82% of top-grossing apps now take payment outside the app store. Patrick Stuart-Constant’s team at Sociaaal ships 4,000 video ads a month, scaling to 10,000. The shift to web-to-app monetization isn’t coming. It already happened.

Part 1 of a two-part panel from a FunnelFox webinar, with Rocketship HQ as partner. Shamanth Rao moderated the conversation with Jacob Rushfinn (CEO, Botsi), Andrey Shakhtin (CEO, FunnelFox), Patrick Stuart-Constant (CEO, Sociaaal), Mike Gadd (VP Customer Success, Singular), and Elise Zareie (Head of Paid Social, Lingokids) on how paid acquisition teams design the path after the click.

On the panel:

Shamanth Rao (Rocketship HQ):   / shamanthrao  Rocketship HQ (partner): https://rocketshiphq.com
Original webinar recording:    • From traffic to revenue: how paid acquisit…  
FunnelFox (webinar host): https://funnelfox.com
Jacob Rushfinn (Botsi):   / jacob-rushfinn  
Andrey Shakhtin (FunnelFox):   / andrey-shakhtin  
Patrick Stuart-Constant (Sociaaal):   / patrick-stuart-constant 
Mike Gadd (Singular):   / mike-gadd-27a16427  
Elise Zareie (Lingokids):   / elise-zareie  

FULL TRANSCRIPT BELOW

Shamanth: Hello and welcome to our webinar, From traffic to revenue: How paid acquisition teams design the path after the click. If you’re joining us, please feel free to drop any questions in the chat for our guests.

I’m excited to be joined by our fabulous guests. To go quickly around our virtual room: we have Jacob Rushfinn from Botsi, Andrey Shakhtin from FunnelFox, Patrick Stuart-Constant from Sociaaal, Mike Gadd from Singular, and Elise Zareie from Lingokids. We have an all-star panel today that I’m excited to host.

This webinar is hosted by us, RocketShip HQ, and FunnelFox. Andrey, if you want to take a moment to tell us about FunnelFox — and thank you for making this possible — please go for it.

Andrey: Sure. Just a brief overview: FunnelFox is a platform for building and launching funnels. It includes a funnel builder, an agentic funnel builder, and payments infrastructure that helps app developers scale on web without any hassle with web payments.

Shamanth: Excellent. And your co-hosts are us. I’m Shamanth, CEO of RocketShip HQ. We’re a creative-driven performance agency making hundreds of ads per month for a number of our customers, so hit us up if you’re looking to scale your creative production.

But for now, let’s put the spotlight on our guests.

Shamanth: I’d love to hear from you all. What’s one shift you’ve seen in the last 12 months or so that you feel most teams haven’t priced in yet? Mike, would you like to get us going?

Mike: Sure. I think the big thing we’re seeing is that most teams are still optimizing for install cost or trial volume, or looking at how many free trials they’ve been able to turn into paid subscriptions. But the real bottleneck is keeping users through that first paid cycle. It’s the first renewal that’s the interesting bit.

It’s getting harder, because we see users actively reducing total subscription spend across all categories. They’re not adding more — they’re consolidating and canceling. So that first payment instance is really important, and it’s what we’re seeing people track more.

Shamanth: For sure. Patrick, what are you noticing?

Patrick: I’m noticing that the falling cost of creative production continues to produce an incredible opportunity.

We started going fully AI on creatives two years ago, when they were still pretty bad. They started getting good 18 months ago, and that’s when we started taking off. Since last year it’s been great, and it just gets better each month. Honestly, it’s a huge driver for any app studio.

Shamanth: Yeah, and I still keep hearing that AI creatives are terrible. I think bad AI creatives are terrible. The good ones aren’t priced in yet. Andrey, what are you noticing?

Andrey: For sure. It’s the massive shift to web-based monetization. That includes both web-to-app, where you run traffic to a web funnel and sell the subscription on web, and app-to-web, where you integrate web payments inside your app for the US market.

Both trends look very strong, especially web-to-app. We recently published a report on these trends, and we see adoption increasing more and more year over year.

