Top
Podcasts
5.26.2021

Radu Lupu – How Adidas uses AR to revolutionize footwear e-commerce?

The Cogniverse Podcast: In this episode, we are going to talk about how you can achieve success in digital product development starting from scratch.

Attila Tóth

Digital Strategist

“You always have to start from solving a user's needs.” Radu Lupu

Radu Lupu is the Senior Technical Product Manager at Amazon, and up until very recently, the Lead Product Manager of the Adidas mobile app, with former experience at Rakuten and eMag.

Radu is a true technology geek, his core strength lies in building digital experiences in a corporate environment, during which he has gained a profound understanding on how you can achieve success in digital product development starting from scratch.

You can connect with Radu on LinkedIn, and you can check the Adidas app.

Listen to this episode on Amazon Music, Apple Podcasts, Overcast, and Spotify.

Please enjoy:

Podcast transcript:

Attila Tóth: Hey, Radu! Welcome to the Cogniverse Show.

Radu Lupu: Thank you for having me.

Attila Tóth: I think we should start with your professional history. So from QA tester to lead product manager in just five years, it looks like you had quite a progressive career development. Could you tell us how you forged ahead?

Radu Lupu: Sure. So actually before starting in tech and doing QA, my university background was actually economics and, you know, during the university once I sort of figured out that becoming an accountant isn't exactly what I'd like to do. No offence to actual accountants I’m sure there's so much love in their job but I just wouldn't. I started looking and trying to figure out what I would like to do, basically. And I've always had sort of an inclination towards tech. The majority of my childhood was basically spent in video games, which obviously at that point of time looked like the worse way to spend your time, but that also sort of helped me develop some form of affinity towards tech. The only thing I was more effortlessly good at maybe during school was programming, was IT, but for some reason, at that point of time, I didn't see it as an actual career path which I could follow, which came to my rescue later once I had the pre-midlife identity crisis, and trying to figure out what I want to do. So I managed to get this job in QA, basically, I worked for eMag, which is the biggest e-commerce company in Eastern Europe right now, as sort of a training step towards building products. Products at the time weren't even a thing since the whole concept of product management wasn't as widely publicised and as popular as it is right now. And then that led me to being in charge of all the tech projects for launching this e-commerce platform in a variety of countries, so I've been working on launching it in Hungary and Bulgaria for instance, and briefly also launching it for Poland. And gradually I started moving to bigger companies afterwards moved out of Romania, and took jobs which had a bit more and more impact, until this current situation with Adidas, where I've basically been in charge of building the Adidas app, which is the flagship e-commerce app of Adidas.

Attila Tóth: Cool! So you've had quite a ride, and I think I would like to dive a bit deeper into your current digital work. So let's start from the beginning: when did Adidas decide to make the first steps in digital? How did this Adidas app project start, what was the motivation behind it?

Radu Lupu: So first off, when I joined, digital was already kind of starting off, I think Adidas has always sort of been trying to go into this direction, but it wasn't a concerted effort until maybe eight-nine years ago when also Direct Consumer as a business model has become in all retailers a much-much bigger deal and a very valuable avenue for growth. So the reasoning behind going into digital was mainly led by that, by achieving a direct sales channel to your customers, and it started off with establishing Adidas's e-commerce site. So this started off a long time ago and gradually got built up. Along the way, arguably relatively late compared to the rest of the industry, the benefits of a direct mobile app and a direct linkage to the consumer became much more prevalent and evident to Adidas, and the main selling point basically of having a native app is that you're much closer to your customers, you have a direct bridge since you're in their pockets, push notifications are a big thing, the majority of traffic anyways moving towards mobile, and the vast majority of mobile users prefer using apps rather than websites. So the overarching idea here was to drive both more mobile growth but also to ensure that we switch from the old way of viewing e-commerce – in which you think of every user as a session, which you need to convert as fast as you can – to a longer-term relationship, in which you start looking at concepts like customer lifetime value and how to ensure that you keep that connection with your customer. There's obviously a lot of benefits then, you know, in terms of marketing spend, which then would go much lower than having to reacquire the same customer 300 times, and overall achieving a better cash flow with that recurring revenue.

Attila Tóth: Very interesting insights. And I think as Adidas is a major brand there is a topic that I would like to pick your brain on because I think, and what you said it really proves that Adidas walks the talk when it comes to digital transformation, but from the outside people might say "Well, it's easy, you’re Adidas". However, based on my experience, I know that even established brands face quite a few challenges when it comes to digital. And I would really like it if you could share some setbacks and, basically, also the dark side of digital because usually, nobody talks about that.

