---

About Elise :LinkedIn

FULL TRANSCRIPT BELOW

Shamanth: I’m excited to welcome Elise Zareie to Intelligent Artifice. Elise, welcome to the show.

Elise: Thank you so much.

Shamanth: Yep. I’m excited to have you because you’ve done some very cool and interesting work with subscription apps over the last very many years, and most recently you’ve also been diving into web funnels, which I feel like a lot of people talk about, but don’t really appreciate the sheer complexity of it.

Shamanth: And I don’t think a lot of people appreciate how hard it is to actually get results versus just talk about how theoretically easy it is. And you actually have gotten some incredibly strong results and growth with web funnels, and I’m excited to dive into that and also talk about how you use AI for a lot of your workflows as well.

Shamanth: So there’s a ton that I’m excited to cover. So let’s get started with what your day-to-day looks like. How would you describe what your primary focus is on a day-to-day basis?

Elise: Yeah, it’s a very good question. So what I really love about my job in UA, I’ve been in UA since 2019, which is kind of dinosaur years in UA because it’s such a new vertical, a new kind of job.

Elise: What I love about my job is that it changes every year. My focus is different. At the end of the day, I want the same things. I want performance, and I focus on the same objectives, but the core of my work changes. So in 2019, I would focus on creating audiences and briefing creatives. Now my work is a bit less operational, and it’s more research.

Elise: So my day-to-day, I’ll check performance. I’ll do some creative research. I’ll do some competitor research, and I spend a lot of my time planning tests, launching tests, looking at results from tests, and that can be on campaigns, funnels, and creatives. So it’s more of that roadmap and project management job that I do day to day.

Elise: At the end of the day the goal is the same, I think. But I have more free time with AI tools, for example.

Shamanth: Yeah, 2019, the good old days when we did a lot more button clicking. I remember those days. So, switching to web funnels. That became a big focus for you in the last year and a half or two.

Shamanth: What about web funnels got you interested and pulled you in?

Elise: It was really the demand of the market. So I’ve been working with some health and fitness apps for the past couple years, and I think there was an appetite for building those funnels from clients I had while I was in an agency, and then for clients and apps I worked with as an independent.

Elise: It’s not new, right? Web funnels have existed for a few years. There are apps like Flow and Headway that have been around for a really long time. But because they’re not as straightforward to build and test, it’s a process a lot of developers have kind of shied away from or not really focused on.

Elise: But we’re getting to a point now where it’s very easy to make an app with AI, so it’s just really easy to make, but it’s becoming harder and harder to advertise an app efficiently. It’s not like you just launch a campaign, put a few creatives in it and call it a day. It’s a lot more complicated. So I think web funnels come as an alternative for advertisers to grow their audiences and reach new people.

Elise: And the other thing is, from an operational perspective, there are new tools in the market, like FunnelFox is a really big one, Web2Wave. I think recently Superwall released a web funnel tool, so it’s becoming easier for people to make the funnels themselves. So it’s both. I think something that interested me, but it’s also something I’ve been asked a lot more, “Can you make funnels? What do you think about funnels?” And then in 2025, I started deep diving into it, actually making them myself, and it’s been super interesting. It’s been such a change for me as a marketer to work on this and seeing it work.

Shamanth: And you said in a preparation call for this recording that you had to learn to almost become a part-time product manager to build web funnels, which is a completely different skill set. Talk to me about, number one, what was that learning process like for you? Number two, walk me through the process of being that part-time product manager. What are the different steps involved in shipping a funnel?

Elise: Yeah, absolutely. I think when I started my career in tech, my manager was a product manager, so I kind of learned the very basics, but that was 10 years ago. And then becoming a marketer, I think as a marketer you want to ship things really fast

Elise: You want the results for your test today, yesterday. You’re super agile, super fast, but you’re not aiming for perfection, and I think that’s the difference between product and marketing people I see sometimes within companies, is that product people are so focused on building something that will delight the user, whereas in marketing I’m just trying to bring the user to the App Store and the Play Store, and the rest I will let the app deal with.

Elise: So I kind of had to shift my mindset from one to the other, and then in order to prepare for that process I did a lot of research on people that have built funnels before. Because I understood the goal of what a funnel is, I understood how to launch a campaign with a funnel from a marketer perspective, but I wanted to understand the conception of it, like, what are the different steps?

