need to consider what it does to their targeting. Everyone pushes UGC to Instagram Reels, which is an interesting placement because there’s a lot of volume there. It’s close to a TikTok experience; people are watching many videos. However, user intent and quality are lower than on other placements because the traffic is younger, and purchase power is worse than other placements. 

MARCUS BURKE: 

It doesn’t mean you can’t run ads at a good return there; it just means the traffic behaves differently. A lower cost per trial or cost per action—whatever you’re targeting—on that placement will mean something different than for people coming from a Facebook feed. So, drill down on a creative level to determine which creatives outperform on specific placements.

Sometimes, by accident, I come across things like, “Okay, interesting, this one scaled on Android on a rewarded video,” a placement that’s never in my top three or four. It’s very seldom that something happens there, but now and then, I find an outlier.

When you find those outliers, you can unlock this additional bucket. You can run those specific creatives and iterations of them in their own ad set and scale that specific placement if it works. Over time, you can unlock one placement after another if you find which creative works best. That essentially turns Meta into multiple channels, not just one, because the Facebook feed has massive traffic, as does the Instagram feed. So, you want to know how to spend your time there.

SHAMANTH RAO: 

A hundred percent. I like how you talk about breaking it out into its own campaign or ad set once you know what’s performing because then you can say, “Look, this ad set is UGC-Reels focused, this one is an animated video, and this one is focused on Feeds.” What I’m also hearing is that it’s a great way to expand your audience because someone on Feeds is probably going to be older, maybe a Midwestern mom, as you like to say, compared to someone younger on the East Coast or West Coast who’s a Reels user, right? That’s very interesting. I haven’t heard anyone describe it quite this way, so thank you for sharing that.

MARCUS BURKE: 

As you said, it comes down to setting up your account structure to allow you to scale on these placements. The issue with Meta is that it often acts on a “winner takes all” principle. If you throw everything into one ad set or Advantage Plus campaign, they’ll put 80% of your budget on your best creative, and the rest gets very little—the last 20%. That’s not your goal. If you want to scale an ad account, you want to scale multiple creatives and placements, so you reduce risk by running across a large portfolio.

It’s not about running your best ad in the best placement to the best audience because that limits your spend. Maybe you can spend $50K a month doing that, but if you want to spend $500K or more, you need broader targeting, a broader creative set, and different angles in your creatives. Meta won’t diversify for you; they’ll always favor the winner. So, it’s up to you to build your infrastructure in a way that enables you to spend across those different placements and never put all your eggs in one basket.

SHAMANTH RAO: 

I would also add that, in subscription apps, Meta doesn’t always do a great job of identifying the best ad. I know you have a strong opinion on this. Can you share your thoughts?

MARCUS BURKE: 

Yeah, that’s another issue. In products where Meta is tracking actual purchases—like in the D2C world or eCommerce—Meta has a very good understanding of who’s buying. So, Meta can identify your ideal customer well. But in the subscription world, most apps optimize for a trial start. From trial to conversion, there might be a 50% to 60% drop-off, depending on how good you are, and this drop-off varies massively by audience.

For instance, in the 18–24 age group, you might see drop-off rates of 80% to 90%, while among older users—like 65+—the trial conversion might be 80% because they actually have the money to spend on a subscription. This means the algorithm skews younger because all trials look the same. Meta sees a trial event and thinks, “Young audiences are cheap,” so they give you more young trials without knowing that those trials convert worse on the backend. 

So, more modeling must happen if you’re running this kind of optimization. You have to guide the algorithm in the right direction. One thing I like to do is exclude younger audiences. I tend to avoid targeting anyone under 25, at least in the early days when you don’t understand how things behave. Sometimes, targeting 30+ is a good call, depending on the app and price point.