Radu Lupu: Right. So, in any large corporation – as you duly noted – basically, you have a huge amount of stakeholders, a huge amount of departments, a huge amount of differing opinions, and different interpretations. All of that is generally mitigated by having a strategy as a guard rail, but that can be easily interpreted in different ways. So, let's say the biggest challenge when trying to establish a digital channel inside of such an organization is ensuring you get both buy-in on what you think as a product manager your users want and your organization is good at doing, versus what the rest of the organization would want to do and what type of different avenues or ideas they might have. So it becomes a really delicate balancing act in a way, in which you need to have a very clear idea of what you want to do and why you want to do it, and then have proof in order to get this sort of balance in place. It's then much easier to say no to certain things, adapt to other requests and try to make – I don't want to say everyone as happy as possible, because that's obviously a false premise, which you'll never achieve, but – trying to keep that buying at a certain level. And we've had challenges, especially with the Adidas app, when it came to what the Adidas app translates as from a product vision perspective to the overall strategy of the company, which obviously got picked and torn apart in various ways by various stakeholders, but as long as we came up with a concrete idea of what we wanted to be, that was clearly an e-commerce focus to happen, our user base clearly expects a certain set of functionalities, this is what they want and this is how they interact with it, then that is much easier. If we did not have that, probably the Adidas app would look much-much different today. The amount of requests I've gotten to rebuild Instagram, Snapchat, Spark AR, and you name it, and the app is a testament, it's that danger. So, I think that's the main one, but on the other hand, there is a big benefit, which obviously then helped our growth, and that's obviously not only been in terms of money because that's actually not that true, we are a pretty lean demand in products. The benefit is the brand basically, which obviously helps drive adoption quite easily. Obviously, if I had the same product in a completely different, smaller company, let's say, we wouldn't be looking at over 40 million users after four years or so, so that's a huge plus of it.

Attila Tóth: It's a great insight, and I think you touched really well on these potential conflicts that can come up in discussions because many people have ideas, and to keep that focus in one direction, I think that's a very hard job and requires a lot of juggling, not in terms of only technicalities but also in terms of communication, aligning with the right stakeholders and keeping that vision and keeping that focus. So I fully agree. And also it helps demystifying the common belief that if you're a major company, digital will be a smooth transition.

Radu Lupu: Definitely not.

Attila Tóth: Now I want to learn more about probably your most efficient achievement. You have been promoted to lead product manager position of the Adidas app in less than 12 months, because you achieved the target KPI by 115%. So how did you do that? What tactics helped you to be that effective?

Radu Lupu: What I was speaking about before around strategy as a [...] and achieving you're having a clear idea and a product vision, once you have those as foundational pieces in place, those can be a good avenue for fast iteration and fast action on execution as well. So you have your North Star, you know where you want to go, you have the equation which leads you to whatever success metric you have, let's say it's revenue purely, then you have a clear indicator of what are the building blocks, which drive towards that. Is it attaining more users? Is it increasing my conversion rate? Is it increasing the retention rate? And so on. If you take those, and those building blocks are clear, you can then come up with specific features, or it's not only features actually, it transcends just pure development. It also means getting more buy-in from a growth team, which owns a marketing budget, and getting that, it can mean convincing different country managers to then launch this app in their country as well. It's all of those things. As long as you know what you need to achieve, it's easy to come up with actions which get you there. Now the second part of it is then being able to action in a fast pace. And something I'm very proud of is the team because we – it's probably a very standard thing to say, which you also maybe hear a lot but – we do see ourselves as acting like a smaller startup in the Adidas space. Basically, how you achieve that is by having all of the necessary tools and people in your own specific team, which help you achieve these things. Of course, it's never going to be a 100% true, so you're going to have dependencies on a lot of other things, but trying to minimize that greatly increases decision making time and execution time afterwards. So, how we achieve a faster pace of delivery, basically we don't view roles as locking you down to doing a very specific thing in the team, the role is something we view as being fluid in a way, so for instance, if we're lacking development power for a specific new project or a bigger overhaul which we're doing. Then instead of just hoping for the best and pushing that one person to deliver faster, we then start pitching in and writing code ourselves. For instance, I go in and also do development actively whenever needed, we have our developers then also figuring out different sort of new patterns, or how to use standard Android interactions from a UX point of view, so there's a lot of fluidity, which you can also view as dangerous, but since we were relatively tight-knit, we do ensure that we do follow a common strategy and we do have a good quality of our delivery there. So I think these would be the elements which helped us go very fast, and still help us go relatively fast because we were constantly going over what we either planned or was planned for us.

Attila Tóth: That's something really to think through. And I really like this idea of fluidity, which means that you don't only focus on your core competencies, but you try to be open-minded and connect the dots and that I think also requires the right team and the right people to have that mindset in place. So that's truly amazing. And, product owners, who are listening to us now: this is the mindset you need! Radu just showed it. And yes, let's talk about the new app, which you have lately launched and it's one of the biggest augmented reality digital catalogues in the industry. First of all, congrats, I think it's a huge thing you have done, and this proves that Adidas takes things to the next level in digital, so it's not just doing what's the industry standard, but going a bit forward. Could you share some ideas, and how this product development strategy was created, and how did you start it from being an e-commerce, then to a mobile app, and now to an augmented reality mobile app?

