Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Strategic Decision Making, The Role of the Data Scientist, and Building Uber | Sundar Swaminathan
30 Plays1 month ago
SUBSCRIBE TO LEARN FROM PAID MARKETING EXPERTS 🔔
The Efficient Spend Podcast helps start-ups turn media spend into revenue. Learn how the world's top marketers manage their media mix to drive growth!
In this episode of The Efficient Spend Podcast, Sundar Swaminathan, Head of Data at Bounce and former Uber data science leader, explores the intersection of data analytics, AI, and growth marketing. Sundar shares insights from managing $1B in ad spend at Uber, the role of brand awareness in scaling businesses, and the evolving skill set for modern analysts. He also discusses decision-making frameworks, the future of AI-driven analytics, and strategies for optimizing media budgets effectively.
About the Host: Paul is a paid marketing leader with over a decade of experience optimizing marketing spend at venture-backed startups. He's driven $250M + in revenue through paid media and is passionate about helping startups deploy marketing dollars to drive growth.
About the Guest: Sundar Swaminathan is a data and marketing leader with over 15 years of experience in B2C growth, analytics, and data science. As the Head of Data at Bounce, a commoditized business in luggage storage, and former Uber data science leader, he has analyzed $1B in marketing spend, built high-impact analytics teams, and now helps startups scale through data-driven decision-making.
VISIT OUR WEBSITE: https://www.efficientspend.com/
CONNECT WITH PAUL: https://www.linkedin.com/in/paulkovalski/
CONNECT WITH SUNDAR: https://www.linkedin.com/in/sswamina3/
EPISODE LINKS:
https://bounce.com/blog
https://experimental.substack.com/
https://www.marketingevolution.com/marketing-essentials/data-driven-marketing
Recommended
Transcript
AI and the Future of Data Analysis
00:00:00
Speaker
My gut feeling is that AI makes the best data analyst even better. Like I think for the first time in in the history, you can have a 10x analyst.
00:00:12
Speaker
Just you could have a 10x engineer, you can now have a ten x analyst because they can automate away all of that work that used to be sort of monotonous, repetitive, etc.
00:00:24
Speaker
What that also does though, is you can outsource the, again, the 90% of analysts that have been sold that just knowing SQL without communication skills is enough. You've now got a tool that can replace literally every single one of your analysts. So have 10 of those analysts that are contractors.
00:00:42
Speaker
Every single one of those the push of a button, doesn't get tired, and is 90% of the cost. But what it can do is sequence the right context of questions at the right time in a way that like actually pushes the needle and makes marketing or product and growth move along.
Sundar's Career Journey
00:01:09
Speaker
Sundar, welcome to the show. Thank you for having me. I'm excited to chat with you today about all things data analytics, AI, growth marketing, Uber, maybe even some Department of the Treasury stuff, which will be fun.
00:01:23
Speaker
I think just to kind of kick things off, would love if you could give a brief but background into into your experience, who you are, what you do. Yeah. you did Yeah. Well, again, thank you for having me. Excited to be here.
00:01:35
Speaker
And um you kind of already touched on some of my background, but and I've had an interesting career where i I started my career as a software developer. Didn't love it. and went into finance, which is where I went to the US treasury and loved that, but just didn't really love but you know what my manager's role was. So i was like, okay, this is not for me.
00:01:55
Speaker
And then finally got into marketing data science and I truly found what I'm passionate about and I've stayed there for over a decade. So most of my career was at Uber where I spent five years across the entire funnel. So when I left, I actually built out their global brand data science team.
Building Teams and Leadership at Uber and Bounce
00:02:13
Speaker
So really, really top of funnel.
00:02:15
Speaker
Before that, I was in performance marketing, before that in lifecycle marketing. So really obscene, the full stack. And then went into a few different product data science roles, product growth data science roles, if want call them.
00:02:27
Speaker
I was consulting for two years. And then as of yeah Monday, I'm now the head of data at Bounce, which is a ah series B consumer marketplace scale up, which is really exciting just to bring back a lot of the stuff I loved at Uber. I'm now doing something very similar, which is awesome.
00:02:46
Speaker
That's awesome. Yeah, and it's you have a ah really unique experience kind of running the gamut from different stages of of companies.
Impact of Career Transitions on Work Strategies
00:02:56
Speaker
i I love chatting with folks that got into marketing and and growth and startups after kind of testing, experimenting with different companies.
00:03:04
Speaker
careers because it also informs how you approach your work today and how you approach your work as a as a data analyst marketer even myself i started my career in sales and then kind of pivoted into marketing and growth marketing and a lot of those elements i still kind of use on a day-to-day in my role and also just in and real world and in life If you kind of think about, you know, some of those foundational elements of starting as a software engineer at Bloomberg and working on at the Department of the Treasury, is there anything that kind of ah like sticks out to you that like informs how you've kind of grown in your career in the startup world?
