How Cash App Scaled to 9 Figures — Without Wasting a Dollar

It's not like, hey, how do I maximize my budget? It's how do we maximize growth of the company overall?

Measurement is nice and all, but it's Robin, and really it's the creative, it's the offer that is Batman.

Within 6 months, we were operating at a 9-figure run rate from a paid marketing perspective.

Insights into the innovations and revolutions of the marketing world. Welcome to the Marketing Scientists Podcast. Hey everyone, and welcome to another episode of the Marketing Scientists. Today we're gonna talk to Dyan Khor, who led growth marketing at the Cash App and then also at Speak, one of the fastest growing YC companies backed by OpenAI as well. This is a special episode. We talk about all things incrementality, marketing measurement foundations, and everything it takes to stand up this infrastructure from scratch and how it works for brands that are scaling beyond just a handful of channels. So definitely tune in if you're interested in that topic. Dyan, this is awesome. I think this is the first time we have a customer or former customer on our podcast, which is awesome. So we worked together, we'll leave the details of that, but I'm super excited to actually learn about all the things that you've been doing at Cash App. And then obviously, you know, when you were at Speak and scaling that sort of company that's now unicorn, and now you've started your own growth consulting practice. So there's a lot to cover. Welcome, and I'm gonna jump straight in and ask you a bunch of questions, but thanks for joining us.

Yeah, thanks for having me, Pranav.

All right, okay, so I'm gonna start straight at the start of your, I think it was the start of your career. So you started at Cash App, right? That was the first thing and then you after Cash App—

No, that was not the start of my career. That was probably a good like 6 jobs in, but definitely the biggest job of my career. But yeah, happy to dig in there.

All right, so you did Cash App, you did, you know, a hot AI startup. I feel like Paul Graham talks about Speak all the time, which is awesome. And then, you know, now you're advising other startups and scale-ups on growth acquisition. So here's a fun one. If your journey were the title or, you know, of a book or a documentary, what do you think it would be called and why?

Oh my gosh. I'm so bad at naming things, which maybe is a bad thing to say as a marketer, but maybe that's why I'm a growth marketer and not a brand marketer. There we go. Challenging question. Really hitting it hard at the beginning. Just buying myself time as I just like— This is great.

This is great. Take as much time.

But no, I mean, I think it would be, um, like something along the lines of like finding the multiple lanes of growth. So I think that like growth to me is about the synergies of creative analytics and strategy. And so, um, I kind of stumbled into this job. I originally thought I wanted to be in analytics and then I realized I was missing this intuitive component. And so I think that that's one of those rare careers that you can find. I didn't know that growth existed. I also didn't know marketing existed, to be honest. And then I was looking for a data science job and no one wanted to hire an inexperienced data scientist. And they basically were like, can you build like an attribution model or something? And I was just like, I have no idea what that is, but sure, I'll look it up and I'll figure it out. And then as I was doing that, I was like, oh, underneath all this attribution stuff, there's the marketing campaigns. In order to do the marketing campaigns, you need to figure out all these And so, it'd probably be about like kind of peeling back those layers, but I'd have to think more on a succinct title of what they're all kind of like bringing those synergies together.

I like it. I like something about the multiple lanes, right? Because there are, like, it's not an either/or, it's really you're tracking each lane sort of independently and all of them have to, so that's, we'll work on that analogy. I like it. I like it.

Okay.

I had no clue that you started more on the analytics side. And I was, I saw this post yesterday from the CMO of Meta, who by the way, also has a title of CMO and VP of Analytics. Did you know this? This is like wild to me. Yeah.

Sick title.

That is a sick title. So I posted about it on LinkedIn yesterday, and I just thought of that as you were saying this, that I think that the future CMOs are kind of going to be like that, where they're sort of multifaceted. Both on the analytics side and the marketing side. And yeah, so that's great. Okay, so I want to jump straight into Cash App. Were you like part of the founding sort of marketing team or the growth marketing team at Cash App? Like, is that like, tell me what was the context when you joined and then I have a bunch of follow-ups for you.

Yeah, so I was not part of the founding team of Cash App, but part of the founding team of growth at Cash App. And so the The interesting thing about Cash App is it's a peer-to-peer app. And so there's a lot of product-led growth when it comes to that. And so think about referrals, and then there's a lot of kind of activations that people will do like on the ground, and that was more brand-led. And so they'd never had a growth marketing team. They had some people on the brand and social side, and they'd had some people on the product side who were thinking about growth, but no one from a growth marketing perspective. And so As kind of the LTV increased, when you introduce more products like card and banking and investing and Bitcoin, there was kind of this question of like, hey, is the LTV high enough to kick it into next gear? How do we hit this next phase of growth? Because Cash App already had, you know, millions and millions of monthly active users, but there was this, these gaps in the network and tons of users left to acquire. And they wanted to, you know, beat Venmo, beat Chime, all these kinds of things. And so Marketing and growth marketing became super top of mind. And so the product organization created a business organization. We had our head of business and I was his first hire. He was like, I don't know how to do this marketing piece. I'll kind of set up, like, he was like, maybe we should have AppsFlyer. So he started setting that up with engineering. And as I came in, it was like my first task to figure out like, was this set up right? How are we going to instrument our first campaigns? How do we bring it from zero spend to a million dollars spent and continue scaling it? How do we build attribution? iOS is about to change. How will we continue to build attribution? How does this start to play into brand? So just lots of pieces kind of coming together over the years. But yeah, the start of it was just how do you build this machine soup to nuts? And kind of goes back to that analytics piece. A lot of it was analytics to start because there was such like a strong focus on return there since it was an existing business that was doing pretty well.