Shamanth: Yeah, and I’ve seen some apps that have basically stopped using the app flow — they’re using only the web funnel now. Jacob, how about yourself?

Jacob: Most of my focus is post-acquisition, and I really see personalization making a huge difference. It used to be all fake personalization — pretend personalization. Now you can actually create real personalized experiences in the funnel, post-paywall, using gen AI to create genuinely unique experiences, and of course to serve the right offer to the right user.

Showing that right offer can make a pretty meaningful difference in monetization. I’ve seen the largest apps figure this out — we see Tinder creating dynamic paywalls, different paywalls for different users. So personalization from the first touchpoint all the way through to deeper experiences makes a big difference.

Shamanth: Yeah, and we’ve always had the data. It sounds like it’s become a lot easier to use that data to personalize journeys, especially with the help of AI. Elise, how about yourself?

Elise: In the last year, web funnels and web-to-app have been a big focus, and it went from being a side experiment to a big acquisition engine.

It only works if you treat it like one, and not just as a side project you’re doing an hour a week. The barrier to entry has never been lower — it’s easier and easier to do web-to-app — but you still need dedicated ownership. You still need to run strong experiments, and not just try it once and then discard it.

That’s been something I’ve observed across different teams on subscription apps.

Shamanth: Yeah. It seems easy, and it’s actually hard. And for those of you who are curious, Elise and Patrick, you’re both on my podcast, Intelligent Artifice — so please check out everything they’ve just said, but in a lot more detail.

Shamanth: Andrey, you had a report recently saying that about 82% of top-grossing apps now route their payments outside the store, which is massive growth over last year. What’s driving that, and what sort of apps do you see that might be moving too late?

Andrey: We researched the market in terms of web funnels and web-based monetization, and the first thing driving growth is attribution. There’s no default restriction like the one Apple introduced five years ago. You can use pixel-based and conversion API-based attribution, which means you get signals about the purchase immediately and can optimize your campaigns quickly and with less budget. You can literally understand which creative converts better or worse.

Conversion rates are also higher, just because of the flexibility of the flow. The whole funnel from ad creative to purchase is more holistic, because there’s no generic app install step in the middle. Users can continue engaging with the quiz right after clicking the ad creative, which increases conversion rates a great deal.

And there’s less commission — you pay 2–5% payment commission to payment providers instead of 30% at scale to the app store.

So those are the main factors driving this.

Shamanth: For sure. Lots of different factors — lower costs, more opportunity for customization. That all makes sense.

Shamanth: Mike, from a measurement perspective, what has changed as more teams have gone off-store?

Mike: I think with more teams going off-store, you’ve got a big challenge around different platforms. Previously you were maybe looking at just app information and everything was on one platform, but now you’re looking at web and app as well.

What we’ve worked on with customers is making sure you have everything going through as if it looks like an app purchase, for example. That means you’re able to better optimize your campaigns, and you’re able to look at all your revenue in one place. That really enables people to optimize more effectively, and it helps give you accurate ROI and LTV calculations.

One of the biggest challenges people hit is that they see their spend on ads across different platforms, and they can’t reconcile that at all to where the payments are coming in and what the revenue looks like. So having a strong way of decoding that is really important.

And then finally, a lot of customers don’t have a strong user ID implementation, at least at the start. If you have a strong user ID, you can knit all these things together and measure everything in a total view. If you don’t have that, you’re seeing a device on web, a device on an app, and you can’t match them together. There’s lots of rich measurement available, but there are lots of pitfalls where you lose that measurement and lose that insight.

Shamanth: So if I understand correctly, there’s one paradigm on the web and one on mobile, and you really need to put in the effort to marry them — because otherwise you’re going to have two very different systems of reporting that will be a mess if you’re not very careful.

Mike: Exactly. We call it platform hell. You get stuck in platform hell, you can’t make sense of what’s going on, and it’s key to try and find a way to get out of that.

Shamanth: To switch gears a bit — Elise, to lob the ball over to you. For a lot of products right now there’s an infinite number of creative directions they can go in. What’s your typical approach for figuring out which bet you want to take upfront?