00:03:44
Speaker
Yeah, that's that a really good question. i they I think just the fact that I was able to bounce between those roles. So ah two two questions to answer, I think.
00:03:55
Speaker
One, i think just learning to be a software developer was just learning to be very independent. Something I was really surprised by is like how much they software developers just repeat.
00:04:06
Speaker
Like you're not like they're basically hit the rule was like, if you're rewriting or if you're building new code, like you're doing it wrong. Like you should just be basically rewriting or borrowing from someone else and just packaging it for whatever your app needs.
00:04:19
Speaker
And so I became really good at Googling. But as a software developer, and it's something like you don't think about, but like, it's really important to be able to just be self-sufficient. Like I can, if I get stuck on a problem, and now AI makes it a lot easier, but like it used to be like, you would have to learn how to Google and go through Stack Overflow. And that's also true as a data scientist.
00:04:40
Speaker
So I think learning how to Google is something I've become really good at. And it's still something like that makes me easily like if I'm curious about something, I just will go Google away at it.
00:04:50
Speaker
So it's a weird skill set to say, but it is something I think I'm quite good at.
Lessons from the U.S. Treasury and Decision-Making
00:04:55
Speaker
As a at the U.S. Treasury, i think a lot of it on like just understanding how important your decisions can be and learning to make really important decisions at a very early age. So i was at the U.S. Department of Treasury. And so we were managing 19 trillion dollars of U.S. debt.
00:05:16
Speaker
We had to predict when we thought the government would run out of money during government showdowns and sorry government shutdowns. And so you learn very quickly, like it's not about the complexity of the work you're doing. It's just like yeah says really making sure you've thought through it, like making complex decisions, like having the validation behind it.
00:05:37
Speaker
And I took those into and to Uber. and And so being independent was really useful at Uber because you had a lot of autonomy, but then also making like pretty big decisions at a relatively early stage career was also something that was true at Uber. So yeah, I think those are the two biggest ones I would say.
00:05:55
Speaker
That's something I hear over and over again when, you know, listening to interviews from executives, founders, CEOs, you know, they even talk about Jeff Bezos talking about this, like the the magnitude of decisions and how you don't need to make a lot of decisions. You have to make a few really, really good ones.
00:06:13
Speaker
As you think about maybe decision-making, your decision-making framework, do you think that's kind of like a ah muscle that you've that you've grown over time to be able to differentiate between what's a micro decision that I need to make quickly versus what's a long-term decision that I have to go on a walk and I have to journal and reflect and kind of think about?
00:06:37
Speaker
Like, what advice would you have for marketers, analysts, as they think about that, making decisions?
Decision-Making: Micro vs. Macro
00:06:46
Speaker
Yeah. Oh, I love it. This is really like, these are some hard hitting questions right right off the bat, which I love. But I think the only way you learn what is a micro versus a macro decision is a bit of experience. I mean, there's some of them, of course, that you have a gut feel for.
00:07:03
Speaker
But if you look at it from business context or a marketing context, if you're afraid to make the micro decisions, it becomes nearly impossible to make the big ones. But if you approach the micro ones as macro decisions are early on in your career, because let's be honest, when you first join, you're not making macro decisions. like You're not being put into situations where you have to make macro.
00:07:23
Speaker
But treat the micro decisions like macro ones, right? like Just because it's a micro decision, it's micro in that the impact might be small and therefore you don't have to think about it a lot. But you still should think about it in an intelligent way. And so treating the micro as like, okay, what if my career depended on this? Just as a hypothetical.
00:07:43
Speaker
Early on, i would overthink it, but that made me really good because over time, those micro decisions, i would think quicker because I've already spent a lot of this time overthinking it.
00:07:56
Speaker
and you like At first, you're going to overthink. like You shouldn't try to avoid it. It's actually probably good for you to overthink. And then you learn, well, I didn't have to overthink this. And the next time I was like why definitely enough to overthink that. And oversaw all of a sudden, by process of elimination, you think exactly the amount you're supposed to for micro decision.
00:08:15
Speaker
But as your decisions get bigger, that framework of I know this isn't overthinking, I know this is, you start to calibrate. And so... I would say in the beginning, treat micro decisions like macro, build up that muscle.
00:08:29
Speaker
Because unfortunately, if you F up a micro decision, depending on the place that you're at culturally, it could not go well for you. like You also have to be in a place that supports you. So that's the other thing too, that you learn over time is it's less about the decisions you making and where the place that you're at, that you're making them, that actually will have the bigger impact.
00:08:49
Speaker
But when you don't know that in the beginning of your career just try to think a little overthink and then just start to calibrate over time and it's the only thing i've i've personally worked has worked for me right you know i i think about that as i think about a lot of things as existing in a spectrum and maybe decisions existing on a spectrum of like You know, maybe easy, simple, small to large, complex, longer term impact, whatever.
