Episode 306 / Izzy Tepekoylu / Zopa Bank / VP of Growth
Zopanomics & Finding Joy in Work
Subscribe to Our Newsletter
Show up, do good work, find joy.
That’s the mantra developed during Covid by Izzy Tepekoylu, VP of Growth at Zopa Bank.
Showing up and being present in the work you do makes a huge difference, says Izzy. After all, a forty-year career is made up of 10,000 work days, it’s not about heroic bursts. Instead, Izzy’s advice is to choose consistency, steady improvement, and importantly, finding joy.
“We don’t have to be serious to do seriously good work,” he adds. Humour helps create relationships and improve our everyday lives, so why not adopt it, regardless of what industry you’re in?
On the podcast, we talk about the importance of simple maths in marketing, but also about the shiny new object of Zopanomics - from the simple to the complicated end of data driven marketing.
Tune in to hear how Izzy and the team at Zopa Bank screen customers to understand where profitability lies for paid performance in fintech. And find out how to develop a growth mindset and find more joy in your work.
Transcript
The following gives you a good idea of what was said, but it’s not 100% accurate.
Izzy Tepekoylu 0:00
The reality is that almost all business problems you'll encounter can be reduced to something as simple as two variables, two equations. The math isn't the barrier, it's the mindset.
Speaker 0:15
Are you a fitness loving foodie? Then defy is next is for you. Defy offers real high quality chocolate you love, combined with functional ingredients, giving you as much as 26 grammes of protein per bag. So you can be both a little bit bad and a little bit good at the same time as a female owned brand, defy dedicates a portion of its profits to other women owned businesses currently sold in the US. You can find us on defy with an i snacks.com and follow us on social at the five snacks.
Tom Ollerton 0:42
Hello and welcome to the shiny new object podcast. My name is Tom Ollerton. I'm the founder of automated creative, the creative effectiveness ad tech platform, and this is a podcast about the future of data driven marketing. Every week or so have the pleasure and the privilege of interviewing one of our industry's leaders about the future of data driven marketing, and this week is no different. I am here in the house of Izzy Tepekoylu.
Izzy Tepekoylu 1:13
Nicely done.
Tom Ollerton 1:14
Thank you. And for the second time, so went to Izzy's office, and we did a brilliant podcast, and I didn't, I don't know where it's gone. It vanished. And that's the first time I've made of school boy error in over 300 podcasts. So Izzy is being so kind, not only to do this again, but to invite me into his like, literally, beautiful home. So thank you so much. I really appreciate your time, and I know you've had a really hellish last couple of weeks, so finding the time to do a podcast, for the second time, it's no small thing. But Izzy, look, if anyone who doesn't know who you are and what you do, can you give us a bit of background?
Izzy Tepekoylu 1:51
Absolutely, absolutely Tom and no worries, it happens to the best of us, happy to do it again. I had a lot of time the first time. Hopefully this will be better. So I'm originally from Turkey, studying mechanical engineering in Istanbul, and found my way to Purdue University in Indiana to get a master's in operations management, and then I joined a company called Capital One, which is a well known company in the US, as I would call them, the OG of fintech. Before FinTech was FinTech, that was Capital One, and I started there as an entry level analyst. The company was small. It was a niche subprime credit card issuer. And I was there for about a dozen years, and by the time I left, it transformed itself to the fifth largest bank in the US, and today they're the largest credit card issuer globally. Bigger dynamics, bigger than Chase. During my time there, I did from kind of account level analytics to strategy to managing different parts of the business, like cross, sell, balance transfer, small business cards. About a decade ago, we moved to London. I made a big shift and joined Google, originally in finance, then marketing strategy and operations, and eventually heading up ads marketing with a focus on, how can we steal some TV advertising dollars to YouTube. About two years ago, I reconnected with Jaidev, the wonderful CEO of Zopa bank. And at the time, I remember him saying, Hey, we don't really do marketing. And I'm like, What do you mean? You have a million customers? And it turns out Zopa had entirely grown through affiliates and aggregators. And so now I lead growth at Zopa, overseeing owned, earned and if you really have to paid marketing channels. And we have big ambitions, we want to get Zopa to 5 million customers in five years.