There's a couple of things that stand out to me, right? Which is the LTV had to reach a certain point where the business is like, hey, now it makes sense to kind of go the inorganic route, right? Or the more sort of, let's put more media behind it. And the second is, you know, you as the first hire with this really strong analytical background coming in and kind of assessing whether we have the right infrastructure to then go and, you know, do a bunch of stuff. So 2 follow-ups there. What were the first few bets when you come into an organization like that? High volume, good PLG, good brand, like everyone's talking about it. And what do you do? How do you even think about, oh my God, how am I going to, like, I would be nervous going into that role and be like, oh, what am I going to do? Right? How do I prove myself? Right? Like, what is the first, like, 5 experiments that we do? So what were the earliest bets and how did you think about them? And would you do something differently now that you have some hindsight looking back at, you In the first few years?

Yeah, so the first bet was essentially just paid marketing. And I think within paid marketing, there were obviously a lot of nuances. It was building out attribution, but it wasn't just about like your typical paid attribution. It was about understanding from the start, how do you attribute paid within the context of things that are more product-led, like invitations or referrals? He also had payments. So it's like, if I just sent you $5, that acquired you, like how do you make these things as measly as possible from each other. And so that was actually probably more of the upfront work, but I wouldn't say analytics was ever a bet. It's the foundation so that you can take the bet. And so I think that kind of starting from that analytical foundation of like the LTV to CAC, like how can we kind of prove this out and like build a foundation? But that CAC then has to be really something that people can believe in that allows you to really accelerate that bet in the long term. And so something that I'm really proud that we invested in from the start. But then on top of that, there was kind of this incrementality question from an organic standpoint. So it was something that we started thinking about very early on. So once we kind of scaled beyond the first couple million of spend, it was like, how are we going to measure incrementality? The iOS questions were also coming up. So the next kind of bet from an analytics standpoint was actually starting media mix model from like a few months in.

Oh, wow.

Very early, this was 2020. Not a lot of apps were doing that at that time. And so kind of figuring out how would we not only like find the right partner for this? Paramark definitely didn't exist. I make it 2020. But how are we going to like actually use this data to make day-to-day decisions? Like you know when you're working with a traditional partner that's providing like quarterly outputs that come like months after the quarter has ended. Like what does this actually mean? How do you make executives who just got comfortable with what marketing is, comfortable with a media mix model as well. And so a lot of those pieces, I think that the bets wouldn't be any different because the foundations there were essentially to go with paid first and to prove how much paid should be layered on top of these referral pieces. I think there were, eventually my team took on a lot from a referral standpoint as well and a lot of experimentation on that front. And I think that that's a really valuable piece for a lot of teams to kind of figure out like, how do you break out of these silos and either cross-functionally or otherwise, like think about the whole picture of what is driving acquisition and make sure that the business is investing in the right things. It's not like, hey, how do I maximize my budget? It's how do we maximize the growth of the company overall? And then once we kind of figured out the scale of those things, within 6 months we were operating at a 9-figure run rate from a paid marketing perspective. We had maximized our referrals. We were like, hey, where do we kind of go from here? We'd started with core channels from a paid marketing perspective. So the next set of bets were, how do we bet on other things like connected TV, for example, and things like that? And that made us all the more glad that we'd invested in incrementality measurement to begin with so that we could make those bets in a way that were a little bit more apples to apples.

Gotcha. That's all very helpful. I have 2 questions there. You mentioned incrementality right off the bat. Was that something that came from the finance team, the executive team, or you and your, you know, GM or head of business? Like, how did that conversation even start? Because most people are like deer in headlights when, you know, like especially back in 2020, it wasn't such a popular thing to talk about. So I'm curious, you know, how that ended up happening.

Yeah, so I think it was a pretty natural synergy between myself and our head of business. He had a lot of questions for me about like the iOS 14 changes and what that would mean. So it's like, hey, you're building this really beautiful user-level attribution. Like I was literally coding user-level attribution. He's like, but Diane, this is going to go away on the iOS side. So like, what are you going to do about it? And so he was like, how else could we get around this? Is it something we could do from a data science perspective internally? So we started kind of like bouncing ideas of like, what could we do from a causal basis? And then it also started to get to the point where we're like, oh, like some of these other companies outside of our space are doing this media mix modeling thing. And it also would solve some of these other questions that we had existentially talked about from an overall incrementality standpoint, both for paid and for referrals. So it's like, would, how do we wanna think about bounties for referrals? Would people have joined anyway? Those are questions that had always existed as well. And so we kind of figured it was a two birds, one stone kind of thing where I think if the iOS conversation hadn't happened, we probably would've waited longer on incrementality. I think it's interesting that it hadn't necessarily come up for a lot of other teams as easily as it did for us. I think probably because of the maturity of the business and the interest in incrementality outside of it as well.