00:09:22
Speaker
And then within any given role, there's going to be that spectrum. So an entry level job, you know, the small decision might be like going, like doing a training or I don't know, you know,
00:09:35
Speaker
ah Attending a meeting, like very like micro micro kind of things. And then the the larger decision for that entry level role, like you said, would be something seemingly simple to ah more, you know, someone in their mid career or something like like that.
00:09:54
Speaker
So now as you exist in you you kind of like grow your role over time, i guess the idea is that you want to make sure you have a balance between like spending time in those small decisions and automating that and making that easy, but also making sure you're allocating time towards thinking through some of those more complex kind
Role of Managers in Career Growth
00:10:16
Speaker
of things. And like, how do you know what the appropriate balance balance Let's just go with the examples you provided because i actually really like them. right So if I were to think about like um like a training decision for a micro as a micro thing, at the earlier set of my career, I was overthinking, what is the audience going to ask? Have presented it right? like Let me go. And I would spend hours rehearsing 15-minute training.
00:10:37
Speaker
And then over time, I'm not rehearsing for hours. I'm rehearsing for a minute, but I'm still practicing. As I've gone into my career, I've really realized like your manager is probably the most important person in your career.
00:10:52
Speaker
Like the whole classic saying, if you don't leave jobs, you leave managers has been very true for me. Like I've stayed because of managers I've left them. And so Part of wisdom, I think, is also realizing you're not in it alone.
00:11:05
Speaker
And so when you go through and you become more mid-career, I would lean on my manager even more because the decisions become bigger. so I'm like, hey, you know I can either do these 10 trainings, for example, or I could do this one- meeting, but it has like a senior stakeholder in it.
00:11:20
Speaker
Which one should I do?" And it's like, well, you go to the senior stakeholder one because you need that visibility. Why do I need the visibility? It's not going to help me today. But yeah, but come come this time where you need buy-in or you need promotion or you need something.
00:11:34
Speaker
you know you know My manager is telling me like that visibility is important. That's not something you know yourself. So Part of it is learning from people that have already done it on what are those micro decisions versus macro.
00:11:49
Speaker
Having the faith that, listen, I've already made these type of micro decisions 100 times. I don't need to spend as much time. i need to get better at the stuff that I'm more insumptuable with. And there's this like natural pull, I think, of if something is uncomfortable for you, at least for me, I know it's bigger than I'm prepared for.
00:12:08
Speaker
And so i like collaborate i mean I corroborate that with someone who has done it. And I'm like, hey, is this really as big of a decision as I'm thinking? Yes, no. And then you you sort of steer the appropriate time towards it.
00:12:21
Speaker
So... Yeah, a lot of it is just, and this is why I think like even marketing data science is such a human function. Like it's so much of it is like gut feeling. It's talking to your manager. It's like, hey, predicting like what I think people are going to react.
00:12:37
Speaker
And then leveraging the experience of stuff already done and be like, okay, that wasn't as big a deal. I saw it play out the way I thought it would. This makes me feel uncomfortable. Let me go towards that. That's kind of how I thought about it.
00:12:48
Speaker
I don't know if that's satisfactory because this is the first time I've had to think through how I think about these decisions, but it's really it's a really good question. So yeah, let me know if that's helpful. ah Yeah,
Communication Skills in Data Analytics
00:12:58
Speaker
yeah, for sure. and ah and I mean, I'm also i'm also thinking about, you know I wanted to get into Uber, but we're we're jumping. I want to stay on this for for a second.
00:13:07
Speaker
There are, there's also, there's, there's decisions, big decisions, small decisions, you have to make it in your role. And then there's also the skillset that you have, and you tend to lean towards making the tough decisions using this.
00:13:22
Speaker
You want to leverage the skillset that you have, right? When it comes to analysts, one of the things that that you kind of mentioned when we chatted last time is that 90% of them are pretty shit, right? And I think what we were talking about is that certain elements of analyst role have kind of been commoditized to a certain extent.
00:13:42
Speaker
And so it becomes more about some of these soft skills of communication. i was reading an article that that you wrote on Experimental about how to have a very simple dashboard and how we can over-engineer dashboards and create too much work. And I see this as something that's like very prevalent specifically in the kind of data analyst function of really smart folks that maybe don't, aren't, it's not, the the human element doesn't come as easy. So I wonder how you think about that, right? Like that kind of the skillset that somebody technical should be thinking about building.
00:14:22
Speaker
Yeah. So an important, cla not even clarification on the 90% are shit, is it's not because the analysts, right? Like that's something I want to clarify is like 90% of analysts are shit because they are trained to be shit and or that is the expectation that they've been sold.