Then, you can also look into which creatives drive which traffic. For example, static creatives typically target Facebook feeds, and older people are on Facebook feeds. So, statics bring in older people, which might need a different cost-per-trial goal because you know that older users are more likely to convert—if you have an age survey during onboarding. That’s why gathering first-party data is super important. You want to ask users for information to help you understand conversion on Meta or any channel.

SHAMANTH RAO: 

Do you find connecting a creative angle to a user’s goal reliable and actionable? For example, if I join an app to improve my productivity but I saw an ad that says, “Get a promotion,” how reliable are those insights?

MARCUS BURKE: 

The goal and angle are much less reliable than age, gender, and channel. Those basics give you a very good understanding. Anything deeper becomes tricky. Not every user goes through this kind of funnel. Someone might see a productivity ad but give a different answer during onboarding. Also, an ad with a productivity angle might mention promotions because you’ve seen that work.

Usually, there’s little drop-off in onboarding questions, so you can gather data to get additional insights. However, don’t go crazy with your modeling. It’s more for direction. For example, you might find that females who say they want to be more productive convert worse than males who wish to a promotion. That gives you direction—maybe use more male creators in your videos because you want to target males while talking about success and promotions.

SHAMANTH RAO:

 Makes sense. Just to switch gears, something you also talk about—and something you’ve worked extensively with—has been Web2App. Beyond the obvious, like “save 20%” or using a different flow, what are some of the pros and cons of using Web2App?

MARCUS BURKE: 

I just wrote an article about that on the RevenueCat blog. Everyone starts using Web2App to save the 30% fee, but in the end, that’s not the real benefit. It’s a catchy hook to get people interested, but for me, the funnel behaves completely differently than an in-app funnel. Every metric looks different, and you use a different algorithm on Meta—the web sales one instead of the app promotion one—which behaves differently and drives different people.

It’s not like you can just compare your funnel metrics and say, “Okay, if everything stays roughly the same, we’ll save 30%, and all is good.” It’s a totally different picture. You’ll probably have to iterate on the funnel a few times, just like you would on your app onboarding.

The main benefit, to me, is that you can iterate much more easily because you’re on the web. It decouples the marketing from the product experience. This means that marketing, using a no-code tool, can build a number of funnels, iterate on them, and test them much more easily than in the app. It saves time and is a fun task for marketing to work on.

You’re also a lot more flexible with pricing. Through Stripe, you can set up anything you want—you can do upsells or offer pricing structures that are much more customizable compared to what’s possible with SKUs on the App Store.

Lastly, a web onboarding can help you stand out. Everyone sends traffic to the App Store these days, so users are very used to landing there after clicking an ad. You can differentiate by having different screenshots, but the App Store is very standardized. The web is not, so you can bring part of your product experience to the user early on, like an “aha moment.” Users can see the value of your product immediately, or you can use quizzes, which are now a standard way to do Web2App onboarding.

At Blinkist, we scaled paid content pages, where the first touchpoint was an article about the service we provided. This was very interesting because that funnel was tailored to certain placements. For example, on Facebook or Twitter feeds, users are consuming news, so a UGC video ad doesn’t fit the environment well. But a written article fits perfectly. That’s why we scaled that placement and ran it on Facebook feed, Twitter, Taboola, and Outbrain.

So, think of Web2App as a way to differentiate and tailor an experience to certain placements, something that’s much more unique than what you could do in the App Store.

Of course, there are disadvantages as well. On the web, you need to gain trust and credibility. The App Store is standardized in a good way—it shows reviews and ratings right away, and people can immediately tell if an app is trustworthy. On the web, you need to work harder to build that trust. You should think about adding social proof, trust symbols, and secure payment options. Offer more than just credit cards—include PayPal or Apple Pay so users can pay as conveniently as they would in the App Store.

So, you have to do some homework, and it won’t work perfectly in the first iteration, but the benefits can be massive once you get it right.