Tom Ollerton 3:22
So quite the career. In that career, what has been the best new belief or behaviour that's really changed your work life for the better?
Izzy Tepekoylu 3:31
Yeah, yeah, we forget about it. But we went through this whole thing called Covid. Do you remember covid? And working at Google at the time, I just just, I remember my days becoming a kind of endless log of video calls, one after the other. I just, like, felt like automaton going through the motions. I'm an extrovert, like I was like the self appointed mayor of the Google office on the sixth floor, type of thing, and no commute to bookend your day. You start your morning, you end you just like it was, like the slog. And I'm like, What's the point of all this? What's my purpose? Blah, blah. And I find people at two categories, like one, people who know what they want to do in life, and a whole bunch of us, me in the second camp, we don't, we don't know what we'd like, what's our calling, what's our mission, what not. And I was listening to Alain de Botton, who I absolutely love, like a very stoic outlook on life is like we're all marred by our childhoods and we're fundamentally lonely and we'll eventually die. Something struck a chord with me, yeah, and I realised my mindset of "what's my purpose blah blah" was too grandiose, too ambiguous, too overwhelming, like, simplify it down, like, dramatically. And so I started one meeting at a time, one day at a time, and I developed a little mantra: Show up. Do good work, find joy. Showing up is about kind of choosing to be present. If you're going to do a call, be there fully engaged, not distracted by email or something else, and make a conscious choice to fully occupy the moment. Yeah, doing good work is not about perfection. It's about delivering quality as I define it. Forty year career is about 10,000 work days. It's not about heroic bursts. It's about kind of discipline, consistency, steady, improvement, and then finally, finding joy. This is the crucial step. It's about being intentional, reflecting on your day, on your week, and noticing what sparked joy. Surprisingly, there's always something. It might be like a beautiful phrase, someone said something and then, like this, one clear chart and slide 44... 34 of a deck that really answers the question or some unexpected, insightful customer comment. Joy is always there. It's just you have to be curious and attentive enough to notice it. So this simple mantra, show up, do good work, find joy transformed my daily experience, not just professionally, but personally. I mean making every day in its own quiet, practical way.
Tom Ollerton 6:03
So it reminds me of a talk by a lady called Lindsay Barrett, who is the Global Head data driven marketing at Mars Pet care, one of our clients. We worked with Lindsay for a long time, and she stood on stage at Madfest and said My principal edict to my team is to be a delight to work with. And so and I said this to Amy, our global head of strategy the day, I said, you know, I've really taken on Lindsay's advice. And she's like, No, you haven't.
Izzy Tepekoylu 6:36
Look, like I tell the same thing. When you think about it, the most amount of time you spend in your kind of living hours is the people who are around you at work, more than your spouse, more than your family, more than your kids. So if you're delighted to be around I think that's that's I guess. That's a great thing.
Tom Ollerton 6:53
Yeah, I always say that if you make a friend at work or through work, that's like one of the best pay rises you can get, you know, because, as you say, you're 10,000 days you're in there, in your time to time, you're doing terribly boring stuff, or stuff goes wrong. And if you've got a friend in that that's better than a than a colleague, then you're like, you're enriched, right? Yeah, you're more joy, absolutely. So I'm gonna push you slightly on joy, though, because it's quite it's a bit of a nebulous term. I remember an agency I worked with years ago, one of their clients, their kind of, their reason for being, was all about being joyful. And no one could actually put a finger on that actually meant, so what? When you say joy, what do you mean? Do you mean excitement? You mean dopamine?