00:14:44
Speaker
But it's not like a, I don't think there's like a unique talent that you need to have to be an all-star analyst, right? This is not like an athletic thing where like, kind of like you're born with the ability to be six feet five to go play NBA.
00:15:00
Speaker
it's It's not like that in a professional career. And so what's important for me is how do you actually train analysts to understand what is important in the role And then how do you fit their skillset into that? That's how you make a good analyst, right? Is here are the expectations for high quality analysts.
00:15:22
Speaker
And here's how we can morph your innate skillset to do that. So just as an example, I've never learned to Python.
00:15:33
Speaker
I think if you were to tell any data, if I were to just like throw my hat in the ring, like, I don't know, like I think 99% of data analysts will laugh at me. But I feel really strong in my career. Like I am an above average analyst because I've got this ability to admit when I'm wrong and just, this is gonna be a weird thing to say, and just like smile.
00:15:53
Speaker
Like I'm pretty, like I'm often like the cheeriest person in the room. And that's my advantage because like I'll admit when I've made a mistake, but then i'm like, hey, let's like work together. like how do we like like how do i and And I'll go and be like, I am sorry. like All these little things that we don't really do a good job of coaching analysts to be like, hey, at the end of the day, it's just a relationship. like ah stakeholder is making a decision and evaluating you.
00:16:20
Speaker
of course, on the data. but they're always willing to forgive that if there's ah relationship behind it, right? I'm not a perfect analyst. And so for me, like the thing that I would coach analysts to like lot on is remember, it's like a human relationship first.
00:16:39
Speaker
And then like, like you don't have to be extroverted, introverted. It's not about that. It's just about, again, it's, it's a human thing. So Like, that's how I like recommend analysts sort of go beyond what they're just reading and be like, I got another certificate on SQL. It's like, well, cool.
00:16:58
Speaker
But like, when's the last time you've actually sat down and and even just have lunch with your stakeholder? Or, you know, like, go do that. And I bet you, you get much further in your career than getting that next certificate.
00:17:11
Speaker
And a lot of it is because I had mentors and coaches tell me that. They're like, hey, you're you know you're not good at saying no you're not good at you know making sure that the stakeholder has like the right information, like all of this stuff. like I was so buried behind like how I would look at the and analysis versus positioning for how the stakeholder wanted it. And so, yeah, that whole framing is is something I recommend.
00:17:37
Speaker
One of the things that I've noticed and I see kind of maybe maybe changing is Within lot of analyst teams, you'll have a lot of subcontractors or functions that are outsourced to different countries. i mean, a lot of time, India, right? different Different places where there is maybe time zone differences, works being done off hours and then presented the future, you know being presented without really being communicated as much.
00:18:09
Speaker
Maybe there is communication, cultural disconnect sometimes, and it's a challenge, right? Because some of these soft soft skills, you know, it's, it's, you can't take a course on it.
00:18:20
Speaker
Right. And i wonder, like, you know You see the ability to take these boot camps. A lot of this education has become commoditized.
00:18:33
Speaker
It's being hired out internationally. And then you have AI on this other side of things, which might automate a lot of it too. So I'm curious, like when you talk about the future of of analytics and we talk about how these soft skills are super important, does that even become more important in a world that I kind of just structured, right?
00:19:01
Speaker
AI happening and just the challenge of maybe outsourcing some of this stuff to save a few dollars too, because you have to pay someone more in America.
AI's Role in Enhancing Analyst Efficiency
00:19:10
Speaker
ah Yeah, this is this is a topic that comes up often. And so what I'll say is my gut feeling is that AI makes the best data analyst even better.
00:19:23
Speaker
Like I think for the first time in in the history, you can have a 10X analyst. Just you could have a 10X engineer. You can now have a 10X analyst because they can automate a way all of that work that used to be sort of monotonous, repetitive, etc.
00:19:41
Speaker
What that also does though, is you can outsource the, again, the 90% of analysts that have been sold that just knowing SQL without communication skills is enough.
00:19:54
Speaker
Because you've got some, you've got a, you've now got a tool that can replace literally every single one of your analysts. So you have 10 of those analysts that are contractors. Every single one of those is the push of a button, doesn't get tired.
00:20:06
Speaker
and is 90% of the cost. But what it can do is sequence the right context of questions at the right time in a way that like actually pushes the needle and makes marketing or product and growth move along. like I have yet to see like an agentic AI that's like, okay, cool. like Hey, have you thought about answering this churn question first and then this and then this and then?
00:20:31
Speaker
And so that's sequencing and sort of human thinking is not there yet and i'm not sure if you can ever completely encapsulate that because it's it's uh yeah like again people need like ai is just is essentially taking a bunch of information and bringing it to the average so that it's you know you're you're you've got the average information at your fingertips which is awesome That means it leaves a lot of room for above average.