Radu Lupu: Actually, augmented reality started as a sort of a pet project of mine,  which I started maybe more than two years ago. Basically, the reason why we started this is to solve two things. Basically, every feature we're prioritizing and we're bringing forward is meant to solve a specific pain point, otherwise it's just work for the sake of working, from my point of view. And the two main issues we are trying to solve with this is: The first one, which is the easier one is having a good preview or understanding of how products look like. One of the biggest problems in any fashion retailer is return rate. Return rates are generally huge, and those are due to on the one hand this issue of the physical product not really matching 100% what you see in a picture, and that since is a very visual thing, this depends on a lot of different conditions: one, the studio lighting of the pictures being taken; second, even the color settings of your monitor can have a huge effect on the final color you see, right? Then when you get the actual product, you see it in your home, maybe you have a fluorescent light bubble or whatever, and it's not the shade of green I actually wanted, like why is this the case? The second is how things look with actual volume behind them, right? There's one thing you see in 2D, and you see a 2D representation of a full object, and it's a completely different thing when you can actually view it from all angles and understand its dimensions in actual real space.

This first one, which is – again – the easiest thing to solve, we solve it with augmented reality, which by the way is extremely easy to put in place, like apple and google have made it so accessible right now, that almost anyone can go in and develop something like this in a very short amount of time. The second problem we wanted to solve then – and this is arguably the biggest one – is sizing and fit. Basically, the problem here is sizing is generally not the same across different retailers. A size 44 shoe in one retailer is going to be probably different in another one, you have no guarantee of that situation. And even if that's not the problem or if that's not the case and you get lucky, and the two brands somehow magically match, the other issue is base off of the dimensions of the shoe and the sizing of your foot as well and the composition of that shoe as well, like is it more elastic, does it have more non-flexible parts in it also determine how comfortable those things feel. So we're now also working on trying to take the information we have from the 3D model of the shoe, which gives us clear understanding of what it's made, we also know what it's made out of, and we can take all of those data points, and then taking the input of your foot or your shoe size, or maybe even going as far as making a 3D model of your foot, which is now easily doable with a variety of ways, so you can just take a video and there's a platform, which actually been applying AI, determines the 3D model and determines the volume and the shape base off of just the video. You can use the iPhone's IR blaster, which is used for facial recognition, also to map a 3D object, lidar, there's a bunch of things, it doesn't matter. You take those two and then you try to match them together into figuring out: will it fit, is it good, do you need the bigger size, do you need a smaller one, and so on and so forth. So this is the second thing we're currently working on actively. We started off with these problem statements and at the time augmented reality had just been made much more easier to use, I think it came out of one of the dub-dubs or one of the Google I/Os as well as an easily plugable experience, and we started doing some testing inside of Adidas and with some of our users, but the main challenge there was actually getting 3D models, which you might expect you already have, but you don't. And that was basically the majority of the work afterwards: trying to partner up with different parts of the organization to then have this as a broadly available feature set for all of the different types of shoes we might have. So it was - wouldn't call it an uphill battle there but it was a lot of work on the side to get enough buy-in, to figure out a way to scale this production in a cost-efficient way.

The first 3D models: they were very expensive because they were being built from scratch, and you needed a full-time 3D artist to figure out how to do those – and those are not generally cheap, as you might expect –, then going into looking at alternatives. And now there's a lot of them on the market. The saving grace we got was photogrammetry, basically. We started installing a photogrammetry rig in our warehouse, and then started simply capturing and scanning shoes from the warehouse, as a standard sort of working process there. And the goal here is to have maybe 100% coverage of products with 3D. So going from seeing a product in just images and video, and then just enhancing that with 3D. You know, arguably it may be even the main asset, by which you see a product. This has a lot of applicability and a lot of places where you could add it in.

Attila Tóth: Wow! It was quite a challenge, and thanks for sharing this because many people still believe that digital has to only do things with technology, but as you said it's a background process and you have to align with different parts of the organization, and then you had the change in the organization to have these scans created in the warehouse, which is amazing. So I think for all of us listening, especially product owners: this is the mindset you need, so don't only think in terms of "okay, what we can do with digital", but actually in the process, what you can do in the background that helps that product grow. That's a really good insight, thanks for sharing that.

Radu Lupu: Arguably the tech side of the things it's probably the easiest side. All depending on your background as well, if you're non-tech then obviously it's going to be much harder, but yeah, tech is probably the simplest part of the whole equation in a big organization.

Attila Tóth: For those people who are listening from non-tech, don't be so afraid, Radu is a tech genius, so that's why it's so easy for him. Going a bit into the current, let's say, economic and social situation: how did COVID-19 impact your work style, your teamwork styles, your meetings, collaborations, what was the change?