Izzy Tepekoylu 7:38
I think dopamine has a lot to do with it, but it's so overworked with the kind of social media addiction that we all have, so we get a lot of it through that. I think it's something that puts a smile on your face, something that makes you think and ponder, or challenges like a core belief that you have, or some hypothesis proven or disproven. It's a very personal, I think it's a difficult claim back to your point, like we're going to be joyful. This whole mantra is about what I define as joy at that moment, at that moment of reflection. So it could be many things.
Tom Ollerton 8:13
Yeah, when we were writing out all of our values at Automated Creative and haven't really changed actually, in the best part of 10 years, and we very much based our our values on the framework that was on Bruce Daisley's podcast. I think he'd say work repeats. And one of them was like, Don't ever force fun. But when it happens, go with it. You can't be like right now, now this is where we like have fun, exactly. But if people are losing their about a thing, then whatever you do, don't stamp on...
Izzy Tepekoylu 8:47
Nothing, like nothing like some corporate, organised fund. Tom, no, I agree with you, I but I really love that, because if it's spontaneous, that's the only time it's going to happen. And just get out of the way. Let it breathe, let it have its moment, and then you'll come back and you'll be better for it.
Tom Ollerton 9:10
So what bit of advice do you have someone listen to this podcast who wants to become a better data driven marketer? What's your one silver bullet?
Izzy Tepekoylu 9:18
All right. So many people are unnecessarily afraid of basic math. I'm talking like four operations, right? Adding, subtracting, multiplying, dividing, like I'm an engineer by training. I'm a statistics enthusiast. I even knew at some point around multivariate regressions to do predictive models. But honestly, 90% of what makes you a great data driven marketer comes down to these four basic operations. Like, if you can do those, you're good. But people have a often, often have a mental block. They convince themselves that analytics or data driven decision making is complicated and scary. It's really not. The reality is that almost all business problems you'll encounter can be reduced to something as simple as two variables, two equations, like, break even, right? Like, what does our conversion rate need to be for this marketing campaign to pay off? Everyone with a high school diploma can figure this out. The math isn't the barrier. It's the it's the mindset. So my best advice is to know, no, don't take complex analytics classes, blah, blah. So adopt a growth mindset. Stop being scared of numbers. Lean into simplicity. When you stop fearing basic math, the clarity, the confidence, the power of your marketing decisions will increase dramatically, man.
Tom Ollerton 10:38
So you're really speaking to me there because I like, if anyone mentions politics or maths, my brain just like, or golf or something do you know what I mean, I'm just like, oh God, really, we're talking about golf. So there's an irony here that I run a business called Automated Creative right, which very much a statistics led business, but I'm not particularly maths focused. I get it. I understand it. So how? What's your advice to me to become a better data driven marketer using those those four principal calculations?
Izzy Tepekoylu 11:06
First, let me unpack the kind of mindset piece, right? Just like you said, right now, I'm not good at math. The growth mindset is about I'm not good at math yet. So that's the mindset piece. The second is the simplicity sometimes in a group setting, in a team meeting, people are scared of coming across, oh, that's a stupid question. Like actually, the simple, the more basic, the simplest questions are the hardest ones to ask, because it takes courage and actually not that difficult to answer. Let's go back to your kind of industry, automated creative. You're looking at two pieces of creative. Why did this work better? It's a simple question. The answer is kind of, there's a numerator and denominator. More people saw, more people clicked. That's the basic division. And then they're like, Okay, why did they do that? Follow that to the next step. That's probably like this many people saw this thing, this many people whatever, as a multiplication so simple questions answered simply, without jargon, and getting into, oh, to this, that and the other. I think that's kind of what I mean by simple maths that helps.
Tom Ollerton 12:24
This episode of the shiny new object podcast is brought to you in partnership with madfest. Whether it's live in London or streamed online to the global marketing community, you can always expect a distinctive and daring blend of fast paced content, startup innovation pitches and unconventional entertainment from madfest events, you'll find me causing trouble on stage, recording live versions of this podcast and sharing a beer with the nicest and most influential people in marketing. Check it out at ww.madfestlondon.com
Tom Ollerton 12:54
So we're now going to move on to your shiny new object, which we've never had before, not in over 300 shows, which is Zopanomics, yeah?
