Nykaa's CTO on Fostering a Culture Open To Change
Good evening. Good evening Thanks for having me here. Looking forward to it. Absolute pleasure. So, Rajesh, I want to since this talk between you and me is about how do we get our teams to embrace technology? Before I talk about your technology, I want to learn a little bit about you. So the quintessential question, take us through your tech journey, please.
Yeah, sure. You know, I have been incredibly lucky to be part of very three distinct opportunities in my career. Obviously, like an on the first leg of my career, I started as a software engineer. I was writing software for Amazon dot
com in Seattle and there obviously like a lot of features in the app, a lot of large scale distributed systems back in supply chain at home management systems. And that was one leg of my journey that sold me very well later. And then there is another distinct opportunity that I got to set up in the development centers and also grow them. Like, you know, the growing part was
more of it insured. But the setting up of the India Dev Center much before even Amazon entered India, and that obviously gave me a very unique muscle and leg and a learning, especially when people are not familiar with what Amazon is. Right. And you have to build products for global consumers. How do you attract talent and, you know, how do you grow the teams? How do you build innovation without really using the products on a day to day basis? I think what the last and the one that I'm truly grateful is the opportunity to build for India considering was whether it was part of Jungle or the Amazon dot and or now Nykaa. It has been one wild ride. I wouldn't trade it for anything else.
The opportunity to transform how India buys and sells. That's an interesting point. You bring up, Rajesh, because you're building apps for 1.4 billion people and at the same time you want to make an app
immersive. How do you make something immersive for 1.4 billion people? That's a great question. And I especially like, you know, since I represent the technology organization, I firmly believe like, you know, there are many ways in which you can run a technology organization. You can obviously be like no business
model focused. You can be very technology driven. And these are all fantastic ways. There's nothing wrong in it. What I saw as well, especially at Nykaa, is to root our technology organization also in deep customer immersion. And I'll share three concrete examples how we do it. Before building anything at Nike, we do profound research. For example, personalization is a big
thing when you come to the app. We don't want to show the same discovery experience to every consumer. You want to personalize it based on who the consumer is. Even before going after the problem and talking about machine learning or collaborative filtering, the thing that we did was we invested a lot of time in understanding personas of users who showed beauty, and we came up with interesting learnings. And the other nice thing is we have fantastic relationships with our brands and they share their consumer research as well, and that helps us identify like key personas, like, you know, someone who would want to be bold with their makeup, the choice of colors, the kind of products they use. What's a someone who would want to be a little more demure and start using that as a knowledge? Then you're going to look at the digital footprint of the people, how they buy, and then map it to the personas and then use that knowledge to personalize their experience.
What, someone buying similar books, right? Similar products. I think that's where magic happens. So research. The other thing that has also served us well is what we call as like, you know, walking in customers shoes, for example.
Most of us do what we call as liking or warehouse towards our visits. I personally went to warehouse. I do customer orders myself. The last time I did, I realized to pick three customer orders I had to walk a lot of distance and I was like, How can I use technology to optimize the big part of the warehouse I saw said, How can they stack the most popular products in their vicinity so that I don't have to walk a whole lot of distance. So walking in customer shoes is another thing that sells as well. There is another third interesting thing that we do. We pay a lot of attention to customer anecdotes and it works really well. Because customers inherently, they have
a very divided discontent in them. Even if things are good, they keep saying, I want things faster, I want things cheaper. And that pushes the envelope, right? In terms of innovation. We listen to customer calls. Recently we started doing something like, you know, for most of our critical review meetings, we start with a customer anecdote, a pain point. These are some of the things that are helping us build a lot more customer empathy, and we use that to delve deeper into the problems and then elevated with technology.
So I want to pick up from something that I'm hearing constantly talk about the customer, walk in their shoes, develop empathy. These are all very human emotions, something that's quite the antithesis of an event that's called intelligent automation. It's something you also alluded to understand the problem before you get into the solution. And yet one of the operating tenets that you have at Nykaa is when it comes to technology, be fearless.
Yeah, but now, with all due respect, it's easy for you as part of a tech team who has your finger on the button to see. Be fearless. But technology. Yeah. Speaking very transparently, my tech team sends me an email. I get scared to open it.