Elise: So, in a company I worked with that was in education, I bothered the chief of product a lot. I kind of became his little intern. I asked him as many questions as possible to understand the tools, because I worked with a lot of productivity tools I’d never worked with before, and to understand what the steps look like.

Elise: So the first step, if you’re thinking about a funnel, you’re thinking about a website with many steps, and how do you visualize that? You visualize that on Figma. So the first step is to build this mockup of page to page to page to page, get the design right, get the copy right, get the illustrations and the overall flow.

Elise: So you will be working with designers and, if you’re lucky, some UX people. Once you have something that looks pretty and that makes sense, then you’ll be working with an analytics person to make sure that every step makes sense from an analytics perspective, and then you can take those steps and translate them into events.

Elise: So that’s the second step. The third step is you’re bringing that to life with a front-end developer. So the front-end developer will have the Figma file. They will have the analytics events, what events are corresponding to what frame, then they will build it. And then, once you have the frame you can test, this is when you bring in a back-end developer, and this is where I had to really push my technical skills because, again, I’m a marketer.

Elise: I don’t have a technical background. So the back-end developer could make sure everything that was happening on the screen was also happening in the backend. So, authentication, sign up, payments. Once everything is looking pretty but also working in the backend, it’s a lot of QA. Does it work?

Elise: Is it nice? Is it firing to my analytics platform? Is it firing to my payment platform? Are people forwarded in the app efficiently? So that’s something else we don’t really do as marketers too much. I think you launch a campaign, you just make sure the budget doesn’t look weird. You make sure your audience looks right, and then maybe you’ll do some check on some creatives to make sure there’s not a mistake in the copy.

Elise: But that deep work of flow and QA, I found to be super important, but it is something that is definitely a skill shift from what I do day to day. And then once everything is working correctly internally, you launch it with a live audience. So you launch a campaign, and then does it even perform?

Elise: Is anything broken? Are people responding to it? So you have to go through all those steps of building it, and then you release it into the world. And I think as marketers, especially in UA, we’re super jaded because we know that 95% of every creative we test will flop. So I think it’s trying to approach this with an open mind and an open heart and just being ready for it to flop.

Elise: And it’s so much more work than one creative so it can be frustrating. But this was my experience building it from scratch, and I’m very happy I did it because I can really appreciate now working with tools like FunnelFox, for example, where a lot of those steps are automated and you actually don’t have to do that yourself.

Elise: But I can appreciate that I know how to do it in a very detailed way.

Shamanth: Yeah. And that’s a lot of moving parts. That’s a lot of very different steps. And I realize a lot of people don’t quite appreciate the complexity involved. And you’re right, that it’s very different from what you would do as a marketer.

Shamanth: So good on you that you were willing to put on the white belt, you were willing to be an intern almost to learn a lot of these new skills, and I think this is something a lot of people don’t quite appreciate about building web flows. And I would say I had a very similar perspective building apps because my first impression was, “Oh, I’m gonna vibe code apps. It’s gonna be easy. Everything’s ready in 24 hours. Amazing.” And then my realization was very similar to yours, that there’s just so many different parts. Just putting them together is actually hard work. It’s not what they make it out to be.

Shamanth: You just don’t throw a prompt and get an app at the other end. It’s never as easy as they make it out to be.

Elise: Oh, yeah. And that was intense work, and it’s exactly that work of orchestrating. I didn’t do the technical things myself, but I needed to know enough about what

Elise: my backend developer was doing to understand when to call the front-end developer, and it was a lot of me sitting on calls and then researching what everything meant. But you do have such an appreciation for the tech side of things after that. Where you’re just like, “Yes, I actually cannot do this myself.”

Shamanth: Yeah. And I also realize you have an appreciation for the fact that the developers don’t really know a lot about what you want them to do. And you really need to hand-hold them and say, “Guys, here’s why we are doing this. Here’s what we need to do.” And I’ve seen in a lot of companies, when I was in-house, oftentimes the product manager would just leave a brief and the developer wouldn’t really get it, and

Shamanth: you needed to over-communicate, and I think that is so underappreciated, that you have to really over-communicate and hand-hold the developers.