Izzy Tepekoylu 13:09
Economics, but Zopanomics.
Tom Ollerton 13:11
Explain that to me, all right. What is Zopanomics? Give me a simple elevator pitch, and then tell me. Why have you chosen it as the future of data driven marketing?
Izzy Tepekoylu 13:21
Yeah. So it's going to be a bit ironic, because I just said, Hey, 90% of the kind of stuff that you're going to encounter at work is going to be simple. This is firmly in the other 10% bucket. Okay, this is complex, but worth it, which is why I'm excited, excited about it, right? So first, a little bit of context on kind of paid performance in financial services, it's notoriously tricky. Like, paid performance is always tricky, but financial services a bit more tricky because there's a long delay between when you acquire a customer and when you decide when you see whether that customer was a profitable one, or a good customer or not. Like, like a scenario right there. So it's summertime, you want to kind of get a loan to go on a nice vacation, you see our ad, you click, you approve, we give you the loan. It'll take us nine months to see if that breaks even, and then maybe another four months after that to see if it's if it's a kind of profitable thing. So now imagine 13 months from now, I look back and say, that was not a good customer to acquire. And then imagine I go back to the algo, Meta, Google, Tiktok, whatever platform send me that lead to say, hey, remember that lead you sent me 13 months ago. That was bad. So this whole delay in acquisition versus kind of the profitability decision being this long, it's it's impossible for the algorithm to work. The algorithm needs something today, couple of days, maybe a week. So you need to be able to tell, hey, this was a good customer. Send me more of these. I'm willing to pay more or kind of don't send me these because they're not they're not paying back. Okay, so here we need to build. We built a predictive model so that we when a customer comes and logs in, when a customer comes to our website, we already know a whole lot of data about them. If they log into our app, we know a whole bunch more about them, and at that point, we apply this model. It's a causal inference kind of machine learning model.
Tom Ollerton 15:21
So explain a causal inference machine learning model.
Izzy Tepekoylu 15:25
Sure. All right, like the way I think about machine learning is it's as if having 10,000 interns at your disposal. If I had 10,000 interns so I could tell them, hey guys, here's my million customers. Each take 100. It is the kind of 10 million logins that I had over the last three years. Tell me which login is the best one. Tell me which one of these logins turns out to be a great customer or not. You can do that. They'll take a couple of weeks. They'll be like, Hey, easy. It is the ones that you look like a high value customer. The causal inference does that at scale and instantly. It's basically a kind of predictive model that statistical shortcut that does what those interns would do for you, and we use that to assign a value to customers instantly, instead of in 13 months, so that we can go back to the algorithm and say, Hey, Google meta, this is a great lead. Send me more of these. I'm going to pay more or like, this is not great. People not don't send me these. We're not going to pay less. One one side track on the algo. Here are you. I said Google too much. Sorry about that. I was trying to turn that off, actually, but yeah, aside on the algo, right? I because I have the benefit of spending some time at Google, and I see I had a lot of customers who didn't trust it, who didn't trust Oh, you guys have your time on your scale. You're trying to maximise your whatever's you have to trust the algo, Google, whether it be Google, whether it be meta, whether it be tick tock. These are really, really sophisticated models that these companies have invested millions of engineering hours in, and your job as your mark as a marketer should be. How can I work make this magical kind of technology work as hard as it can for me? So that's my little kind of detour on how one thingone should think about algo, and it's gonna get, gets gonna get even, even kind of more stronger. Zuckerberg is out there saying, why? The end of next year, all you have to give me is your product. We'll do your targeting, we'll do your creative, we'll do your campaigns, all. We don't need to do any of that. So that's where the algo is going. Being able to know what signal to send to it now is going to be even more important in the future.