Radu Lupu: So, obviously, Adidas is a huge brand, right? And that you normally would be buying Adidas shoes, and the barrel and anything else is in retail stores, that's a very big part of the business, obviously. Now with the pandemic, obviously, those things started being shut down, going away in terms of overall pandemic lockdowns, which then obviously affected the business a lot, but luckily, as I said at the beginning, Adidas had started investing in digital, way before this ever happened, and we already had a very strong channel, our own channel, which we could have as a bridge for this time period. Without any doubt, retail remains a very important part of the business, and will most likely stay for the long term. You can't beat shopping in a retail store when you're trying especially for apparel and clothing, to see how everything looks like, and it's also a social and personal thing as well, right? But what we've seen – like also every other retailer, which had an e-commerce presence as well – we've seen a huge shift towards digital, which was expected, right? That definitely helped that investment, and paid off for the business quite well during this time period.

I don't want to imagine what would have happened if we did not have a known channel, we did not have any other digital partners as well, because we're also working with the likes of Zalando, Asos and so on, which then represented a big boost for us. In terms of how it impacted actual work though, my team, not as much, to be honest, because we were already constantly running at full speed, essentially, because we kind of really enjoy the process, and we enjoy seeing the results of our work. So we didn't have a need to get even faster or any of those things, but there was a definite shift in priorities, you had something like the Olympics, which was a very big deal for us, and we were preparing a lot of different experiences and feature sets for that specific event, which obviously never materialized at the end. What one of the teams did do then is adapt some of the features we built for this experience into something, which would then help in this kind of situation. For instance, managing queues in stores. A lot of our customer base comes in to pick up products with click and collect from our stores, or when we have larger launches for very-very popular shoes, you end up also with relatively huge queues in front of it. So in order to counteract that, we adapted one of the features we had built for this other situation into a way to manage queues in stores, and basically, have time slots being bookable in order to minimize the amount of people there at any given time. So there were some stories of adaptability there, and which in the end of the day are used even now, and i'm also very proud of the team, which came up with this in terms of instead of looking at it as wasting effort into something, which would never materialize at least until this year hopefully, taking it and thinking outside the box and adapting to change.

Attila Tóth: So that also requires quite an agile mindset, right? So to reuse what you have planned but put it in a different way on the market.

Radu Lupu: Exactly, but the whole job is that, basically, even if you don't have a pandemic, you release features, you reiterate, you figure out what the impact of these features is, you look at your data, you look at your direct customer feedback as well, and then you adapt to whatever the market dictates in a way, or your organization dictates, depending on the direction you're going. So without being able to adapt and to be able to come up with 100 different plans and directions, you're not going to make it, I think.

Attila Tóth: Just one curiosity here: you mentioned that you roll out the feature, and then you analyze it. What's the average time frame you do this? So how much you wait 'till you decide "okay, we need to revise this feature, or we need to actually drop it and do something else", what's an average time frame to do that?

Radu Lupu: That highly depends on the feature, and normally we look at it from day one, and we actively monitor customer feedback and trying to see if we missed an obvious thing. You don't need 1 000 data points for a common-sense topic, if you then see a clear anecdote saying that "this thing doesn't work as I expected it to" or "why can't I view something else", whatever particular perspective you have, and then you should take it and analyze it and then figure out if you need to apply that, and you don't need to wait a long time for it. You either have direct feedback for those points or you see it in some data with a sharp drop or increase depending on whatever you want to see or whatever the objective was, or in other situations you have a gradual evolution of the usage of that feature. And that then requires obviously a long-term monitoring of what that is and then adapting to it. It's not also only adapting from a technical perspective but for instance, if you're delivering content in your future, it's maybe about changing the process in which you deliver that content and changing the type of content we show, because from a technical point of view, you have an endlessly scrolling list. Great, that's not something you need to necessarily change, but what am I showing in that list, how are people interacting with it, do they expect to see it, do they use it at all, does it provide any value? And that's something you would see longer-term by monitoring. Three months is the standard we should be using for every feature. I think it highly depends, and I think you should be looking at your product constantly and adapting whenever a red flag appears or whenever a very good idea pops up in feedback, which you simply didn't think of.

Attila Tóth: As we discuss this, I think we should touch on a topic, which I saw and I still see in many brands doing this: so be it, they have a very well-defined strategy, and then they want to stick really to that strategy and they fail because they don't adapt, as you said. Or quite the opposite: so some people confuse agility with ad hoc decision-making, so thinking strategy is not important, the north star, as you said, is not so important, and then "okay, this is the feedback, we do it, comes another feedback and we do that". So what's a good balance to tackle that? Because it's clear you mastered that skill pretty well.

Radu Lupu: Well, I've had experiences with both extremes and like with anything in life you then figure out that the balance is the way to go, but essentially, I think strategy and having a clear vision on a company level is key to driving all of the other initiatives that you have, be it products or new product lines. Whatever they may be, that needs to be clearly communicated and understood, otherwise, you have a bunch of different stray cats running around, trying to get their pound of flesh through different means, which ends up with the sum of the parts becoming less than the individual sum. It's important to have that rallying cry or that foundation, on which to build up your organization. The tricky part is in translating that broader strategy into what each team is in charge of doing, and how they go about helping towards making that strategy successful, and following it through, and that highly depends on getting each of the elements of the organization to have a general understanding of what each is doing, and trying to come up with ideas, which help you work together. And, especially, not duplicating work; there's a lot of that going on in any company, where you have a different team trying to do product A which is very-very similar to product B in another team, which you can look at, and think of it like that's wasted work or effort, or you can look at it as sort of internal Darwinism, organizational Darwinism as well, and see which one fights to survive, but that's a very hard thing to master I think in getting all of these departments aligned. That's something which I honestly don't have a solution for, it's something I've been thinking about as well, but I haven't seen anyone come up with a foolproof solution to that working.