SHAMANTH RAO: 

Over the last year or year and a half, I’ve seen more developers adopting Web2App and doing quite well with it. It really feels like—

MARCUS BURKE: 

Like I always say, when SKAN came out, everyone said, “Okay, we’re going to spend on the web now and not use SKAN.” But then they realized it wasn’t that easy. It feels like people have figured it out in the last year or so, and more are moving spend to the web, which is cool to see.

SHAMANTH RAO: 

Just to latch on to something you said—it’s not easy. You said it requires homework and preparation. How easy is it for an early-stage app to adopt Web2App? Is there a certain scale where it makes sense for an app to start looking at Web2App?

I ask this because we worked with someone very early on. They had a budget in the four figures per month, and they wanted to do Web2App. My advice was, “It’s too risky with such a small budget; it’s probably all going to be gone.” So, at what scale does it start making sense to pursue Web2App?

MARCUS BURKE: 

I’ve seen it work well for even small advertisers. My RevenueCat article was about leveraging Web2App to go from zero to one on Meta because any funnel you want to do work will take resources and require multiple iterations. You either need to iterate on your app onboarding or your web funnel.

In the early days, your product resources should be focused on improving the core product. There’s probably a lot of work to do outside of onboarding—getting rid of bugs and adding user-approved features. Marketing doesn’t have to compete for those same resources. Instead, they can build a web funnel, improve it, learn from it, and send traffic that way. Plus, signal quality is better on the web, and you don’t need to deal with the complexities of SKAN or AEM.

Of course, it depends on the product. Older audiences tend to perform better on the web, while younger users are more used to the app funnel. And it depends on the resources you have in-house. If no one knows how to work with the web, then maybe it’s not the best place to start. But if you have the resources, it can be a good strategy early on.

I’ve seen startups scale to $100K or $150K in profitable spending on Meta without ever running an app store ad, and it’s worked very well for them.

SHAMANTH RAO: 

Are there examples that come to mind that you’d feel comfortable sharing?

MARCUS BURKE: 

Yes. I spoke with a language-learning app called Jumpspeak. They were only doing Web2App. It was a small team—the founder was running the ads. They knew what they were doing on the web, which worked well for them.

For example, they didn’t do a free trial; they did a 100-day, no-questions-asked money-back guarantee. Users made a direct purchase, which is an amazing signal for Meta to find the right audience. They even had an upsell after purchase. Their pricing was quite expensive, over $100 for a yearly plan, and they had a lifetime deal for $300. Then, they offered an additional video course for another $40. They built a neat package with a lot of pricing psychology, which is easier to do on the web. They charged a lot more than most apps, which I found interesting.

SHAMANTH RAO: 

That’s interesting. You’re right—I’ve noticed much higher pricing for products that use a web-based flow. And you’re saving 30%, so your net margins are higher.

MARCUS BURKE: 

It’s probably not infinitely scalable, and there’s a ceiling. So, you need both the Web2App funnel and the direct-to-app funnel at some point. The app store is great for volume, and volume is what you need to gain rankings through app store optimization. If users don’t pay right away, they’re gone on the web. They won’t go to the app store and search for your app. They’re just going to leave.

So, you won’t get the install volume or ratings you need to rank well in the app store. At some point, you need both approaches to reach scale.

SHAMANTH RAO: 

100%. This is quite a rabbit hole, and I’m so fascinated every time I learn from you—whether I’m reading your posts or speaking with you directly. Marcus, I want to respect our schedules, so this is a good place for us to wrap up. But before we go, can you tell folks how they can learn more about you and your work?

MARCUS BURKE: 

You can find me on LinkedIn— https://www.linkedin.com/in/marcusburke/ I post content about app growth and Meta ads three to four times a week, so check out what I have to say—it might be helpful.

SHAMANTH RAO: 

Amazing. We’ll link to that in the show notes and send it to our email list. For now, thank you for being on the show. It’s been a pleasure and a privilege to have you.

MARCUS BURKE: Thanks, Shamanth.