Elise: I really like this question, because there’s a classic answer, which I’ll share, but this year I’ve been seeing a lot of new approaches for creative that aren’t the classic route.

To me, the classic route is: don’t try to reinvent the wheel. Look at what competitors do creatively in your vertical. Look at top-grossing apps that are doing really well, even if they’re not exactly your vertical, and try to lift and shift their winning formats to your app to get started.

What that means is taking top creatives in the Meta ad library, for example, and seeing how you can adapt them to your specific messaging. Your audience will dictate the kind of creatives you’re picking and the placement. If you have an older audience, it will skew towards statics. If you have a younger audience, it will skew towards Reels — quick videos without much editing, very organic-looking. You can’t really convert someone who’s 50-plus with Gen Z-type creatives, and the opposite is also true.

Something a bit harsh I’d say: chances are, when you first get started, no one will care about all the benefits of your product, even if you’re very proud of it and want to share them with the world. It works better, at least in the beginning, when you focus on a specific persona and a specific problem that your app solves, rather than talking about all the benefits. Once you get some momentum, you usually split your creatives between brand-new concepts and iterations.

So that’s the classic approach. What I’ve been seeing this last year is people sourcing more organic creatives from TikTok directly and from creators directly. They let creators dictate ideas and come up with creative directions themselves, and then they take those directions and use them in paid.

It’s a bit newer. It doesn’t work for every vertical, but I find it quite interesting, because it adds a lot more variety to ideas. It’s fresher when the audience themselves imagines how to sell your app than when your creative team does.

Shamanth: What I find interesting in what you said is the two very different perspectives. On one hand you said, “Look at TikTok creators and see what you can borrow.” On the other, “Look at what’s working for your competitors and see if you can repurpose that.” One way you’re mitigating risk by taking something that’s proven — and the other way, with TikTok, you’re also taking something that’s proven. You’re not reinventing the wheel. That’s what I’m hearing from you.

Elise: Yes. And I think as you scale, I’d really encourage you to be more adventurous with your creatives and push the envelope more. But when you first start, you don’t want to be making hundreds of creatives that will flop. You want to get something that will work from your first rounds of creative testing, which is why I wouldn’t necessarily reinvent the wheel at the start.

Shamanth: Patrick, you spoke briefly about how you’ve scaled creative quite dramatically. Talk to us about how you decide which bets to take when you can basically make infinite numbers of creatives.

Patrick: The way I see it, creatives are a search space. The old-fashioned way of making a few creatives a month — or a few thousand creatives a month — only searches tiny bits of that search space.

The idea is that you want to cover as much of the search space as possible, and there are two implications of this. One is that you really want creative diversity. You don’t want endless variations of the same creative, because otherwise you’re always in the same little spot: 100 attempts, but all in the same part of the search space.

As we scale up, you want to be trying lots of different areas — and when one looks promising, that’s when you do lots of variations to really explore that part of the search space fully. That’s how we navigate scaling up our creative production. At the moment we’re at 4,000 video ads a month, scaling up to 10,000 within the next couple of months, hopefully.

Shamanth: That’s a huge number. But what I’m hearing you say is that you’re not making 4,000 that look mostly similar — you want those to attack very different parts of the search space.

As a follow-up: if something’s working, how much are you iterating on what’s working, versus going to a completely different area of the search space?

Patrick: We’re constantly doing both in parallel. We continue doing variations of one ad, trying to change it as much as possible, until it stops working. And in the meantime we’ve been exploring other parts of the search space, stockpiling brand-new marketing angles and brand-new concepts. Then we start doing more variations of the next one once one has run out.

Shamanth: You said you’re at 4,000 and on your way to 10,000. What does that mean in terms of how your team has changed or grown?

Patrick: I think that’s a classic Jevons paradox: the cheaper something is, the more you consume it. Our creative team is one of the teams where we use AI the most, and it’s also our fastest-growing team.

We’ve doubled the size of the company over the last few months, and we’ve tripled the size of the creative team. That’s because there’s a lot you can automate with AI — but fundamentally, ads that are truly creative and original, with fresh ideas, those are the new concepts in the search space that will perform really nicely. So we’ve needed much more human creativity as we’ve scaled up the number of ads.