You can hammer in the strategy as much as you want, but at the same time all of us are individuals, and everyone interprets things in their own way. It's a matter of the course correction, basically, and that becomes, I think, a full-time job. On the other hand, on the things which you can impact is driving that sort of mindset in your own organization or team, and trying to think of ways in which you can achieve these synergies with other departments. And this AR example is a very good one, in which we couldn't have set up everything ourselves, we needed buying from different stakeholders, from different parts of the company, which you needed to then convince that this is the way to go, and it's worth doing, which obviously was not an easy thing. The same with the Adidas app as well: not everyone was convinced, and not everyone wanted this exact way of presenting it to the wider audience. So, let's say, once you have that strategy set up, and once you have this clear idea, it's about defining what's the success metric for each, like "why am I doing things, what am I trying to achieve, and how do I measure it?" Just broadly defining product success or company success is not really enough, because that – again – gets interpreted very broadly by anyone who's listening. You need to have a clear idea of where you want to go.

You also need to be very careful about what you set as those KPIs and how you measure them because I think the most dangerous thing you can do in a data-driven organization, let's say, is looking at the wrong data, looking at the wrong indicators, focusing maybe on vanity metrics: like focusing on my number of installs, but not focusing on my retention rate. It's important to find a good balance there and fully understand what are the trade-offs of looking at one KPI vs another, and how to construct that equation of success with all of those things in mind, and being able to monitor them. And then it's all translating those objectives into actual actions, which help achieve that. But it's easy to say, much-much harder to do, and it's a very generic description of the process, but it highly depends on a lot of factors: it depends on your organization, on your culture, on your team, on your product, on your user base, especially. It's hard to come up with, but once you do, it's much smoother sailing than without it. I can tell you from experience that definitely, it's much easier.

Attila Tóth: Great insights, thanks for sharing! Just to sum it up for our listeners: so agility doesn't mean you don't have a strategy, so first you need the strategy and the north star, but the strategy doesn't mean that you have to be rigid, so you have to be agile to get feedback, but agility doesn't mean you have to do things on an ad hoc basis, so you have to have clearly defined KPIs that you can also measure and understand, and have the right KPIs. Just to sum it up, I think that's the core, and I really love how you build this up, and as we are talking about product development, I get a lot of questions: "okay, what are the rules and responsibilities?", and there are three roles that people tend to confuse, so let's clarify these. So there's the project manager, the product owner and the product manager. Radu, could you share what are the differences between these rules?

Radu Lupu: I actually started doing project management initially. There are similarities between a project manager, but very few, honestly, between that and the product manager. Product owner and manager – to my understanding – got exactly the same thing, it's just different companies naming them in different ways, and also the product manager role is highly different in each individual company. The nature of the role itself is very fuzzy in a way. The broader description of it is, if your product was a company, you're the CEO of the product. Now, what does that exactly mean? Up to personal interpretation again, I can definitely say that product manager specifically is aware of many hats, you're sort of a Jack of all trades on the one hand, understanding all of the different aspects of your product, going from marketing to tech, to UX, to UI, to pure business, to user research and so on. Understanding all of those assets in order to be able to make decisions, and to help grow or make your product successful. On the other hand, the importance of each of these sets of all of these disciplines, let's say, is different, depending on the stage in your product life cycle, for instance. In the growth phase, it's much more important to be growth savvy, marketing savvy, and then figure out ways to push your user base into that critical number, which then snowballs into future growth. When you're in more of a maturity phase, it's much more important to have a balance of all of these and figure out how to, first of all, keep your product interesting – and that implies a lot of user research, and looking at your data and analytics –, and then defining the technical solutions and so on. So that would be one aspect of product management.

The other one, which is extremely important, is that you're basically a nexus of inputs, which you then need to balance out, and come up with the best possible solution for all parties, and by that I mean you have many important stakeholders: the most important stakeholder you have are your users, basically, that's the main input source that you should have for your decisions. But they present one side of the story. That story that needs to be matched with what your organizational strategy is, and what the company's core competencies are, and then where it can be successful. So you need to take that input, the other input, and match it together into something, into an idea, which achieves both goals and creates the best possible intersection of those two needs. And then the other part is the technical aspect of it: "How do I take those inputs, and then create something which is actually possible, feasible, and also that makes sense, and is extensible. So, I think that's the other aspect of product management, which is also a fuzzy definition, but essentially drives the overall vision with all of these inputs. If you ignore any one of them, you're then bound to fail with your product, because if you ignore your users, you're not going to build what they want. If you only take what your company wants and you're going to build a Frankenstein of different things, which will not necessarily mesh with your market.