Elise: Yes. And I remember when I built this specific funnel, we had daily check-ins, which as a marketer, I’m living my life, I’m doing my campaigns, I’m doing my things, but

Elise: I needed to keep my eyes on the prize to make sure every area was completed. And then I also had to communicate with product, because with product, they were like, “Oh, why don’t we do this prettier flow? Why don’t we bring some of our app onboarding in it?” And I also had to make them understand I’m looking at it as a marketer.

Elise: I don’t want the super complicated, fancy thing right off the bat. I want a very simple flow so I can prove something, and then we can always add to it. So a lot of moving pieces, but I think it can only benefit you to deep dive into another area of the world of app, if that makes sense, and to just not stay in your lane.

Elise: It can only benefit you. It’s a great experience.

Shamanth: Yeah. And something you also said resonated with me, because you were like, “Look, if somebody’s giving you ideas, you just have to push back.” Which is very different from marketing, where you can run 100 ads and no one’s gonna stop you.
Elise: Yeah, yeah.

Shamanth: So how do you justify running one funnel over the other? What are some of the things you look for? If you could theoretically have 100 ideas for funnels, your product team, your developers, your marketers, your CEO, everyone might come up with ideas and say, “XYZ is doing ABC, please copy that.”

Shamanth: What’s your process for evaluating what to actually do? Because you can only ship one.

Elise: That’s a very good question. In the ideal world, I would do 100,000, but then you also have to think about when you’re testing a funnel, you’re not testing a creative. So it’s not like you can put $100 behind it on Meta and be like, “Oh, it works or it doesn’t work.”

Elise: There is a long learning phase. There are some tweaks you can do in your funnels. There’s so many more variables to validate your experiments that testing one funnel is actually going to be a bit more expensive than testing a creative. Also going to require more work, so we can’t test absolutely everything at the same time.

Elise: One thing I really do when I look at building a first funnel in a specific vertical is that I will look at what the top advertisers do in terms of funnel in that vertical, and identify a dominant funnel or a dominant flow that advertisers in that space tend to do. Because it means that if they’ve been doing it for a little while they have data behind it, there’s a reason why you see that flow over and over.

Elise: I think it’s very similar to creative research. Sometimes if you go in language learning, the ads are looking very similar, the funnels are looking similar, and there’s a reason for that. So when you’re starting with your first flow, you should prioritize the best ones on the market, and then go for it.

Elise: And then if you want to explore other flows, I prioritize it in two ways. The first one is I’m looking at the tone of my creatives. So if I have a dominant tone on my creative, if there’s a buzzword, if there’s a hook, if there is for example, we’re working on a meditation app and sleep issues is what works the best with my hooks, one of the first funnels I explore is a sleep issues kind of funnel.

Elise: Because I know this is what brings users in. I know people click on it, I know people convert from it, so why not build a funnel that follows that theme? And I already have proven data on my creatives.

Shamanth: And how would you identify the dominant funnel? And I say that because, for ads you can see what ad has been running the longest.

Shamanth: You could download the app, go to their webpage and look at the funnel, but maybe they’re testing five other funnels at the time. So how would you identify the dominant funnel?
Elise: Time. And I’ve worked with people that make themselves little libraries of links and competitors over time, and if you have several advertisers and you see the same kind of funnel, then this is the dominant funnel.

Elise: I think just like with creatives, there will always be something that is kind of out of the norm, or people will try a design that’s completely different or a flow of questions that’s completely different. Usually, if you look at a lot of funnels, and it’s exactly like creatives, spend a lot of time looking at your competitors, spend a lot of time going through other flows to understand what works.

Elise: Usually it’s not like I’m going to see exactly the same questions or the same colors or the same designs, but I’m going to see an overarching structure that tends to be the same. For example, for health and fitness, the first part of the funnel, the first 10 questions will be like, “Let’s get to know you better.”

Elise: The second part of the flow is going to be, “Let’s answer those questions so I can craft your plan.” There’s a goal to answering the questions, then there’s a sign-up to receive the plan, then there’s a paywall, “Your plan is ready, activate your subscription to start your plan.” So when you start seeing patterns in the funnel that tend to look the same, and then in some verticals, like language learning, everything is extremely close even by design, this is when you can identify an out of the norm idea or something the team tested versus the thing you keep seeing over and over.