Tom Ollerton 17:51
So can you tell me more about what goes into Zopanomics. So obviously, not without compromising your intellectual property. But you talked about people arriving at the website, they're going to put in a certain amount of detail, their name, their age, location, when they got an IP address. And then you also said the app as well. So all of those, I'm going to push you on this, all of those things seem like they couldn't possibly predict what a good financial outcome could be. So without giving away too much, what kind of things are you collecting that you can build a model on?
Izzy Tepekoylu 18:27
When somebody comes to zopa.com and let's say they clicked accept all to the cookie banner. At that point, we know 40 to 60 pieces of data, and that they not just Zopa, by the way, everybody, we all do. And then when they click and say, download the app, and they log into the app, at that point, we also know their address, their PII, and their income as a like as a financial institution, we know a little bit more about customers, but at that point, you're at kind of 100 data points, and that's a lot. That's a lot of data to for everybody. By the way, any advertiser out there, putting the kind of PII aside, even from a simple website visit, you can glean so much more. And for whatever it's worth, this model, it learns over time. So if it's not 100 it doesn't have to be 100% accurate. You're not lending money. You're not like making massive investment decisions. Here you're trying to decide, should I pay a pound or pound 20 for the next lead? So the stakes are a bit lower, and you have the benefit of it will learn on itself. It will learn, oh, that was a good decision. That was a bad decision. Let me go revamp. Let me go. Do my little feedback loop to improve myself and the so that's kind of one piece. And for us, Zopa is kind of, we've been using machine learning for nine years. We have data scientists. All of our data is in the cloud, so it's a bit more natural for us to be able to kind of pull up a data scientist for a couple of months and then kind of some tech folks to make the tracking work. But even, even, even if you don't have that, there are companies out there, one of them I'm not affiliated with. I'm not getting paid for this. It's called boyantis.ai, out of Israel, and they sell a, kind of a value based bidding kind of this Zopanomics solution as a in a box, like a SaaS solution in a box. So you can actually go talk to Boyantis. They can come kick it off. You'll need some budget, some scalable budget, about 50 conversions a week. So whatever that budget is for you, for your business, however much that costs at that scale, they'll be able to get you up and running in a couple of months.
Tom Ollerton 20:49
So let's go back to the comment you made about the Zuckerberg announcement that you give me, give me your pack shot and your bank account, I think you said, and I will just, you know, I will deliver, get we'll chuck it in the black box, and we'll, we'll give you cash at the other end. So what's your take on that? Because that sounds like an engineering solution. So you've got half the, not half the industry. Some of the industry are going, you can't, but you can't provide an engineering solution to a people problem, which is what advertising is. So, but this here, what you're talking about is a machine using the mindset and the skills of your pre existing team to just make sure that your bidding is correct. So what's your take on this?
Izzy Tepekoylu 21:31
You're gonna like this answer. But I'm not giving it to you, because you do what you do. Look imagine, let's go to a world where all the advertisers are doing exactly that, right? And so everybody's using kind of Zuckerberg, I have a feeling all ads would look the same. I mean, they look different, but like some version of, kind of modified AI slop, right? So it will be, yes, they will work, and will be more on the merits of the products and whatnot. But so many products are the same out there. Kind of the all the FMCG, fast moving consumer goods, products come to mind, right? So I still have a core belief that everything you do in marketing, two thirds of your success will depend on the creative. So all these things, Zopanomics, I'm trying to eke out that last third, right? That's kind of about but if I don't have something that that's compelling to a user to stop in the middle of their feed to say, Oh, what is that? Let me click. If I can't get that, I'll do a third kind of my Zopanomics and the machine based, value based bidding and all that will be a third. Two thirds will still be creative. So like, going back to your question, I'm sure there's going to be a whole bunch of companies that says, yeah, man, we don't, we can't do creative. We're not that good. Just to give it to kind of Facebook, whatever, but the ones that differentiate themselves will still be about creative is still be about why does this product kind of exist in the world? Why is it kind of reason to be famous? How should we talk about it that's differently than everybody else is talking about it? So that's kind of my, my kind of prediction. So that's where it's gonna go.