That's very cool. It's a very like business-first type of thinking, which is always, you know, and I keep telling marketers, put your investor hat first and then like take it off and put your marketing hat second. So going from a couple of million in spend to a 9-figure budget, like fairly quickly, Like I have to imagine that was search, that was social, right? Like it was basically Meta, I'm assuming. I don't know if TikTok was big back then, but like Meta, Google, like those are the primary channels and you just found wins right off the bat?

Yeah, so we started on primarily Google and actually the ad mix is social. We were running TikTok, Snapchat, Meta to some extent. That actually didn't work very well for us to start. It was an interesting one. It was around the time where Meta made the changes to their measurements. And so we were struggling with like the changes to AAA campaigns, no historical data, like all these kind of interesting pieces. But, um, yeah, it was mostly the core channels. And then when we started to expand to maybe say like an 8-figure run rate, we started putting DSPs in there. So just like display ads programmatically.

Right.

Uh, and then once we kind of capped out on that, we were like, okay, well now we're pretty close to this 9-figure run rate. Uh, and then you start thinking about the marginality of it. And that's where we started, uh, going like, is this marginal dollar actually smarter to spend on Google or is it smarter to try something like TV that's going to be drastically different? Is it podcast sponsorships? Is it influencer? Is it something else? And so that's kind of the natural progression. Um, I do think Cash App obviously had 2 very like unfair advantages from a paid scaling perspective. One is that it's a utility app that almost anyone in the United States can use. The second one is, or maybe there's 3 then. Second one is that it's free. And the third one is that there was a brand behind it already. And so I think that that allowed us to spend a little bit faster than the average business would. But I think like kind of having a portfolio allocation approach and saying like, hey, like how, what's doing well? How do we allocate across these things? And then like as we allocate, How is the marginal dollar performing and what more can we do? What more placements can we max? What more levers do we have to pull was what allowed us to scale pretty quickly. And then from there it was like, uh, how do we optimize into optimal performance?

Got it. Super helpful. I like how you talk about those 3 things and I completely agree with that, but it also brings up a challenge, right? Because you acquire a free user, may or may not work out. And so I'm imagining you had some pretty sophisticated data science to predict some type of LTV, or is this user going to be good or bad? Or how did you, like, was that again, it took a lot of time to kind of perfect that, or you had a good inkling of the some aha moment, and when you reach that aha moment, you're like, you're going to be good.

So this was actually a really interesting thing, and maybe this would have been an interesting discussion for bets, but at the beginning, so Cash App has a variety of products. You have peer-to-peer card banking and investing Bitcoin. Those are kind of like your main lanes. And you could say like, hey, what's the most obvious thing to advertise? And one of the main things that people would say is like, well, Cash App is a peer-to-peer service. Like, you would advertise peer-to-peer, right? Peer-to-peer is free for users. But the other thing about peer-to-peer is that you're not going to use it unless your friends are already using it. You're also, if your friends are already using it, they're probably already going like, hey, Pranav, like, you need to pay me back, like, join Cash App, or I want to pay you back, join Cash App, you know, that kind of thing. And so, The referral mechanism is already there. And so I think the natural intuition of like more of our spend would go there was like the first thing that was said to me. As we thought about it, we're like, actually, that doesn't make as much sense. And we saw that come through in the data. And so when we looked at, uh, like what was performing out of like the first sets of ads, we ran everything. We're just like, let's just make like basic value prop ads for everything and just run it to see what happens. We looked at the entire funnel. We started out optimizing campaigns towards some Because sometimes people don't take these actions until a month later. You and I have had these very deep conversations about the latency of events before, but the latency of Cash App events was sometimes in the order of months. So it just wasn't always possible to optimize towards those events, because if someone doesn't want to send money today, they're not sending money today, no matter how many push notifications I send them, right? And so it was just one of those things where we were like, okay, do we optimize towards the top of the funnel? And then how do we understand the value thereafter? And we're like, we'll just look at it post hoc and let's see what's like driving value and maybe test events along the way from there. What we found is that, uh, when we advertised the card, that was actually the thing that was driving the most combination of people who ended up what our main indicator was, which was moving money. So making like some kind of financial transaction, which could be across any product.

Mm-hmm.

Um, and having the highest LTV. And it also was actually a lower CPI, 'cause I think like people were like, oh, this is something new, this is something interesting. And so that's what ended up really working for us. So it was kind of this like diagnosis understanding of like what would be useful to advertise. But also when you advertise something, that doesn't mean that the person's going to go actually do the thing that you asked them to do. So it was a continual push and pull of like, hey, as we change our strategies, as we change our creative, how is our LTV on average changing? Is it because our product attach of card is going up? Is the distribution going up? Is the average LTV of a card user going up? If we talk about these value propositions and just getting really granular with the user-level data and the segmentation to see where we could continue to push the LTV up without necessarily like pulling the CAC lever hard.