Shamanth: So sounds like you read funnels for fun and pleasure. That’s the secret.

Elise: I do. It’s a bit sad. Yes, but I do.

Shamanth: No, no. That’s why you’re awesome at what you do. Yeah. Excellent. I tell people I watch ads for fun and pleasure, but I think it’s the same thing.

Elise: Yeah. A thing in common I see with people in this space is that we’re all a bit obsessed with things. We all get very obsessed and mine last year was just all about funnels. So exactly like you with ads. I spent hours just looking at it and trying to deconstruct it. But yeah, go in the ad library, click on things, make yourself a little library, and then you’ll have an idea of what works and what doesn’t work.

Shamanth: Yeah. What I’m hearing you say is you have a pattern recognition so you know exactly what the overarching theme is that you can use for your own funnels. And that’s oftentimes a starting point for something you do and work with.
Elise: Exactly. Exactly.

Shamanth: So let’s just say you have a funnel, you’ve tested it out, it’s working. You’re like, “Okay, how do we improve this?” What are some of the big places you look at?

Shamanth: Assuming the first one you have already built based on the dominant funnel. What are the levers you look at for improvement?

Elise: So this is where I see teams I work with panic because they think that they have to deep dive into every single question to actually move the needle. Those funnels are sometimes 60 steps. They get very, very long.

Elise: But the questions, once you have it out, don’t matter as much. There are three steps I’m looking at to move the needle. The first one is the landing page. The second one is the signup process, and then the third one is the paywall. So the landing page is when you click on the ad, this is what you’re landing on.

Elise: And then what I’m tracking in the backend with analytics is how many people that landed on the landing page click on the second page. Like, how many people did we retain? And based on that percentage, I’ll have a good idea of whether my landing page is any good. And I think the goal is trying to get as many people in from the first page to the second page, because once they’re on the second page, they kind of started the flow and there’s a sunk cost fallacy where they’ve already started answering the questions, so it’s not like they’re going to necessarily drop off once they started answering. But it’s really about getting that first page right.

Elise: So one big thing is having a logical link between your creative and your landing page. In an ideal world, you’ll have a landing page that is customized to your creative. Very simple in theory, harder to implement. You can implement that through UTM tags, for example. You can put your creative in a specific category and a theme.

Elise: What I’ve done with an advertiser I worked with is that I categorized the top creatives in four different themes and then based on the UTM tag, there were four different versions of the landing page to have a reminder of the language that was used in the creative, not necessarily the visual.

Elise: You can also take the visual and implement it as an image if it’s aesthetic. But you need that very cohesive connection. If, for example, the ad is about your cortisol spiking and you click on it and the landing page also says, “Here’s this to reduce your spike in cortisol,” it’s just going to create that element of trust where you’re like, “Oh, this is actually made for the thing I clicked on. There’s a direct link, therefore I’m just going to go into the flow.”

Elise: But if it’s completely unrelated, it just creates some friction. The thing also about landing pages is you don’t have the testimonials from the App Store or the Play Store. You don’t have the legitimate platform people trust and land on, and they know it’s a real app.

Elise: It’s like a random website at the end of the day. So the more trust you create on that landing page to make people understand, hey, this is not a random website, this is very linked to what you clicked on, this is trustworthy, then the better. So you have to balance that trustworthiness, and you also have to pique their interest at the same time.

Elise: And usually, if I launch a funnel, this is the first thing I’m looking at experimenting, because it’s much easier to experiment with one page rather than an entire flow. So I’m trying to tweak and A/B test on this first page first to get people in. And there are good benchmarks. I think you want something above 40% usually from first page to second page, then you’re in a better zone.

Elise: If you’re below 40%, this is where I would spend more time tweaking it.

Shamanth: Interesting. So the funnel itself might be great, but it might underperform if it doesn’t match the ad.

Elise: Exactly. Exactly.

Shamanth: That is so interesting and I can see why a lot of products might lose out because they’re spending a lot of time optimizing the funnel, but not really looking at the continuity with the ad.