Tom Ollerton 23:14
It reminds me of two things. It reminds me of, first, Jeff Goldblum in The Fly and secondly, Adidas advertising around 10 years ago. So, so Adidas quite famously, went really big on performance ads, like 10 years ago, eight years ago, and it didn't work. Well. It worked initially, and they converted all of the brand equity that they built up, but then, because they didn't invest in the brand, then, you know, obviously the conversion slid, right? You know the story. But then equally, and I think that's where we might go is that everyone goes brilliant, AI slop will do, I can sack or I'm stop paying these marketing idiots over here. They can't prove what they do anyway. So we'll just give them a pack and off we go. And you know, the slop will convert to begin with, and then it'll slide right. But then I was listening to a podcast for my podcast challenge I was talking about earlier, before we started, I was listening to a podcast about The Fly and all of the horror gore fly effects are man made, and it's all, it's physical, right? There's no this. Even the title, I don't know, the titles probably have some digitization in it. And there was this whole range of horror and gore movies that happened from 2000 to 2008 ish, where it was all CGI, and then eventually they went, No, this is awful. We're really sorry. We've made lots of really terrible films. And now, you know, there's this, there's this move back, because there's something about the craft of making special effects that works better than CGI and everything. When you think about, you know, those early, you know, the Star Wars, what is it? Um, Phantom Menace, you know, just the old JarJar Binks. And I think we're going to go through a same thing with advertising, where, as I said on when we got here, like this, market is drawn to the easy button. Where's the way? How can I press the easy button and press it really hard, and they want to sit back, but as soon as you press the easy button, the quality goes down, and you have to fill up the rest of your day with stuff. Anyway, I'm dubious, and I'm just gonna say metaverse.
Izzy Tepekoylu 25:04
Yeah, no, let's see. Look, it's about zigging, what everybody else is zagging, right? Like, if everybody else doing that, you have to still stand out the way it goes there. So how are you going to differentiate? Yeah, what are you going to do that's different? Maybe it's just going to be black and white and text and, well, maybe, maybe, maybe you'll send, like, flyers, because nobody's sending flyers. I bet you flyers are kind of effective now, nobody's doing them anymore. I don't know. I haven't looked at it, which kind of, I'll take a note now, but yeah...
Tom Ollerton 25:30
Well, it's, I mean, well, the difference might be Zopanomics, right? You know, your first party data, your own data, your own intelligence, your own...
Izzy Tepekoylu 25:37
Exactly, exactly how strong the signal that you can build on the margins. Everybody's bidding like a kind of an obvious good customer. Everybody's bidding, especially in financial services, where you're up against the big six high speed banks with deep budgets. But what are those roughs in the gem? Gems? Yeah, gems in the rough that you can actually say this one doesn't look great customer to everybody else, but I know, because I have more of those guys, and I can predict looking at that, so I'm going to get that one, but it's kind of like a bargain. It's a bargain.
Tom Ollerton 26:07
Izzy, thank you so much for doing this podcast again. Thanks for the Turkish tea. Thanks for having me over to your house. If someone wants to get in touch with you, where and how?
Izzy Tepekoylu 26:14
Absolutely look. Humour is my unlock. Levity is underrated, especially in our world. People assume serious work has to be done seriously. I strongly disagree. We don't have to be serious to do seriously good work. Humour recenters us, diffuses tension, open our minds so much broader set of solutions. Stretch leads to tunnel vision. Laughter breaks through all that. So if you want to reach out, send me something funny, a joke, a meme, a clever pun, a hilarious observation about marketing or life. If you make me laugh, you got my attention, and probably the reply.
Tom Ollerton 26:45
Beautiful, thanks, Izzie.
Subscribe to the ‘Shiny New Object’ Podcast on Apple Podcasts, Spotify, YouTube and Soundcloud.