This is great. So much gold in there because I like 3 things that you called out. One is just because your base product is free, doesn't mean that you have to advertise with that free value prop. In fact, I would argue we've seen the same thing with other PLG companies where you got a free plan. You probably don't want to advertise the free plan because, hey, people are already coming to you for that free product. You want to take what are the things that are going to get them to actually do something more valuable and advertise those value props. Again, this is not a cookie cutter rule, but I've seen every PLG company struggle. with activating ads because they're driving everything to PLG. So that was interesting. The second thing was, you know, this constant sort of look at what is that, you know, actual activation metric. And you said, you know, some type of money movement, but again, most companies don't have a very clear sense of like, this is the activation moment, especially if you're in that, you know, free freemium sort of environment and people don't invest enough time and energy in getting that, you know, You know, really figured out. What was the third thing? I forgot the third thing, but those are two big ones. All right, so, you know, you came at it from— sorry, the other— yeah, now I remember the third thing. You kind of just brushed it off and you said, we tried all the basic static ads for all the different things. That's really cool, right? Which is you did not just go in with an opinionated, like, oh, I know exactly what's going to work and I'm going to do that. You tested a variety of things and then you went with what's actually getting picked up by the algorithms and also then showing up in results, which again, if you're starting from scratch, great strategy to put lots of bullets out there and see which one actually hits. Maybe that analogy is not great for military usage, but lots of candy out there and see which candy people may— maybe even that's not a good one, but anyway, you got the point. How closely were you sort of thinking about creative early on? And I wanna transition a little bit to creative, right? Because I say this very often to people, like measurement is nice and all, but it's Robin and really it's the creative, it's the offer that is Batman, right? So when you think about, you know, growth marketing or sort of acquisition, how do you think about sort of making sure that you have enough creative or the room to really sort of go, you know, broad and deep in terms of that? Like any tips, anything that you saw working really well? Now, we can talk about Cash App, you can talk about Speak, you can talk about, you know, all the other roles that you've had.

Yeah, so I think there's a couple pieces. When I joined Cash App, Cash App already had a really strong brand, and a huge part of that was design. And I think Cash App to this day is like known to be an incredibly design-led company. And so there was an interesting kind of like tension to pull on of like, how do you integrate performance into an extremely design and brand-led company? Um, which like, there's some huge pros to that. Like you have like these like wonderful, like 3D assets to pull from and like all these like really descriptive, like brand stories and all these things. But they weren't used to, you know, like putting a bold CTA on something or like a bold value prop. And so there was kind of this like very direct involvement of being able to build the relationship, but also, uh, just kind of say, how are we going to approach this? And the interesting thing there in particular was that Cash App had a lot of creative directors, but not a lot of like hands-on designers who were producing like production-level assets. And so we worked with some agencies and things of that nature. So it was a balance of like having those creative stakeholders push the creative direction, but then being able to bring in the best practices and finding the balance of like letting it be a feedback loop and saying like, okay, here's where I'm willing to compromise. Here's where we need to test this. And then like after we test it, we can come back and say like, Was that worth it? Were we both willing to compromise? All of those things. But I think like across the board, as things advance from a marketing perspective, creative is probably like the most important lever you have to pull. Like analytics is the foundation. Once it's there, then you have creative as your lever to pull. Like strategy is kind of like an overused word and like campaigns like is like the third lever that I'll talk about sometimes. But it's like, I mean, there's only so much you can do in the campaigns, right? It's like change your optimization event, like Okay, like you're not doing that every week, right? And so the main thing that you're inputting into it is like either you're shifting your budget around or you're changing your creative. And usually the creative is a much bigger lever than shifting your budget around because you can only do that so many times. And so I think just being really involved with those teams is important, but I think that's a skill that I've developed over many years of my career of like creatives don't talk in, you know, mode dashboards or Excel spreadsheets. So being able to say, Like, how do you bridge these things together and make the work really rewarding for them? Because it's something that drives a lot of value for the business, but it's not always the first priority. But yeah, kind of like distilling it down and making it really digestible from a performance marketing standpoint. You can do things like creating naming conventions and parsing all these things out. Like, we're going to label all of our different hooks. We're going to categorize influencer versus UGC. We're going to categorize our different products and we're going to say like, okay, like for one, at a production level, are we producing enough of the assets that are doing well? Which kinds of assets are doing well? Is it certain combinations of these things? And so I think that there's a very meticulous approach there. And then once you have that information, it's no longer this subjective, like, well, we look at a million ads and like this one did well. Why did it do well? It's like, oh no, we mean, we have all these factors to it. And it's like, hey, creative team, like we really liked when we talked about this value proposition for this product. And we had it in a static ad, but like, how might we approach this for an influencer ad? Like, we haven't ever had that work yet. Do you have any ideas for that? And then you can really have a generative conversation.

Oh, that's amazing. Okay. So posing it as like, you've done the work to structure the data, to group the data in a way that would be actionable. You've got the insights from incrementality, from MMM to say, this one seems to be popping. You take that back to your creative teams like, hey, can we do more of this? Or can we do more variants of this? to unlock even more growth. That's awesome. Did you have sort of a structured process to engage with creative or was it sort of more ad hoc? And looking back or even guiding teams now who are just starting on this journey, what would you recommend in terms of the interaction between sort of the media buyers and the growth team and the creative team? Like how should they work and operate? Like is there a cadence? Is there any best best practices that you have found?