00:20:58
Speaker
But anyway, my point there is like, it will change and I hope it does change the analytics and data science field in in that it removes the the need for like, you know, just being able to have SQL.
00:21:12
Speaker
But if you are really good and you have really strong communication skills, I think you become even better. Like I imagine if I had an analysis now, it would take me 20% of the time it used to take me, which now means i can churn out five times more analyses.
00:21:30
Speaker
Each one of those then makes me more impactful because not only am I doing the same volume of analyses, but I'm doing it shorter time. But you have to have been a good enough analyst to understand to ask the right questions in the first place. And so even a lot of the conversations I hear around, okay, we're going to give everybody access to like, you know, a prompt that'll write SQL for you.
00:21:50
Speaker
Well, cool. I mean, people have not been able to ask analysts the right questions for two decades. They don't think you're all of a sudden going to be able to just because you have this AI tool, right? It's always, were you asking the right questions? Yeah.
00:22:04
Speaker
even understanding stuff as simple as like CAC, like people still have like a thousand different definitions of the CAC and like not everyone's computes the right way. Like I could go down to so many different examples of it's never, was never the the the answers that were wrong. It was always the questions that were wrong. And I don't think until people learn how to ask better questions that it,
00:22:24
Speaker
It completely eliminates them, but it does make, it makes people, if you're a bad analyst, you should be scared. If you're a really good analyst, you should be happy because like you can, you can become a 10X analyst or data scientist. I, by the way, I use those terms relatively interchangeably here.
00:22:39
Speaker
Right. Yeah. if If AI is very good at answering questions, you need to become really good at asking them and then also kind of understanding what is a realistic answer and and what to expect and and learn how to to massage that.
00:22:57
Speaker
And the skill set of learning how to ask the right questions is something that only comes with experience of working with an organization. seeing the growth challenges, being able to identify, like, here are the areas to type prioritize.
00:23:12
Speaker
um I think that's a good good transition to to
Strategic Marketing at Uber
00:23:16
Speaker
Uber. the The headline at at Uber is that you built the brand data science function and measured $1 billion in in spend.
00:23:25
Speaker
that's That's a lot of spend and a lot of experience in understanding What are the elements that led to the growth of that organization? What I want to get your thoughts on to to start is thinking through how to how what your perspective is on the ways deploy media budget to grow and organization do you have a framework do you have a philosophy on that after seeing so much spend go through uber
00:24:02
Speaker
So what's cool is that I was able to use data science to verify that like marketing is not changed in ah hundred years.
00:24:14
Speaker
And the principles of that have not changed and I don't think will ever change. And what I mean by that is at the end of the day, you need to communicate in the most effective manner, how your solution solves a customer's problem.
00:24:31
Speaker
which means in a very simple sense, you have to understand what their problem is, and then you have to build a solution for that. The best example I have is not even from brand, but I'll talk a little bit about performance marketing here.
00:24:44
Speaker
The best performing ad for Uber for a very long time, and I'm not sure if it's still true, was push a button, get a ride. We could not beat that.
00:24:54
Speaker
Like just could not, no matter what you did. And at the end of the day, it is so simple. It is so fundamentally simple. And I'm probably sure, I'm pretty sure if you research a bunch of like, talk to customers, they're like, why do you love Uber? It's like, man, like I pushed a button and a ride showed up.
00:25:10
Speaker
There you go. There's your language. Now you don't even have to write the copyright. So what I what i was able to see is Yeah, it's just getting that solution. And so you could not like, you know, at the beginning of Uber, the best performing market channel was referrals. Okay, well, it makes sense, because you've essentially productized word of mouth.
00:25:30
Speaker
But like, was it incremental? Nobody ever asked, but it was still there to prove that word of mouth was going. Then eventually we like, we're on paid social. Well,
00:25:43
Speaker
Eventually, we were so saturated that we didn't need paid social for rider acquisition in the U.S. So we turned it off. Okay, well, then what became the issue later? The issue later was that there was in every country, there's many competitors.
00:25:56
Speaker
We're losing to DoorDash on Uber Eats in the U.S. So what do you see Uber do in the last four years? You see a lot of push towards brand awareness, which is where the billion dollars of spend came in. All of those principles were true from the same. And in fact, if Uber had been brand building from the beginning, there would be zero room for any of the competition.
00:26:15
Speaker
Like you could not have gotten Lyft to succeed, right? If people had loved Uber from the beginning, Lyft couldn't come in as a counterposition to be like, we're better for drivers. And at the end of the day, now Lyft and Uber do the exact same thing. So drivers are arguably better or worse off in the same way. But what I learned was like, you know, nothing has really changed.
00:26:35
Speaker
We just have gotten better at the science to back it up and prove it. But even now, like Bounce is this is ah is a I will say a commoditized business in that it's luggage storage.