Shamanth: And you and I have talked about this — it’s been our experience too. Our creative team has grown 2x, just because as we have more throughput, we need more humans to manage the creative process.

Elise, what are you seeing in terms of how AI is impacting the staffing of teams?

Elise: I think it’s an amazing help for repetitive tasks — resizing, localizing, generating B-roll, drafting copy. All the things that aren’t the concept itself, but are the small iterations and adaptations. That’s one side of it, where it’s been extremely helpful.

The other side, which I love, is creative analysis. When you have huge creative volume, you want to understand what’s working, what’s not working, and where you have blind spots. I’ve done it manually, before AI, and it’s super time-consuming. Being able to do it with a tool has been amazing, because you can speed up your creative cycles where you communicate with creative strategists and designers.

Where I think we need the human eye is for taste, judgment, and quality control. It’s very easy to produce a high volume of AI work that isn’t slop, exactly, but isn’t super tasteful — things that are uncanny valley. Everyone can make a static with AI now. Everyone can make a thousand statics with AI. But what feels human? What feels good? What will convert and sell your app? That’s where you need the human eye.

So I don’t think it means a smaller creative team. It just shifts the workload more towards creative direction rather than execution.

Shamanth: Yeah, 100%. And I like that you emphasized sourcing of the assets, the B-roll, the raw assets — because AI can do a great job of that, but that’s not a finished ad.

I’ve seen a lot of people talk about how this makes ads completely autonomously, and the results are not great. Even with models looking very realistic, you still need a lot of human eyes and, as you said, the taste to really understand what to run. I think that’s very true.

Shamanth: To switch gears again to the funnel side of things — Elise, staying with you. A lot of teams try web funnels, see that the early numbers aren’t great, and decide to kill it. What separates the teams that really make it work from the ones that give up?

Elise: It’s very cheesy to say, but it’s ownership and troubleshooting. With web funnels there are so many more things that can go wrong. It’s not just your creative, your app store, your app. You have different pages, you have the payment, you have the sign-up, you even have the setup of the campaign if you’re running it through Meta.

Teams that give up early will run into a problem, won’t see the data populating, won’t see it working, and then won’t necessarily investigate every single part. Which is back to my first point: it’s easier to run web funnels now, but you still need that dedication, and you still need people who will dig into what’s wrong and not offload it to other teams when it gets crazy or technical.

It’s easier said than done, for sure. But the teams that make it work are the ones that troubleshoot and constantly try to improve small parts of the funnel and the flow to improve the numbers.

Shamanth: And that’s similar to what you said about AI creative, because it seems easy. If you were to read LinkedIn you’d think, “Ah yes, this is easy.” Web funnels, creative — but ultimately there are so many parts you need to get right.

Shamanth: Mike, what are you noticing as you work with a lot of your customers?

Mike: One thing is that you’ve got so many different user flows you can go down, depending on what source you’re coming from. Are you going to hit a landing page, or are you going straight from a web ad to the store? There are these initial questions marketers have to ask themselves about how they want to funnel users towards downloading an app and then signing up for a subscription.

And then, how are you going to measure that whole thing? We get a lot of people who say, “We’ve got web-to-app from a landing page, and someone’s downloading the app.” But then: “We have no idea how someone’s getting to the landing page, because Google Analytics isn’t telling us that.” And they can’t bring the two together.

So actually thinking about that full-funnel measurement — how do you go through each stage, measure each step, know where someone originally came from, and get out of platform hell so you know where that revenue came from and can tie all of it back.

It’s not a completely straightforward flow, so it probably does need an expert to talk you through it. If you’re going off on your own, it’s not the easiest thing to implement straight away. And you want to make sure you can measure all the way along before you suddenly scale up. You don’t want to start spending loads of money on web ads if you haven’t got all your measurement in place and you’re not very confident that it’s working well.

Shamanth: Yeah, 100%. You need the conviction that your money is well invested. And I’d add that you either need an expert, or — like Elise said — you need to invest a lot of time and effort into understanding the different pieces and what’s changing.