Those would be  the two aspects of it. Then additionally, also depending on the company, the standard situation with product management is that you're leading your product team without directly leading your product team, so you're not in direct reports or you're not directly managing people, the people which you work with. It's very important to lead through with the product vision as the main incentive, to get your team bought into the idea, to get your team excited about what you're building, and especially, depending on the people you're working with. What generally works very well is showing the impact of what everyone does. Everyone seeks meaningful work and giving that meaning through showing results and driving the communication of results is another way of basically keeping your team motivated. That's another part of product management which you also have to do at the same time. And there's probably a bunch more, but the last thing, is super important that I've grown to realize the importance of it much more throughout my career, is being – I hate the word, but – being the evangelist, basically, of your product and driving that same sort of buy-in which you do for your own team for the rest of the organization, and trying to get everyone under the same rallying banner, to work towards the same goal. It's very important to create this understanding of what your product is, why it's there, and what functions it fulfills inside of the organization in order to then best use the organization's resources in order to make that successful. It's key towards getting the majority of the growth which we've had without full buy-in, even though we could have developed our own app in a complete vacuum, we would have had a very tough time getting all of the different great types of content we have, launching in all of the markets we've launched and especially in getting all of the feedback and inputs from the parts of the organization which know their specific user base best. If you have your stakeholders over in Japan, which have intimate knowledge of their own market and what their own customers want and need, then that's something which should then transfer to you, and you can only achieve that by, one: them understanding what type of feedback is actually valuable, if someone really wants to integrate, to create tinder in e-commerce, then maybe they understand that this is not what the role we fulfill and that's not valuable feedback, so understanding, what type of feedback is good and lastly, knowing how to use these tools, so we've built a global app which is in the hands of… I don't even know a number but I'm even scared to know the amount of employees inside the value that which then has a direct impact on how the app is perceived, used and how successful it is, and it's important for all of those members to know how to use it properly, what its strengths are, what its weaknesses are, and to plan with those accordingly.

Attila Tóth: Well, that was a very complex walkthrough, and I think you already described that, basically, the product owner is the key responsible for growth.

Radu Lupu: Basically, yes. Even if you don't control it, even if you don't have direct control over the growth of I don't have my marketing budget under my red button which I can push whenever I need more and more traffic, it's our responsibility then to ensure that we get that from the organization. You're ultimately responsible for the success or failure of your product, if it's misunderstood – your fault, you need it to do a better job in doing that, if it's not bought into and you don't have the support you need– also your fault, because you didn't adapt to what was needed from organizational perspective or so on.

Attila Tóth: You said two interesting things here: success and failure. How do you define these? So what you consider a success and what you consider setback?

Radu Lupu: So, I think the nature of the job itself is – again – besides all of the things I talked about before, it's about iteration and adaptability. I think it's almost impossible to fail, if you keep this sort of fluid mindset in which you try things and then adapt based off of the feedback you get. You can build anything and make it successful if you listen to what's happening and don't just throw something out there, and then hope for the best. So you always need to monitor and to change your course depending on the feedback you get. Only failure I would see, it would be a failure to adapt and to change, and that's how I see most failures happening: ignoring the size of the stakeholders I was mentioning before, and ignoring those important elements which would drive the success of your product, placing responsibility on those outside forces for your lack of progress – is probably the only way which I can honestly see you fail. Success, though, means the exact opposite of that, so as long as you're constantly moving and constantly adapting, with enough iteration it's impossible for you to not get that success metric which you've looked at. And even the success metric, interestingly, can also be a form of iteration, maybe you're looking at the wrong thing, maybe you're measuring the wrong indicator and you're pushing towards the “doomed to fail” direction, which then means maybe our strengths, overall, lie in another area, we can definitely be better at customer lifetime value, rather than maximizing session-based conversion rate, for instance, right? So even that level of iteration would be a way of adapting.

Attila Tóth: It's very interesting, and I want to highlight it for the audience that actually what you said that failure is not about, let's say, failing with a feature, like you had an idea, it didn't work. Failure would be if you would stop working on that idea to improve it or to change it or to adapt it, so yeah, it's a great mindset. Coming a bit back to your current work, how do you see the augmented reality market tends to go in the next three years? So what are the trends and, if you can share, what is the percentage of the Adidas app audience of people who you can say are AR friendly?

Radu Lupu: So, AR is still in its infancy, it is still in the beginning of its lifecycle, it's becoming much more and more easier to use and much more prevalent. Definitely, the adoption of it will become greater with time and with better advances in technology. I think the main blocker to using it in your everyday life is the fact that it looks strange, you're looking through your phone at the world and it's perceived as a very strange thing to do, and people then avoid being judged from that perspective. We've had a lot of different trials where we could observe people trying AR out, and whenever there's people around it's a bit weird for them, and it's not as normal as you might see it, but at the same time, that could have been said for any piece of technology, which is now super successful. When it started out, it was strange, it was perceived as being weird, and if it was truly solving a problem, or it was truly well designed, then it would then gradually get more adoption.