What? What? What changes are going to be announced in that? How do I have to change my life to accommodate for this? How do you how do you marry the need to be fearless with technology, with everything else that you spoke about developing empathy, not just for your customer, but also for your teams who need to embrace that change? You know, that's a very great question, Janet. So I think at least the way we look at building that fearlessness with technology is in three ways, right? I mean, in the previous panel, we had extremely good insight on how Geneva is like, you know, or really like transforming things. And if you look around, there are way too many external trends that that keep happening. Before that, we had like blockchain and then like in the whole Mecha and Web3. And
there are two ways in which companies react to these external trends. Right. And sometimes when you are in a certain phase, you also might not necessarily have a lot of tech muscle or sometimes when you are too big, you might also be resistant to noticing an embrace in some of these trends. And the problem with that is you risk becoming obsolete. And I'm not saying you should jump on
every trend, but some of the ones that are really they are unmistakable. You will never miss them like we are talking about genitalia. And one of the things we need to do is the sooner we embrace them versus resisting, fight them. I think you have a great chance of making that trend as a tailwind than a headwind. And on the same thing, right, like in more recently, we have done a hackathon in partnership with Microsoft on the specific theme generating with at least five super interesting ideas and you don't need massive theme.
We had like a handful of family engineers and they spent like couple of days and some of the ideas were almost like ready to at least try it as a pilot and put it in the hands of the customers. So that's what I mean by being technically fearless. One aspect of it embracing trends. What's this? Fighting them? Then there is something that I truly believe in. Like, you know, as a company,
when you start growing bigger, the scale of your experiments should also naturally increase. And which means with that, the rate also will increase because experiments by nature have a slight probability that they might fail. And how comfortable are you right, with experiments and failures? Almost to the extent that can you even celebrate some of your failures.
So that's that's another thing like and on that we consciously later not try to like at all watch out for our experiments and even sometimes recognize and celebrate the failures and the learnings. Oh, and there's one last thing I would want to add is the commitment to engineering excellence. It's not easy. Like, you know, when I say engineering excellence, there are a lot of jazzy things you want to build on the hybrid new features, but everything that happens behind the curtains to make your app resilient so that it doesn't crash when you open it, it doesn't take a lot of time to load or making it super secure so that your information doesn't get leaked. All these things, they take a lot of time and in the end, like customers might not even immediately notice it. But your commitment to creating a super
strong, resilient, secured performing app, these are some of the traits I believe as engineering leaders we need to champion. Rajesh, you spoke about being an engineering leader. Yeah, but I'm sure it's different for different organizations.
I've been able to spoke about it during his panel that for a company that's a cloud native company, it's probably easier to adopt. Such a change was as a Kotak Mahindra, for instance, which is a regulated entity. They have to explain certain things to the regulator. You've spent the initial part of your career at a 0 to 1 company, generally Amazon, and now you're at Nykaa, which is going on the next stage from 1 to 10. So other tools that you use to manage such change different at the life stage of an organization. I think that's a great question.
I think something I think that really. I think it's a measure of your leadership is how well can you navigate change and how well you can carry like, you know, not just your teams but also your stakeholders. I don't know if I have a magic silver bullet.
Jeanette I'll start with this powerful quote from Greek mathematician Archimedes and a physicist, of course. I think we all know this code, right? He said, Give me a lever that is long enough and a place and the fulcrum. Like, no way I can put that long pole and a place to stand and I'll move the earth. And that's a very powerful
cork. So a lever has the ability, intrinsic ability to move something at a distance. And I use that analogy to kind of explain what you're just alluding to. Right. Like, luckily, there have been some organizational levers that have served me well. I'm sure there are other ways, like, you
know, in bringing about some of these change. It's very interesting. If you look at these levers like the Archimedes analogy, there are some really long levers they can actually make big changes. And the challenge with exercising long legacies, you get a lot of resistance in the organization.
And then there are these medium sized levers and there are these small size levers I specifically like, know have this toolkit like an all in my toolkit. The long levers are typically aligning at a mental model level, like, you know, getting organization, buying through decision making frameworks, operating techniques. All these are the longest levers. So the harder the changes, I fall back on those levels. But those are the levers that get the most resistance as well. So you have to invest a lot of time. And then the medium size levers are typically around goals, incentives, rewards, again, very useful.
They might not move things as much, but again, supercritical and then slightly lower would come like, you know, what we call typically things around like resources, headcount, budget. And the key is to flex all these levers in your nature. Let me bring it to life through a small example played very recently.
