Return to site

52 FromNaive to Unicorns: Sharad Gupta's Journey in Building Digital Empires

· AI,podcast


broken image


Podcast with Sharad Part 1



Summary Part 1 with Sharad:

In this podcast, Andrew Liew interviews Sharad Gupta, adigital business growth and product leader with a remarkable background in
founding unicorns like Lazada, Zalora, and Foodpanda. With over 20 years of
experience in fintech and e-commerce, Sharad shares his journey from naive
thinking to building successful ventures. Discover how his curiosity,
open-mindedness, and willingness to explore new opportunities led him to
exponential revenue growth and pioneering work in AI and data science. Tune in
to gain insights into the challenges of startups, the importance of
product-market fit, and the power of perseverance in the ever-changing digital

In this podcast, Andrew Liew interviews Sharad Gupta, a digitalbusiness growth and product leader with expertise in AI, machine learning, data
science, and risk management. Sharad has been involved in founding successful
companies like Lazada, Zalora, Foodpanda, and Tokitaki. He shares his journey
and insights from his various ventures.

Sharad discusses his transition from a corporate career to thestartup world. He initially started a consulting firm, hoping to offer
McKinsey-like services to tier two companies. However, he learned that building
a reputable brand takes time and recognized the importance of brand value and

Moving on to e-commerce ventures like Foodpanda, Lazada, andZalora, Sharad emphasizes the significance of product-market fit. He highlights
the need to focus on customer feedback and fulfilling genuine needs, even when
the initial numbers may seem small. The trajectory of growth and the validation
of solving a significant problem are crucial indicators of being on the right

Sharad also mentions his experience with Tokitaki, a deep techcompany in the financial industry. He discusses the challenges of entering an
emerging market like AI and machine learning for risk and compliance in banks.
Despite the lack of precedence, he relied on customer feedback and market
validation to navigate the path towards success.

Overall, Sharad's journey emphasizes the importance ofcuriosity, open-mindedness, and a willingness to explore new opportunities in
the business world.

[00:00:00] Andrew Liew: All right. Thank you, everybody.Welcome to the show. Welcome to the AI of Mankind. Today, my guest is Sharad.
Sharad is a digital business growth and product leader with specialization
around artificial intelligence slash machine learning, product management, data
science application, financial risk management in banks.

[00:00:19] Andrew Liew: He has actually over 20 years ofexperience in building fintech e commerce ventures and digital products and has
drove exponential revenue growth with strategy, digital innovation, data
science, risk management, and digital marketing initiative. One of the
interesting thing that I know about Sherad was that he has always been a
founding team of unicorns like Lazada, Zalora, Foodpanda, and the recent deep
tech company, Tokitaki.

[00:00:47] Andrew Liew: He manages 15 million dollarsprofit and loss and almost 100 plus people across Foodpanda and Kuni. At
Tukituki, he has built almost 20 plus people data science team, [00:01:00] developed native products aroundartificial intelligence, anti money laundering solutions, risk products, and
has been advising banks around AI in risk and compliance.

[00:01:12] Andrew Liew: Other Participant achievementsinclude building and launching the KPMG new digital fintech venture, achieving
20% month on month growth during the foodpanda early days, building the data
science functions for Zalora, Lazada, across Southeast Asia, achieving 0 15
million annual revenue, and build 100 plus people b2b b2c e commerce and retail
business across asia pacific around kuni he was also previously in mackenzie
supporting strategy and risk management consulting for several banks last but
not least He's also listed in the Stanford Who's list and an alumni of
Institute of Technology, IIT.

[00:01:54] Andrew Liew: Okay, so that was a handful ofadventures. So let us start by asking [00:02:00]Sharad so Sharad, the first question I want to ask you is, tell me How do you
get here from the day that you did your first

[00:02:07] Sharad Gupta: startup? Serendipity. But yeah Ithink that's a chance. But I think it's always about being curious and being
open minded about new opportunities.

[00:02:19] Sharad Gupta: And always exploring and try todo a new thing. So that's how I've reached out. In my early days in my first 7
10 years of my career, I was an out and out corporate guy. Worked in GE, worked
in McKinsey, and post that, I felt like after working in McKinsey, we were
naive enough to think that we can build something like McKinsey for Tier 2
institutes, because not everybody can afford.