Yeah, so I think it depends on the pace of business. So like a business like Cash App, we weren't refreshing creative nearly as often as like a business like Skeet was. I think despite spending more, Cash App, you know, it's like, it's a financial app. There aren't that many like net new value propositions. We had a good amount of static and animation that we were able to leverage. And so it was a pretty different cycle. And so the kinds of meetings that you have in those cases are a little different. Like you don't need a weekly performance meeting with your creatives, but you do need to figure out like what the maybe monthly or quarterly planning cycle looks like, what the resourcing looks like. And so it was kind of higher level planning that was going on when you have a big org like that of like, okay, if we need over the course of this quarter X many assets and this is the resource we have internally and these are the assets that have been working well, who are the partners that we wanna work with and who's going to be accountable to those on a day-to-day basis? I think for a company like Speak that was working on a lot of UGC, which tends to fatigue a lot faster, there was a very fast feedback loop of ideally talking about these things up to twice a week of like, here's how we're cutting the sprint versus here's the results. And so just having that rapport between the designers and the marketers. And I actually see this increasing trend in a number of businesses where there's actually an entire role that's centered around this of kind of delivering the insights from a creative perspective. So it's like somewhere in between where this person isn't necessarily managing the campaigns, but they're thinking about creative testing strategy, the insights that come from it, the like how you resource creative and sitting in between those 2 teams. And I think it's an incredibly valuable one. I think I accidentally did this job at the beginning of my career, but it's not one that I had seen as a role very often. And I hired someone to do essentially this role at Cash App and didn't know what to call it because it's not something that I'd seen in a lot of teams before. But I'm starting to see more and more people hire into it. And I think that that's going to be a trend as like we talked about, it's such an important lever. And so it's just like, well, if you don't need, you know, like 15 junior media buyers anymore, you do need someone though who's accountable to what is this roadmap, how is it doing, how do we bring people together around it?

You are full of insights. So you naturally led me to my next sort of question, right? So you talked about sort of hiring. There's a couple of directions we can go. You talked about things that are changing because you're not going to go hire 15 junior people. I'm assuming you're referring to all the AI stuff. The second thing was also like, hey, you know, it's, things are changing quite fast and it's hard to define what these roles sometimes are. and you're just kind of figuring it out, right? You either spike on data or you spike on creative or you spike on like some other thing and then the rest you fill in the blanks. And you've also been doing a bunch of mentoring and coaching. So what are you seeing from people who are coming to you for advice, for coaching? What are the common questions that people are asking right now in terms of what's going on with the marketing profession? And if there's any advice that you can give early career people or even, you know, people who are mid-career who are looking for guidance on like, hey, what the heck is gonna happen with this profession in the next like 2 years? What are your pieces of wisdom there?

For sure. The coaching stuff has been super interesting. I think that like it's always something that I enjoyed in the job and now talking to people who are like coming from all different places, it's been really valuable. So I've always enjoyed the hiring side, the mentoring side, and this is like kind of the outside perspective on it. But I think that in particular, there are a couple things that stand out to me. I liked what you said about spiking. That's a term that we used at CashM a lot that a lot of people have told me is confusing, but it's the idea of what is your superpower? Where do you really go off the charts? And I think that finding people who are super spiky in one area is actually something that I don't believe in. I think that finding, though, people who are evenly across everything is also not a good solution, what I've generally found to be true of the best players on the team is that they spike in 2 places. Maybe one is a little bit stronger than the others, but that thing we were talking about earlier with marketing being 3 core lanes, at least from a growth marketing perspective, you have to have some complement to the other. These days you can't do creative without being, you know, somewhat in tune with campaign or analytics. You can't really do campaigns without being in tune with analytics. Like you can't really do analytics in a growth marketing team without understanding campaigns. So there's all these synergies between the things. And so that's what I tended to test for when I was hiring onto my teams was, yes, like I'm hiring an analytics role, but how well does this person understand the marketing context? How well can they pick it up? Yes, this person's being hired into a creative role, but can they think about either the strategy or the analytical pieces? Are they extremely organized and systematized? Are they going to be able to like push it in a direction that really at least touches analytics, which is not to say that that analytics person needs to know SQL or that the analytics person or the creative person needs to know SQL or that the analytics person needs to be an expert in the nuances of typography. Yeah.

Yeah.

But yeah, there's, I think like the kind of like where things overlap piece. And I think when coaching people, that's one that I like to challenge people on where it's like, how do you want to position yourself? Like what are the, the things that you've done so far in your career and how cool you draw on those strengths. And people will usually gravitate really strongly towards one thing, but when you read a job description, there are usually like 8 things that they actually want from you. When you distill those things down into main categories, you wanna be able to really clearly point out like, where am I like absolutely strong? And then how have I demonstrated these other skills? And I think being able to point to like 2 big spikes. is what I tend to coach people into, ideally through a narrative. So mine is that analytics plus creative component, and it's like that leads to strategy. I don't think strategy is usually a core skillset that you can really be like, I've always been great at strategy, but I can tell that through a story of how I arrived at my career. So bringing people along for that of what is their personal version of that narrative, I think helps them, you know, get their foot in the door because there's so many people who are so well qualified these days. So being able to differentiate yourself through that positioning is so key these days because I think people are looking for those decision maker jobs and to be a decision maker, you have to really bring something special to the table, I think.