Attila Tóth: That is the new normal. Yeah.  

Radu Lupu: That’s the new normal, right? Also like thinking about AirPods, when AirPods first were launched with – just thinking about the the most recent and very simple example – everyone was making fun of them as looking like toothbrush heads or whatever. Now everyone's wearing them, and they've become like a fashion statement as well, people don't take them off at any point of time, it's a bit crazy right? So, I think the way in which we will get to widespread adoption of AR, is having a company figuring out the perfect balance of the perfectly designed product which becomes something like an object of desire for others, like Apple is obviously working on an AR device which will, hopefully soon, figure out. There's a bunch of things which obviously will then determine the adoption of that device as well, and I think from a technological side we're not yet there in terms of all of the small pieces required to make something as sexy as a normal pair of glasses, which have a built-in, either a projector or a display and those are coming down to batteries which have not really evolved in consumer electronics enough in a very long time and now they're finally moving forwards again. Processing power, bandwidth, all of those things need to come together in order to get that star product off, in order to make it standardized. But now looking at the overall usage of AR, I see it being used a lot more than it used to, the impacts on different types of industries is – I can't even think about all of the use cases because it's just so broadly applicable.

Like just recently I was searching for a new apartment, and one of the viewing modes was viewing the 3D scan of that apartment, and then also seeing it in AR, and then being able to go through that apartment, through your apartment in a way which is a super simple application of it as well, because just placing an object in 3D is very easy to do, but thinking of that use case, and applying it is the goal. In fashion now we have AR try on, increasingly growing as a standard, let's say of seeing items. The adoption is getting also bigger and bigger, it highly depends how you present the information, especially when you're looking at the new feature, which is not widely understood, not everyone knows what AR is, you sometimes need to – and we've done a couple of tests with this, as well – you need to clearly show what they're going to get out of it in order for it to be a compelling use case. Just saying try the shoe on will be like, people won't really get it, maybe even ignore it, so at this stage it's very important to explain what the benefit of it is, and give sort of a preview of what AR would do for you. In the near future where you also have things like full body tracking, using AR together with artificial intelligence, with machine learning, is unlocking a lot of different avenues, a lot of different potential features. One thing we're looking at is overlaying a 3D version of yourself on top of you, and then allowing you to change apparel, for example. So you can basically switch your clothing within a virtual you, and that also is becoming easier and easier to do with built-in tools by Apple and Google.

Attila Tóth: Interesting. What else besides AR is in your watchlist, let's say, for 2021 and beyond?

Radu Lupu: So, besides AR, I think we're looking more towards machine learning and personalization, and that also is becoming much easier to use than before, especially establishing simpler models of understanding what type of product is successful in what situation due to what you use they're based off of what information. We're starting to use all of the info we're gathering, basically, instead of necessarily relying on third parties to figure out what our users sort of want with their own models, we were starting to look at that on our own, but that's very early days and the actual effort required to move into this direction is relatively big to do it in a consistent and successful way for the future. Normally, you would need a fully dedicated team to build such models and to constantly tweak them. They're their own product in and of themselves. Your own machine learning model is something you would be constantly tweaking and adding more data points to figuring out how its accuracy grows, and even though you can achieve some form of self-learning model with deep-learning or whatever, we're pretty far away from being able to do that consistently in non-highly technical companies. These would be the two key points which we're looking at, in terms of completely new technologies.

Attila Tóth: Thanks for sharing, and because a significant share of our listeners are key decision makers, I would like to get your thoughts on how would you advise key stakeholders to think of digital products in general?

Radu Lupu: I would make them think of them as tools which they can use to be successful with their own strategy, so thinking of the Adidas app as an avenue for growth and a tool which you can use to achieve whatever objective you have set up. It's a living and breathing tool at the same time, which you need to kind of be careful with using, and understanding its limitations so understanding that you wouldn't be using a hammer for painting a painting for instance, the same would apply to a product. I think that's also the way we're communicating it internally and that's what we've seen as being successful. Otherwise if a product is viewed as its own thing, in its own silo, with its own rule set and access to it is cut off then you're losing that momentum which you can build within your organization. And you're always going to have people who want to have something to do with your product, who will want to be able to use it in one form or way, and it's better to get those people bought into the proper way of doing it, rather than leaving room for an interpretation and then having uphill battles in terms of stabilizing the strategy of your product.

Attila Tóth: So basically, what you say is that, when somebody decides to start a digital product development, then they should take it, not as a separate idea besides the core business, but something that works inside the organization, and it's kind of like using the synergies of the organization – at least in the first phase – until it's mature enough to grow by itself. Is that what you want to point out here?

Radu Lupu: Even after you reach maturity you always will – again, I'm talking from my experience which obviously does not apply to every situation – you always need your organization to be there and to support, because even though you're building this thick product, it needs to be used, and it needs to be kept up to date with whatever type of information you might be forwarding, you might be showing to customers. For that, you always need synergies in place, arguably even more down the line than in the beginning, so yes, it's definitely about connecting the product to the overall organization and ensuring it's proper use.