Elise: Yeah. And again, it doesn’t need to be this really granular thing where you have one funnel per creative because it gets really crazy. But you can categorize your creatives in themes, or in main hooks and copy and then have a few different versions. And this is a small thing, but it goes a long way.

Shamanth: For sure. But if you’re also spending six figures, seven figures, you could technically have eight, 10 landing pages certainly. And we’ve done something similar with custom product pages on the app flows. It’s very similar, but of course this has many more moving parts on the web as you can imagine.

Elise: It’s the same thing as CPPs basically. It’s the same thinking process.

Shamanth: So Elise, you talked about the very many moving parts and the complexities of building web flows. Talk to me about how AI has changed or impacted your workflows.

Elise: So the thing with funnels is that they’re a piece of content that will be consumed by a user and as a marketer or a project manager when I’m leading this process I can deal with the technical complexities of setting up the funnel, but I think what really matters is the content of the funnel, and it takes a really long time to build content that is logical, like questions that really achieve something and create an emotional connection between the funnel and the user.

Elise: So AI has really helped with that copywriting task. I find it quite hard to deep dive into a task of building this copy and this continuous flow because as marketers, I think we tend to want to execute really quickly and test things quickly. So I’ve used AI to basically build the whole copy and the general flow of where that funnel’s going and what it is saying.

Elise: And basically my methodology is that I take the top creatives I work with, so if they’re static images or videos, I take screenshots of them. I put them on Claude, and then I put also a description from the app store, like what the app is. And then, I ask Claude to categorize different themes based on those top creatives.

Elise: So for example, theme one is social anxiety. The second theme is anger. The third theme is ADHD. It categorizes the different creatives. It recreates the tone and the feel of the creatives, and then I’ll put, for example, the Figma screenshot of a competitor funnel, and it creates a full flow of questions and assessment tests and basically everything I need as a basis to build the funnel.

Elise: And the output is going to be an Excel file with each theme with different questions mapped out that I can then take and populate in different funnels. So it’s really doing the work of that copywriting, creative, emotional side that I think would be super time-consuming and honestly would require skills that would be beyond what I’m doing.

Elise: It’s also great because it allows you to test a lot more themes and funnels than you would necessarily. I think when you’re working with web funnels, you’re doing two things at the same time. The first thing is you usually have your dominant business-as-usual flow. You keep tweaking, you keep making iterations of it, so through A/B testing on the landing page, the signup, the paywall.

Elise: And then you’ll test completely different flows where you’re trying to test a specific theme, or you’re trying to talk to a specific persona. Before AI, having something like that on the roadmap would take over a month, and then you’re not even sure it’s going to work because it’s further away from your business-as-usual flow.

Elise: It’s kind of like we don’t really know. When I tested that with a specific client, one of those new ideas worked out of maybe seven or eight. You wouldn’t do that if you were doing the funnel from scratch. But if you’re doing it with the help of AI, not only on copywriting but also on assets like images, PNGs you’re using in the funnel, illustrations, etc.

Elise: You’re able to actually test something quick, and you’ll know in two weeks of testing if it’s scalable or if you should kill the idea. So you’re really speeding up your ability to test those complex projects.

Shamanth: Yeah. So if I’m hearing you right, number one, you can move faster. But also you can come up with emotionally resonant and deep copy

Elise: Yeah

Shamanth: that would be harder for a marketer to come up with. Also just because there’s just so many screens, so much to think through. The sheer complexity can be crazy. Interesting. And, now that you are able to ship more funnels in the same time, what are you able to unlock? What are you able to do that you weren’t able to do before?

Elise: I’m able to basically test more. You’re able to test iterations much faster. Like for example, I had this funnel that was very heavy on the illustrations. I wanted to test it with a different style of illustration that was a bit more mature, a little less Gen Z.

Elise: Before I would have had to brief a designer, it would have taken a really long time. I was able to test it really fast in a couple of days. I was able to switch all the illustrations that were all AI generated, and then I was able to validate something and scale it. I think what it really unlocks is the ability to test faster and just crack new audiences as well.
Elise: Like for example, with an app I was working with, one of the wild cards was building a male-specific funnel for an app that usually has 80% female users and 20% male. They are high LTV, but it’s just harder to reach them due to the Meta algorithm. The ads in that vertical skew female.