Love it. I love everything you said. I like the idea of 2 spikes because you're trying to really understand, can that person kind of, you know, cut across multiple functions? Because anyway, if you're in a high growth environment, you kind of have to do that. So I like that. Okay, AI, we're gonna talk about AI. So I have 2 or 3 questions for you. One is all the ad platforms are, you know, kind of evolving to, hey, just tell me what you want and I will get it for you. And then we use all the ML to figure out the right targeting and the right creative. Are you a fan of that? Are you not a fan of that? What are you seeing in the market? Like you've been an operator, you've been in seat, Is that the direction that everyone should go in or like look at it with some skepticism?

I think you have to look at it with some skepticism because, uh, the question that I would say is, are you feeding it enough data for it to drive the outcomes that you actually want? And is it possible to feed it enough data to drive the outcomes that you want? So an example that I would give is like for Speak, but like for all subscription apps, right? And so, uh, if you were to use Meta and to use automated placements, You would probably get a lot of Instagram, a lot of young people, because the obvious optimization event for a lot of subscription apps is a trial start. But when you look at trial start rates for young people, they're relatively high, and that's relatively good intent. This is not like a low leverage signal. You're not looking at just an install, but on the other hand, they have a lower payer rate than older audiences, but they also have lower CPMs than a lot of these older audiences. And so there's a lot of nuance to that where if you're not able to pass a value signal that's deep enough in your funnel, you're not going to get the results that you want. And so being able to slice that data afterwards, maybe it's introducing age into your onboarding funnel so that you're able to look at at least like total conversion rates and total payer rates by age. And then you can do some nuance to how you want to target your campaigns. And maybe you do still lean on that automation, but you separate out and have a different willingness to pay for those pieces. So I think it's like kind of similar things like at Cash App, like we had a money movement event overall and that was just a structure, whereas like that's what our eng team was able to do. But we had different willingness to pay for CAC for Cash Card versus other products. And so it's just kind of separating those things out and not just saying like, hey Google, take all of our ads and like do whatever you want. 'Cause that would also probably end up in a funny place where it's like, hey, that headline for Cash Card is in GoWPAD. Oh my gosh. Like add, and you would hope that it would learn that pretty quickly if that's the case. But I think just kind of creating the best set of circumstances, there's a lot of options in that. And I think the specificity will change over time, but those are some like maybe easier basic examples.

Gotcha. That's a good one. What about on the creative side? Are you seeing anything on the creative side where, you know, everyone says this, right? Like, all the creatives who are using AI gonna get better? Like, are you really seeing anything on the creative side where you're like, this blows my mind and it's 10x'd my productivity from a growth marketing perspective? Or it's more like, ah, it's there, it's, you know, we're trialing it out, but it's not working yet.

I would say it's more like trialing it out. I think there are like components that are valuable, like copy iteration, for example. Like, but I think that there's still always going to be a layer of like human intuition that's valuable on top of that to say, What do we wanna prioritize? Uh, like why might this work and how could we maybe make it better? Because otherwise everyone is putting in kind of the same things, uh, and then you're getting out relatively similar things. And I think what actually gets you the best results in long term is differentiation. And so if you don't create that difference, uh, at first maybe if you're the only one who's using AI, then maybe it's like, oh, I increased my output and therefore I'm doing better.

Mm-hmm.

But then if everyone starts using AI and everyone has a lot of ads, but all the ads start to look the same, I think that you'll start to see a swing in the opposite direction where it's like, oh, well, this crappy ad that I made myself that is like, oh, maybe potentially even dumb idea, it just looks so different that people are like, oh, thank God, something different, right?

Yeah.

And so, yeah, I think that just being able to objectively look at the layer on top, I think the tools will empower people to work faster in different ways, whether it's Producing more iterations on top of something that was created by a human or taking something that's created by a human and making it better in some way. But yeah, I mean, I, I, I'm, for example, like not a fan of like fake influencers, like the AI-generated influencers.

Oh my God.

Okay.

Wait, wait, wait. We gotta talk about this. There are people who are saying, hey, we are doing AI-driven UGC. And I'm like, no, it's not UGC. It's AIGC. Right? Like, how can you call it AI-driven UGC? Like, it makes no sense, that whole category. Sorry, rant over.

Yeah, it's super creepy, and I think it's dilutive to the overall thing. I think that the scary thing is people, like, at some point really won't be able to tell the difference, which is like a huge credit to the AI companies in the quality of work that they're doing. But I do think that there is something to, like, the cadence and the way in which someone decides To present something. Like if you're giving an influencer a script and they're reading a script, yeah, maybe an AI can read a script just as well as an influencer can write a script. But if an influencer is coming up with their own idea that would fit in well with their feed, that resonates with their audience, and it's actually genuine, I would like to believe that people will still care about that. And so I think that like, if we stop caring about that, then the whole UGC thing goes away altogether. And then people won't trust it and we'll lose the ability to use that creative type. So, team, use it.

Okay, this is fantastic. I am going to put you through a rapid-fire round and we will bring it to a close. So, I have 3 questions. I don't know how rapid-fire they will be, but I think they're all interesting questions. So, let's say you've hit a creative block.

Right?

You're just like, hey, like we're hitting a plateau. What's your strategy for breaking through a creative block?