Attila Tóth: I loved how you said it before: "not to live in a vacuum and not to build in the vacuum", I think that's a great way to think about product development. As we're getting closer to the end of this episode, I want to tackle a more, let's say, a visionary question. So, if you'd be able to travel ahead ten years in the future, what would you see? What would you be doing, how would the future look like?

Radu Lupu: That's a tough question but at the same time, I have a few visions of the future which I'm excited about, but the one which we've been also talking about here is the impact of augmented reality, as well as virtual reality on the world. I expect augmented reality to start catching up maybe, and being much more used in the next, let's say, five years. Not only in an e-commerce perspective, but from the way in which you interact with physical objects around you. My vision – the one I'm excited about is having a single device which basically adds an interface, and adds an extra layer of interactivity to your current version of reality. Imagine not needing television, not needing buttons on your stove or whatever, not needing a way to a physical representation of something, which can be replaced by an augmented reality version of it. Imagine having hand tracking already built into whatever augmented reality device you might have, you'd be able to interact with anything in the world then with this augmented reality there. Besides this, what we've been talking about from e-commerce before, think of the case in which you no longer need to go to an office in order to meet your co-workers. You just plug into your home office and you're standing near whomever you normally sit next to at your desk right there beside you, by achieving that you'd then also no longer need the same amount of office buildings, you wouldn't need the same amount of infrastructure to support this mass-movement of cars between point A and point B from 9 to 5, right? This sort of development could then impact the entire landscape of your life. I definitely see me working at least either – not necessarily as a full-time job – but definitely building more experiences in AR, which would go towards this goal, and it needs to be ready for when this actually does happen, and it is much more adopted in the market.

Attila Tóth: That's a great vision and very innovative thinking to reshape society with the benefits of AR and maybe have more trees and more green space in our crowded cities. It's a very inspiring vision, I love that. I think this vision, on one hand proves that you have a great feel of what the future may look like, but I think this is not something that you're just born with, I think you have developed this skill, and I want to know more about how do you continue to learn in order to stay on top of your game, especially in your role, which is – as we discussed– changing and adapting, so what are your personal strategies to learn more and keep up the pace with this ever-changing environment?

Radu Lupu: Basically I enjoy the process of building things, what I spend most of my free time on is playing with new technologies, learning them and using them. What I find most effective when wanting to learn something new is then setting an objective for yourself, either I want to build this type of app which does XYZ, or creating this sort of experience, or estimating what the price of Tesla shares will be in the next six years based off of stock market data, and then trying to learn the technology which you want to apply to then solve that specific problem. I enjoy doing that because I enjoy the result of these things then, I don't see it as a chore or as something I'm forced to do to keep up today, but I enjoy doing it for the end result. That then translates to any other objective you might set for yourself. Now, for instance, I'm learning Flutter, which is a new sort of framework to build native apps, and I'm building my own unknown app which just came out of a discussion with my wife, for example, because why not, and now I'm actually quite obsessed with that, so after the work day ends, I'll jump back in to figure out that last piece which doesn't quite fit yet.

Attila Tóth: Cool. When will it be up in the App Store?

Radu Lupu: I don't know, honestly. I've changed the design of it three times already, I tried to go full typography-based design which sounds very sexy, but when you try to apply it and you have no real skill in choosing font types, you quickly realize it's doomed. Then I started doing crazier gradient based interactive designs which sounded also very cool, which I also then killed, and now I'm honestly back to very standard App Store style design which looks the best and I'm very happy with it, so hopefully I won't change it again, and after that I'm 90% done with it, honestly. I just keep wanting to tweak it.

Attila Tóth: Sounds cool! Please share your link once it's live, I'm really curious about it. It's great background strategy, and I love how you have even these KPIs for learning, I think that's a very efficient way to learn, not just to spend time on different topics, but to have actual goals, so everybody who is listening to this: this is a very straightforward way to keep learning in a very changing environment. As we are reaching our last topic, I always like to conclude with one simple question: If you had just one piece of advice for people who want to create digital products, what would that be?

Radu Lupu: I think there's a bunch of different things that I would say, but if there's a key one, it's always starting from solving a user's needs. Whenever you devise a product, whenever you devise any project, you should start from having a clear understanding of what you're trying to solve. It's very easy to fall into the trap of doing hype-driven development, for instance, using that new technology, putting blockchain in the name just so that you can attach that and then sound innovative, but you're never going to be successful unless there's a real use case which is validated. I think that's the key to starting anything new, and then it highly depends on your environment along the way, how you grow that product, and how you adapt to change, and to the change in the need of your skill sets, let's say as well. Which then will make you successful.

Attila Tóth: Thank you! That's a great closing note. Thanks for sharing, and really looking forward to see what you do next.

Radu Lupu: Thank you, it was a great talk.