Elise: We were able to launch it in, I think, two weeks or something like that, and then scale it and prove that it works. Whereas before it’s not something that would have been prioritized because it was much more far-fetched of a hypothesis to actually make it work. So yeah, it’s really that testing aspect that is super interesting to me and really exciting.

Shamanth: Yeah, you can test faster, you can unlock audiences you wouldn’t have gone after before. You can test ideas you wouldn’t have been able to test before. Interesting. And specifically, if you look at AI apps, what are some of the learnings that you have around web funnels that could be very specific to AI apps around the landing pages, the questions, paywall copy? How is the strategy or the planning different?

Elise: So I think in the business community, we’re very excited about AI because it’s made our work so much easier and faster, and I’m very excited. But for the general public, and when we work with consumers, I think we have to be careful not to alienate them.

Elise: We have to remember that there’s a human at the other end of the phone, and what they’re seeing is something that we should check, and we should make sure we’re not creating some kind of rejection by using too much AI. And what I mean here is when you’re using AI visuals, for example, that are a bit too uncanny valley, I’m seeing that with video generation sometimes.

Elise: Yes, as a marketer you think, “Oh, I’m being super agile and quick. I’m testing all those new creatives,” or “I’m testing this new funnel with all those new pictures.” But you also have to ask yourself when the audience sees it, does it feel trustworthy? Does it feel too AI generated? Am I harming the quality of my product and maybe my brand by putting too much AI generated output in front of them without really acknowledging it or noticing it?

Elise: So what I encourage is, if you have the ability, get some real user feedback on flows where you use AI imagery, and you’ll actually be surprised by how people respond and how suspicious they can be. Because we are getting excited by the capabilities of AI for image generation, but people also get more and more suspicious about it.

Elise: So I think being tasteful and careful about how much you’re using it. When I think about creatives and AI, for example, I’m seeing really good examples with B-roll AI, but talking heads, for example, is still a bit too uncanny valley, and it’s something that is completely being caught by people that are being served the ads.

Elise: So I would say use a human check and really think about it, out of respect and out of wanting to have this connection with the users, rather than trying to ship something very fast without thinking about it. So that’s been my position working with different AI apps and working with AI to make creatives or web funnels.

Shamanth: Yeah. So what you’re saying is just because marketers get excited doesn’t mean everybody else will get excited. Also, I think a word that you’ve used multiple times through this interview is human, which is a very interesting choice of phrasing. I think you just talked about this when you talked about the time, effort, and thought that goes into the web funnels.

Shamanth: Now you’re also talking about how you want to be nuanced enough around AI products, and ultimately, I think what you’re talking about is really how important it is to be human as a marketer because that’s something that’s easy to forget when you can throw 100 ads into an ad set and go home and sleep.

Shamanth: And I think that’s the meta point I’m taking away from a lot of this because even if you’re using AI, even if you are building a ton of funnel assets, I think the most important thing is to put in the thought around the interconnectivity of the different parts that will really build the system, and I think a lot of that is just being human, and I think that’s what I’m taking away from a lot of this.

Elise: Yeah. No, exactly. And you’re never safe from a hallucination. A part of my day is trying to catch hallucinations from the AI engines I’m using, either on the data or on the insights, and I think they’re incredible tools, but you need to remember the incredible skills of developers that know the overall structure. And at the end of the day their craft is also making sure there’s no hallucination. So I’m super excited about using AI, and I use it every day, but use it with care.

Shamanth: For sure. Yeah. And that’s perhaps a good place for us to start to wrap this interview. Those wise words. Thank you, Elise, for sharing your wisdom, your knowledge, and just also opening the curtain to show how the sausage is made, and it’s not as easy as it looks. Before we wrap, could you tell folks how they can find out more about you and everything you do?

Elise: Yeah, absolutely. Just very simply on LinkedIn, you can add me at Elise Zareie, my name, and you can always message me.

Shamanth: Wonderful. We will link to that. And thank you so much for being on the show.

Elise: Thank you so much for having me. Such a pleasure. Thank you so much.

Shamanth: Wonderful. Excellent. Take it easy.

Elise: Thank you