Yeah. So I think some of the things that I've seen over time are that like people get like too far down one tunnel. And so like the zoom out. So we would do this thing at Cash App where, uh, like once a quarter we would try to like, this, this sounds incredibly environmentally unfriendly, but like print every single ad and put them on the wall. Um, and like when you look at all the ads in total, like you start to notice some things about them. It's just like, hey, what haven't we tried recently? All those things. I think you can also do that from a data perspective if you set up the infrastructure we talked about earlier and see, are there any meaningful gaps? Where have we really spent a lot of time? And it's like, well, maybe we should spend time somewhere else is the answer to that one. I think the other piece of advice that I tend to give people is a lot of companies do the thing where, in a mean way, you could say that they're ripping their competitors off. In a nice way, you could say, the number one inspiration for most companies is their direct competitors. I think that that puts you in a pretty small box because then it's just like you are behind the curve where all of a sudden, like, not only are you copying the best performing ad from your direct competitor, but like 5 other people are probably copying that best performing ad too. And the trends get more and more same in that echo chamber of your particular industry. And so like for Speak, for example, like they would see this in Korea because they were the number one or are the number one brand in Korea where every language learning app would do whatever the top performing ad was for, for Speak, right? And then that dilutes Speak's, uh, ad, but it also dilutes all those brands' ads thereafter. Like the second one goes first and then the third one goes, you know, thereafter. And so I encourage people to look outside their industry. And so I think some people take this to an extreme, like Cash App would do this thing where it's like, how do we be like Nike? And it's like, okay, maybe Cash App and Nike is like a little bit aggressive of a parallel for a variety of reasons. I think there's a lot to learn from like aspirationally large brands like that. And Cash App to some people probably is the aspirationally large brand, but there's like multiple hoops to jump through there in terms of like type of business, all those kinds of things as well. But I think that being able to take a parallel and say like, for example, should Speak think about other kinds of subscription businesses besides language learning when they're thinking about like how they could change their creative? Or do you even take a step back and say like, What is like happening in creative overall. The last thing I'll say is I like to create group chats and like force sharing for all the people who are involved in creative and like have that as a mechanism for those creative readouts. Everyone's feeds are different and like you may think like, oh, well, we've already tried the thing that I've seen, but maybe you're not communicating exactly what you're seeing. And I think that like seeing different people's feeds will allow people to understand things better. Not everyone's on TikTok, not everyone's like, you know, spending their time scrolling the same things. And so I think it creates a lot of value. We're all different humans. That's why diversity leads to more creativity.

I love those amazing tips. I heard the last one, the same exact thing from another agency, an e-commerce focused agency. They said the same thing. They have a group chat for the entire agency and all they do is just creative inspiration sharing, right? It's like, hey, this is what I'm liking on my feed, or this is what I'm not liking on my feed. Them and so on and so forth, which is awesome. Okay, if you were not in growth, what would you be doing?

So this has been a hot topic lately. Um, this is just a— now I'm just rambling, but, uh, you know, I was traveling for, for Speak so much, and I was spending my time in Korea, in Japan. Um, and I've always been a big food person. Um, and everyone at Speak knew this about me as well, and everyone at Cash App too. And when I left Cash App, everyone was like, Dyan, don't start another job, open a restaurant, right? It's just like, no, I think I'd be a little bit more fine in like tech left for me. And then when I took the job at Speak, like they were just like, gosh, like she's, she's starting another tech job. And then, um, but when I was traveling with Speak, uh, I got inspired like so deeply by some of the bakeries and stuff in Japan and Korea.

Mm-hmm.

And so I have a little concept idea brewing. Um, and so I'm thinking about spending a little bit of my time, uh, maybe like staging, like, which is kind of like apprenticing at, uh, a bakery. Since I now have more flexible working hours as a consultant to kind of get more into the life of that, because I don't underestimate how difficult it is to be a baker.

Oh yeah.

Yeah, I think like something in the food space, I think like it's kind of crazy the way that you think about things sometimes, but spending time in AI, spending time in tech, like things change so rapidly and I don't know exactly where I think everything is going. But one of the things that I do think is going to continue to matter, if not like me being involved in tech and like being at the front of it is, uh, people are still going to care about the things that make them happy. And I enjoy being able to help people or being able to make people happy. And I think like food is one of those ways that like really connects people. And so that's kind of the fun one, the fun take on what I'd be doing otherwise.

I love it. There's one industry that's not going to get disrupted by AI, that's food. I hope, I hope it's not going to get disrupted by AI.

I, I still want to eat real food, so.

Exactly. Same here. Same here. Count me in. So, whenever you start something, you've got your first few customers lined up. Amazing. All right. Last one. What's the weirdest or most unexpected experiment that actually worked?