00:26:47
Speaker
And so the same principles are there, right? Like you you can only really win on trust. And maybe you have more luggage storage available to you than in competition. And so it becomes another brand play. And so, yeah, that's that's really what I learned and that the metrics and frameworks are all about that. Like if you don't, if you want to measure the impact of brand awareness on business metrics, well, first you have to prove that your campaign actually moves brand awareness.
00:27:14
Speaker
pretty simple logical like thought pattern then you do a brand live study okay great we're moving awareness well then you set up a six month geo holdout and you prove that it's moving the business metrics but How do you know that your campaign is moving brand awareness?
00:27:31
Speaker
Brand lift study. But then the briefs, the campaign, like they were all the same. Like, you know, you still had to write a campaign brief. You still had to draw from an insight. You know, you still had to think through through what was a problem that I was solving for a customer. And so, you know, for me, it's, it was really cool to just validate a lot of the literature and research around marketing is still the same.
00:27:54
Speaker
One of the things that that i I love about what we do and I i love about my experience scaling media mixes for for different brands over the years is that it it is very deterministic in a way because there there really is kind of a clear pathway to getting to scale.
00:28:14
Speaker
And the elements of that are you need to have a good product, a great product. You have to go after demand audiences and then move up to mid demand and low demand and and build that that brand.
00:28:33
Speaker
And there's really a clear pathway, but it's hard because it wouldn't exist if you didn't have a really good product to to start with. I want to hit on the the the Facebook us spend in a second.
00:28:48
Speaker
But before I do that, I think one of the things that a lot of folks listening to this might struggle with right now is that we're in the era of efficiency and trading dollars from really good CPA, CAC-looking things into video view, ad recall things is hard to sell to a CEO and a CFO.
00:29:17
Speaker
And so how do you do that?
00:29:23
Speaker
um I don't think they're going to love this answer. i mean, you essentially don't. if that's the If that's the argument that you're having, I'm not saying don't try, but you've lost because i don't know how many more examples of this we need to share.
00:29:38
Speaker
Right? Like, it's like, okay, pick, like, you almost have to get confrontational with your CEO and CFO about this. Like, okay, well, what shoes are you wearing? I'm wearing Allbirds or Nikes.
00:29:50
Speaker
Why? Okay. And like, you know, yeah how many, like you have to go through this, like, so my point there is like, I know it's like, it might feel like a cop-out answer, but like, it's such a, it's such like annoying question, not from you, but it's such an annoying question in general in that.
00:30:05
Speaker
I don't know. I just get sort of like fed up of having to be like, why do we have to keep proving this? Like, i would almost say the then yeah go to somewhere place that actually gets it because that it's like such a taxing endeavor to do.
00:30:21
Speaker
That by the time you get that buy-in, you're almost like so uninspired to then go build the marketing strategy because it's going to be as soon as you launch that brand awareness. Well, okay, well, when are we going to see awareness go up?
00:30:33
Speaker
It's like six months, three months, you know, like, okay, fine. Well, okay, three months later. Okay, well, fine. Well, has it gone up? Yeah, it's gone up like 6%. Well, is that a lot? Is that a little? It's like, I don't know, like I can't like that was I'm not able to map out and predict.
00:30:48
Speaker
And so you like the foot is always on like this isn't working. Let's stop it. And that like attitude is so hard to beat. That, yeah, the only way i think you can do it is like you go here's how much budget I need.
00:31:06
Speaker
Here's the metrics that I think will move. And then we don't revisit this for six months. like unless like The goal for you should be to get like a six-month breather. If you get anything less than that, then you're it's just it's not worth pursuing.
00:31:20
Speaker
right So getting that buy-in from your CFO CEO, if you want a heuristic to think about is... Get a six month runway and just say here are 20 case studies.
00:31:30
Speaker
Let me know if you want more. Sign case studies within your own business, within your own industry. And yeah, and that's that's it. And be like, oh, by the way, we have to make sure our product doesn't suck.
00:31:43
Speaker
We have to make sure our brand doesn't change. We have to make sure yeah there's not an existential crisis, et cetera. And then cool. like I feel pretty confident that over six months, I'll show you that brand awareness lifts search volumes, which means we can do...
00:31:58
Speaker
Higher volume of throughput at the same CAC. So like here's here's all the evidence, here's all the case studies, but like, yes, so that's
Balancing Performance Metrics and Brand Building
00:32:06
Speaker
kind of my answer. i like i get this asked all the time and I just, I feel bad.
00:32:10
Speaker
i Like as soon as someone asks, like, how do I convince? I'm like, I don't know, dude, it's like kind of like as a data scientist, I'm like, how do I convince my CEO that like being data driven actually proves anything? Like, I don't know, just look at the top 30 companies in the world and they all invest a ton in data scientists, data analysts. Is that not enough evidence? Like, so yeah, that's kind of hard thing to push for.