One of the things we try to do at Nike as a big change was to bring about more differentiation in our talent collaboration. And what we wanted to do was like to identify a small set of people whom we would call as the top two years. And again, like, you know, try to identify the bottom performers and create that separation. And we knew that was important because it's very natural. Like, you know, you you as you grow, you have some bottom performers that you need to eventually invest and turn around things similarly on the top performers and probably like, you know, invest more on them and. We spent a lot of time on mental model alignment, like in other ways that we wouldn't have been able to achieve it. Like, you know, we had to do multiple
discussions saying why this differentiation is important, how we have to fast track a few people. Eventually there is only going to be a finite amount of money how you want to give it away. And that helped us, right? Later on then we set goals saying like, if you have a team of a certain size, you have to make them all achieve that kind of differentiation. But that I think brings out like in a whole, we were playing at multiple levels, using all this legwork to drive that change. I used these levers typically to try to win the organizational change, but Rajesh, as a CTO, you only have limited visibility on the organization. So while you have the levers and while
you probably want to take a decision about which of the three levels to use, how do you as a CTO with this limited vision nor which levels to move and which opportunities? That's a great question. I think it depends on the type of the challenge. If the challenges I typically look at challenges as leadership challenges and more like a managerial challenge, leadership challenges typically are very hard. They typically involve mindset changes
that typically involve cultural changes. And invariably you will have to fall back on this larger level. You will have to strive for a lot of alignment, mental model levels, principles. You debate a lot, you get resistance, you go back again, you try to understand other counterpoint or communicate. And then there are other challenges.
For example, let's say for our new hires. We traditionally give a backpack grade for their laptops. Tomorrow we decide to change that to something else, like a laptop sleeve. It's not a great leadership challenge. I don't get a lot of resistance, and
those changes are handled very differently versus maybe changing our policy on return to work remote versus hybrid. So depending on the type of the challenge, we decide which lever to use, maybe a combination of them. Drilling down into the subject matter. Rajesh. A lot of the terms are used interchangeably, right? Automation, artificial intelligence, machine learning.
How do you as a leader think about these three terms? How do you and more importantly, how do you train your team to think about these three terms? And what do you want to implement in your organizations? Yeah. Oh. So at the crux of it is data to begin with, and you'll have to obviously invest a lot in having like clean data and democratizing the data to make a lot of decisions.
And then at least at Nykaa, we have a three pronged or a three tiered approach. I'll briefly talk about like know, we obviously have a large analytics team and this is the team that is actually looking at data and doing a lot of deep product analytics and identifying the right signals. For example, if a customer comes to the website or the app is here, premium customer, things like, you know, does this customer have a propensity to reject or return items quite often. And that's what the analytics team is doing.
And then we obviously have our data sciences machine learning and data sciences team, and this is the team like and all that is more often than not building models that directly like no interface with the customers, whether it is in terms of personalizing your search, showing you like the right kind of product recommendations going into the supply chain aspects like not trying to figure out which products to stock up and which way. And that's what the machine learning is doing right now. The third function that we are, it's a small early stage like, no, no, we are getting into generating we as well. I just talked about a couple of experiments that we are doing.
So depending on the kind of the problem, it is tackled by one of these three, if I may use the word parts in our organization. And yeah, automation invariably like, you know, it runs through our company. We have gone through an evolution very early stages when tech is a premium right there, bandwidth, and you have to build a lot of features, so you'll save a lot of tech bandwidth. So you try to solve some of these challenges using people, right? We typically use the phrase throw people at the problem, and then you slowly get the courage and say, okay, let me prioritize that one task, which is very laborious, which is very rough, drawn, then slowly volunteer like, No, let me graduate from picking tasks to identifying an entire workflow that can be automated. Rajesh With an eye on the clock.
Yeah, I'm going to I'm going to take on the wrath of the organization because I want to push this conversation a little bit more. I feel we were hitting the crux of it. Right. Start by throwing people at a problem. Yeah.
And then train the machine. Train the system? Yeah. Which has led to this fear today that inevitably people will lose jobs when it comes to air travel. And this is probably the source of a lot of resistance against it. Yeah. Is this something you're facing as well, and how do you tackle it? I haven't faced such a resistance in my organization because for us, fundamentally, the people today who are doing some of the manual work are very talented people as well. And what we have always realized in our
experiences, I think we leverage their expertise to qualify the rules and then their expertise to take those rules and then evolve them into models as well. So in some sense, what is happening is it is giving them a lot more bandwidth to help us better our heuristics, our role based engines, and also to help us train the models. Let me give you a very quick example. Right. Like in the past, whenever customer orders had a certain criteria that could potentially make them fraudulent, maybe like an already higher order value or from a certain PIN code or whatever, we flagged them and before shipping them like, no, we would have a team that would go and call the customer, do a lot of verification to make sure like, you know, this is a genuine order and we should be shipping it.
And over a period of time, obviously, like, you know, we built rules to automated. You also build models. But you know, the best part of it is all these people who have been calling the customers, figuring out whether it's genuine or not, how big so much knowledge that they are the ones who really helped us build these rules and heuristics and models. So at least for us, it has been extremely symbiotic and it is allowing them to do more higher order work. So for us, we haven't been facing such a concern adding value to the value chain.
Thank you very much, Rajesh. Very, very insightful conversation. Thank you. Handing over now to Rookie. Yeah. And we carry on with the discussion on automation through the evening. Thank you so much. Thank you.
2023-08-26 14:49