The weirdest experiment that actually worked. I wouldn't say that this is like particularly weird, but one of the things that I think was like very, went against the grain of what I believed in at the time was when we stood up our media mix model at Cash App, we were like, how do we put everything into the media mix model? And what is going to maybe surprise us about the things that we've already been doing? And as I mentioned at the top, Cash App had been doing a lot from a brand perspective. Some of those things are more or less measurable than others. PR is kind of hard to put in unless you're doing a lot of PR stuff like that. But, and like social, we were putting in impressions. Um, we were doing like tons of money in giveaways. We were getting tons of volume of impressions. We actually weren't seeing that drive, um, as much as we thought it might be. And we're like, that's fine. It's not necessarily like meant to, but, uh, that was one where I was like, I'm curious if it would. But one of the ones that did actually drive a ton of value from an acquisition perspective, and we saw it proven quarter and quarter, uh, again, 'cause at first I was nervous about this. In the first readout, it's like, well, maybe it's small data. I don't know. But we'd also had years of historical data before this was actually podcast. And like, I think some people will be like, this is not that hot of a take, but the way that Cash App did podcasts was like fully sponsored. Like we work with Pod Save America. We bought them their first mics. They talk about us in this really integrated way. So there was no native measurement for this. And like we would put codes in the episodes, but the codes weren't always used that strongly. Because, you know, a podcast is like, maybe you're listening in your car, like there's like all these factors to it. But we had kind of heard this before from users where they'd be like, oh yeah, I heard about Cash App from like X podcast or like something like that.

That's cool.

And so it was really cool to see that come to fruition, but the scale was just much more than I thought it would be, you know? 'Cause it's like you always hear these like bright spots where someone's like, oh, I saw your TV commercial, or I like heard this podcast thing, but it's like the relative scale of it is often hard to compare to performance marketing or referrals or something like that. but it was really valuable. And the experiments that went after that were like, how do we categorize different kinds of what we started to call episodic? So we would work with like RuPaul's Drag Race and like, that's an episodic, but it's on TV.

Mm-hmm.

And then we have these like podcasts that are either on YouTube or they're like audio podcasts. They're ones of different sizes, of different genres. And we started doing experimentation and we continually saw that investing for a reasonable span of time repetitively with these episodic was driving not only brand value, which we were measuring through a separate version of our marketing mix model and feeding into the menu mix model, but also acquisition results as well. And the more that we experimented, sometimes the more impactful we were able to be with that. And so that was something that was really exciting. So I think sometimes brand teams and growth teams can be a little at odds in terms of like, hey, I don't think your thing works or whatever, which was never the intention. But I was like, I didn't think of it as an acquisition channel. So the opportunity To be able to come to the total marketing team and say like, I think we should invest in this more. Like, don't give me more money, give them more money, but I would love to be a part of how we experiment with this and how we can engage with this, uh, was really cool and allowed me to work with some people that I wasn't normally working with on a day-to-day basis on how we could collaborate, how they could understand more about analytics, um, how I could understand more about the, the brand space as well. So that was one that I really enjoyed and learned a lot from.

That's amazing. All right, there you go. Podcast ads. I'm a big fan of podcast ads. I listen to a lot of podcasts. I've like a bunch downloaded all all the time. And I've been diversifying my podcast listening habits and seeing like Oracle advertising on The Economist, you know, and you've got like all, and I'm like, whoa, okay. So somebody at Oracle is like going, oh yeah, our audience probably listens to The Economist and we should advertise and there'll probably be business there. So it's pretty powerful and I think it's still pretty underinvested in as a channel. Yeah.

I think the hard thing is like in doing it well. Like Speak also saw podcasts work really well, um, but I think that you have to do it natively. When you just do like programmatic, people skip, or like, I guess it depends on what your demographic is. But, uh, if you're like a high-value product, you're probably a paying listener and you don't even have ads. Like, I don't get ads on my Spotify, um, and so it has to be the native post-read and it has to be one that I'm actually going to listen to. But then you also have the potential to tell a story, right, where it's like Because I, I listen to like all these in a row. It's like, first we're gonna introduce you to the brand, then we're going to introduce you to something that'll make you consider, then we're gonna tell you about these product features. And if you can make that a really native integration that feels like something that the host wants to talk about, I think it really makes a difference. So it's like that creative—

Delivery.

Yeah. Which like then people won't skip it. And so it's like, there's a, there's a big difference in performance. It's like very different than your typical ad.

I completely agree. All right, Dyan, this has been fantastic. You have jam-packed a lot of insight for everybody. How do people work with you? Let's say somebody's looking for a growth consultant, like where do they, how do they reach you? They just reach out on LinkedIn? You have a website?

Yeah.

Where do we send them?

Yeah, all of the above. So you can definitely find me on LinkedIn, have a pretty easy name, which I realized I had a capital in there. That's actually crap. But my website is also My name, so dyan.khor.com. You can reach out on there, happy to grab time. Yeah, really excited to also take on clients and meet people who are doing interesting things. There's also some forms on there for if people are looking for career advice, but also if you're not ready for consultancy and just have questions, there is an AMA in there and I'm trying to get better about answering things, writing things, things of that nature. So. Happy to chat with folks, even if it's not the right time yet. Always like to jam with people. So yeah.

My AMA is going to be when you launch your first baked good or whatever you end up launching. That's what it's going to be. So I'm excited about that. Dyan, thank you so much for joining and sharing all these insights and looking forward to catching up again in a few months and we'll talk about a different topic. Whatever's resonating at that point. So thanks again.

Awesome. Thanks so much, Pranav. Great to see you.

All right, folks, that was Dyan Khor, the, a former head of growth marketing at the Cash App and at Speak, and she's taking on new startup clients right now. So if you need help with your startup's marketing efforts, go reach out to her on LinkedIn and tune in again next week for another amazing guest we'll be hosting on The Marketing Scientist. See you soon.