00:32:35
Speaker
to protect to to push for To me, I feel like it's you're operating within certain financial constraints in the aggregate. You have ah CAC target that you're trying to hit or a ROAS target or ah revenue target, whatever.
00:32:52
Speaker
And if you're hitting that and you can allocate more room in the budget to doing things that don't have that CAC target because the other things you're doing are getting you there.
00:33:03
Speaker
I think it's a challenge when, that I've seen sometimes where you have a CAC target and then every channel is beholden to that same target. Or they we try to think in that manner, right?
00:33:15
Speaker
Well, we need to get a three ROAS and so everything needs to be around three. And if this thing's at one, we cut it kind of thing. Yeah. So I think there, right, what you, the argument becomes like just understanding the three levers that you have to pull within marketing, which is i can, I forget the three now exactly, but like you can keep the same CAC and then therefore keep the same volume. and Like I can't give you same CAC and higher volume unless I do something more up final.
00:33:49
Speaker
um or I can do you know conversion rate optimization. But the point being is like, there's three things that you can pull, but you only get two things. You can only get like good CACs in decent volumes, but but then the third thing like has to stay.
00:34:02
Speaker
And so the point there is like, listen, if you're happy with our CAC, and you're happy with our volume, and you're happy with our growth, great. Fine. Let's not have this battle right now. But just to know, like we estimate that we're going to tap out of our demand in six months.
00:34:16
Speaker
So I'm going to put on paper now that in six months, your caps are just going to start to spike, but we're going to keep the same volume. So like that's the hard part is like you're not prepping for now, you're prepping for six months.
00:34:28
Speaker
And you're also not like, Like it is about efficiency, but it's got to be efficiency in terms of like, like in the context of the business, right? Like if you're a high AOV, high order value business, maybe you can get away with not doing brand spend because you just need the one or two contracts and you're good.
00:34:50
Speaker
But if you're like, if you're in this like small AOV business, like you just need that to be able to sustain higher volumes and then your entire business model designed off volume.
00:35:04
Speaker
And so again, like, you you just have to find a case study, but yeah, again, not an easy conversation.
00:35:11
Speaker
I think about this world in terms of putting dollars across the demand spectrum and in any given product or service, there's going to be a large number of people that have low demand for it, a smaller amount of people that have kind of middle of the road demand and a very small amount of people that have very high demand for this product.
00:35:33
Speaker
um And so you spend dollars on the high demand, smaller area, and you capture that market. And then eventually you get saturated, and you move up and you move up. And then you're in this place where you're kind of spending across all of these different demand stages.
00:35:48
Speaker
Obviously you focus a lot on the lowest demand stage, people, audiences that either didn't know the brand existed, didn't give a shit, maybe had the problem, but weren't aware that this was a real problem for them.
00:36:02
Speaker
At Uber specifically, and maybe some other places that you've been at, what have you seen from a messaging perspective work really well and in brand or what type of experiments have you run to see, oh, this actually did really well?
00:36:16
Speaker
Yeah. So part of demand is that a lot of people don't realize inherently that the demand that they have. Just for example, again, like most people, when they think of Uber early on, thought of it as like, I need it for the weekends.
00:36:35
Speaker
But then it's like, well, okay, do I really want to go park my car at the airport if I go away on a trip for two weeks like and pay $200 to leave it at the garage? Oh, wait, or I take like a $30 Uber.
00:36:47
Speaker
So all of a sudden, we would start to create demand across use cases. Right. So there's, you know, a lot of it is understanding our customers. Right. Which is the primary use case early on was weekend nights.
00:37:00
Speaker
So who are those type of people? They're either they're not college students because we were a little bit expensive at the beginning. They're probably young professionals, really, you know, solid money. Okay, well, what other the uses schedule do they have? they They could take it to and from work. They could take it during the week from like their social outing. And so we would message that. Like, hey, don't forget, like after like yeah three drinks at a bar, Uber is a great ride home.
00:37:24
Speaker
During the week, though, right? During like happy hours. So... That's that's like a very early example, but like over time, it's it's just introducing more and more use cases because that's where that's where the demand is. The demand is, like yes, it might be low demand, but it's low demand, high volume.
00:37:45
Speaker
And if you look at in terms of aggregate demand by just multiplying demand times volume, it's always at this low demand portion that you have just like massive, massive demand.
00:37:56
Speaker
And every company solves the problem of how do we go vertically? Like, how do we go more upmarket? How do we get them to more use cases? right For Airbnb, it might have been a great example for the once once every two week trip with your family, but now like they use Airbnb for work.
00:38:12
Speaker
So now I'm using Airbnb for work on the daily use case, which makes me think about it for the long term. so So a lot of the messaging we would do is just how do we introduce new use cases?