Andrey: I’d also add that we saw a strong correlation between the number of iterations and the pace of iterations. The quicker a team iterates, and the more iterations they do, the more likely they are to grow in terms of revenue.

Shamanth: I recollect this was a graph I noticed your team put out, Andrey. Do you want to talk about what you learned from that?

Andrey: It’s a dependency between the number of funnels teams launch and the volume of revenue. We saw the relationship between the two.

Shamanth: A very direct relationship. And if I might ask a follow-up: what do you see as the most important types of tests run by teams that test a lot?

Andrey: Basically, you need to optimize the whole funnel and find the right combination for your audience. That’s mostly ad creatives and your storytelling, then your onboarding flow, and then monetization. Those are the three parameters, and you need to find the right combination for the right segments. So you need to iterate on all three simultaneously.

Shamanth: Staying with you — are there teams that shouldn’t try web funnels? Are there teams or situations where it’s a bad idea?

Andrey: I think the most underestimated thing is payments, because payments are a completely different thing from what you have in the app store.

You need to take care of refunds, chargebacks, and payment optimization to get the highest payment acceptance rate. It’s tough. It’s a rabbit hole that can’t be missed at the very beginning, and it has the biggest impact on conversion rate and final performance. It’s a complex thing to build.

So my suggestion is not to try to build a payment stack on your own. It’s a completely different thing.

Shamanth: Right. So if you’re trying to DIY everything, you might find yourself in a rabbit hole — especially with payments, that could be a problem.

Jacob, what are you noticing in terms of folks who maybe shouldn’t be trying web funnels?

Jacob: I think it really starts to make a lot more sense once you’re above a million in revenue and you’re past the 15% commission on Apple. That’s when the proceeds start to pay off a lot more.

That’s usually a good bar in terms of paid marketing scale too, because you probably need to be okay losing 10K or 20K a month on paid ads while you’re testing and optimizing first. To their points, you need to iterate, and you have to assume that maybe for the first few months you’re not going to get it right and you’re not going to be profitable. So you have to have some larger margin. You have to have a growing business at a minimum level of scale before you can invest in that learning period.

And a lot of it is because it’s all driven by paid traffic. If you’re having organic growth, you’re probably not investing in paid marketing — you’re probably doing fine, you don’t have the ROAS constraints, and it’s fine. The questions are: are we above the 15% commission, the 30%? Do we have a decent enough level of paid scale to lose money while we’re in that learning phase?

Shamanth: I tell people this is a lot like DSPs in some sense — there’s going to be a learning phase. Out of curiosity, in that learning phase, what are some of the more common mistakes you see people make? And are there ways apps could accelerate that learning phase, in your experience?

Jacob: I think there are probably three elements of the learning phase.

There’s the acquisition side — are your campaigns different in terms of what’s successful at driving towards a web funnel? I can’t speak to the different setup on the paid side as much; everybody else can probably speak to that better than me.

Then there’s the actual funnel, the onboarding flow. Can you sell the product successfully? A lot of times you’re moving from app to web, and in your app you can easily show the value and the aha moment, because it is the product. On web, you’re still often directing people into the app. So how do you sell the product well? Probably to start, look at competitors. I think we all hate saying “just copy somebody else to start,” but oftentimes that really is what you need to do — and then try to differentiate and iterate beyond just copying, to build your own value.

And then there’s the paywall and monetization tactics. Again, figure out the paywall and monetization tactics that are working decently for other similar apps. Looking at what’s working well in the market is usually the fastest place to start.

If you have a long onboarding funnel — and if you’re in the health and wellness or education category, you probably do — a lot of times taking that as a V1 into web will work all right. You’ve already proven that it converts decently. And then a lot of it is how you improve and iterate faster on web with that funnel, to convert, motivate, and sell the product better. You then often iterate on monetization: the different lengths of products, and the value of the products, which is usually different on web.

Shamanth: This basically echoes what Elise said earlier — there are just so many moving parts, and that’s what’s contributing to the learning phase.