00:38:22
Speaker
How do we you know make you realize that the solution that worked for this... is the same is the same form, it's the same function of a solution, but mentally we paint it in a different light.
00:38:36
Speaker
That's also why you see more products like Comfort versus UberX. It's the same supply, it's the same drivers, just better rides. And so you might be willing to splurge like when I'm like, this is either paid by my my my employer or listen, I'm on a trip, I'm going to do Comfort versus X, right? It's just taking that same form, function and messaging it.
00:38:59
Speaker
That's how you would go further and further up demand and you would create demand. That's the other thing too. you So many people don't realize they they need Uber, not need, but they want Uber Eats or they want Uber X. And so it's our job as marketers to sell them on that vision.
Uber's Marketing Vision
00:39:17
Speaker
Why do you even need a car? If i lived in I live in the city, why do I ever need a car? you know Calculate insurance plus lease costs plus gas, and it's more expensive than an Uber. So like right now, all of a sudden, you create this demand of...
00:39:35
Speaker
Yeah, we're just going to replace your entire car. that that's like That was the vision. And that's that's still, the I think, the ultimate goal. So that's that's how you that's how we thought about it in marketing.
00:39:46
Speaker
Cool. I know we're we're running on time right now. I have one more question and then then we can kind of kind of wrap up. You mentioned building out different use cases as a way to expand maybe the the spectrum of of of demand that that exists.
00:40:04
Speaker
I'm really curious tactically, and I'll just give you a very specific example, like I'm on Facebook. I have a certain amount of budget. I'm doing purchase-optimized campaigns, direct response, right? And First Self, the credit-building app I work for, have low credit, build it with Self, get started today, like that this type of messaging.
00:40:27
Speaker
And then... I want to shift some of my budget, you know, 10, 20% of that budget towards reach optimized. I want to get lower CPMs. i want to get a bigger audience.
00:40:40
Speaker
And I want to say something to this larger audience so that they eventually get into that bucket of the performance optimized campaigns. And that's kind of the thing I'm asking about. You kind of mentioned use cases.
00:40:54
Speaker
There's a challenge sometimes where maybe you can't just say here's seven different ads for the seven different use cases. Like I need to have one thing. How do you think about something like that? Right? Like I'm trading higher CPM, more intent based ads for lower CPM, low intent ads.
00:41:20
Speaker
Yeah, it's still in the service of eventually being able to serve more volume of customers that are reasonable see reasonable, because if you never spend the time building the demand or even creating that demand on the low CPM high reach, you just naturally saturate and tap out.
00:41:42
Speaker
And so what this is doing is saying, were we're going to take a ah very short-term hit on our business right now, but to create a more like sustainable funnel, knowing that we're trying to create, like with this low CPM, high reach strategy, someone who will eventually look like this like high CPM target audience.
00:42:07
Speaker
right And so the reason like you make that sacrifice the way you justify it is, yeah, otherwise we're just going to run out. And the like you could take a 10% to 20% hit now, or you could wait, completely saturate, and then you take a 10% to 20% hit for like four to six months because you have to rebuild that funnel.
00:42:30
Speaker
So we're going to take 10 to 20% hit now for like a month, but eventually that 10 to 20% funnel catches up versus taking a hit six months down the road.
00:42:41
Speaker
That's the way I think about it. And in terms of like picking use cases and all of that, like, again, it just goes back to like the roots of like knowing your audience, like what's the cell what's the value prop that... convinces people to join the most. Like you have to know why people are joining now instead of thinking about like, well, how do I convince them? Well, don't try to convince someone.
00:43:00
Speaker
Tell them what's already working. Like there's no like for like for Uber, it's like a million people have now taken this trip to the airport. okay, well, I never thought about it, but if a million people thought about it, then okay, cool, maybe I'll think about it the next time I have to.
00:43:15
Speaker
There's no reason to reinvent the wheel for that low CPM audience. so Just tell them what's working with the high CPM audience and put imagery that puts them in that situation, right? like Like what is the evocative image that says, oh yeah, so annoying. Like, like you know, maybe it's like the bill at the end of the 14-day garage airport and it's like, oh my God, I can't believe I didn't pay $200.
00:43:41
Speaker
Well, you could have avoided this if you had just taken $25 Uber. But you know the airport use case is good for your business. You know that's how a lot of people take Uber first.
00:43:52
Speaker
So just take that and use that as your use case. Like don't Don't reinvent the wheel. just Just do what's working, but position it in a way that will eventually work. I think it's the best way to think about it.
00:44:05
Speaker
Awesome. Sundar, thank you so much for coming on the show. Where can folks find you? ah You can find me on LinkedIn. And I've also got a newsletter. And soon I will also have a podcast. So yeah, a lot of exciting stuff coming around.
00:44:19
Speaker
Cool. Thank you so much, man. Have it going. Thanks for having me, Paul. Yeah.