[00:02:46] Sharad Gupta: I think over there, the luck offof being just a knife, I think innocent on thinking that what really the effort
maybe that was my starting point into the start of journey and [00:03:00] then getting jumping into Foodpanda andRocket Group of Companies, Lazada, Zalora, was more around my wish to explore a
new area around e commerce and then sold my company and joined Foodpanda here
in Foodpanda Lazada Zalora, at that time it was a combined entity.

[00:03:19] Sharad Gupta: And did something which I didn'tdo earlier. I was a, I was more of a risk strategy guy. Then suddenly was doing
digital marketing. And it was a phenomenal learning. So then, so taking
learning from there, being just growing multiple, multifold in terms of
revenue, then moved on to do CUNY and another venture.

[00:03:41] Sharad Gupta: That was a corporate venture.It's a large Swiss company. And then in, by accident, I met Abhishek, who's a
founder of Tukitaki. And we started working on an idea. And really out of
Singapore, we will, we won the first Fintech Prize. And then there was no
looking back. And [00:04:00] now we are ahundred plus people come.

[00:04:02] Sharad Gupta: And so my learning has been justbe curious, be open minded and always start exploring. You never know what
might come to your view.

[00:04:11] Andrew Liew: Wow. That was an interestinghandful of stories. Let me go back to the early chapter so that the audience
out there can really learn. You mentioned about that naive thinking.

[00:04:20] Andrew Liew: What was that discovery thatallows you to think about that it was naive? When you were in the early days,
you started out your Consulting to be a second tier consultant. What was the
original conjecture that you were having in your head? And then it turns out
that, ah, this is naive thinking.

[00:04:37] Sharad Gupta: So what we noticed that thedemand for structured McKinsey like consulting was huge and McKinsey was
always, and even until today, continue to have more supply than what they can
cater. And what I noticed that it is not, McKinsey is nothing but there are
brilliant management practices.

[00:04:56] Sharad Gupta: use, how they consult issomething. But [00:05:00] that sort of qualityis not available and practice is available for other consulting companies. So I
thought, we thought we are McKinsey. We were three guys. We are ex McKinsey.
Why don't we start doing that and apply to, to tier two firms? And we got
decent success but the learning.

[00:05:17] Sharad Gupta: The, so that was a theconjecture. And. I think the learning was the other part which we missed was
around in McKinsey, a lot of time people buy McKinsey product and they take
their advice because it's McKinsey. It's a brand. So that's how it started and
that was the conjecture.

[00:05:36] Andrew Liew: Ah, so you were thinking that,okay, I was previously from McKinsey, so I start out my own brand. I They
should buy from me because I'm is equally as good as my previous company at
McKinsey because I'm still me, right? I'm still sure But then you realize that
when you guys were going to market the tier two companies They don't [00:06:00] see you guys as the McKinsey guy.

[00:06:01] Andrew Liew: They see you guys as individuals.Is that the surprising thinking

[00:06:06] Sharad Gupta: that you realize? Not anindividual. We had the 50 people company. And we could hire equally good
people. But I think the brand and the kind of work which you have done over 100
years is is a great value and great level.

[00:06:22] Sharad Gupta: Which is the investment you needto do over a period of time. Building reputation, building, giving the
processes and the knowledge is easy. But building, building reputation over a
hundred years is something is not easy. And you need, if not a hundred years,
but few years to build that.

[00:06:42] Sharad Gupta: It's very different in productversus services.

[00:06:47] Andrew Liew: Ah, okay. So now coming back tothe second part I wanted to ask you in that journey was after you left, like
you said, consulting services like the ex McKenzie, the ex company that you
tried, and then you move into almost like a product centric kind of [00:07:00] company zada, Zorra Fu Panda. And so whatwas the thing that you have. I've learned that you realize that what was the
original conjecture when you go into these companies and what was the thing
that you learned?

[00:07:14] Sharad Gupta: Yeah, I think so. Originalconjecture was that, hey, I've done too much of consulting and let's get the
real stuff.

[00:07:21] Sharad Gupta: Let's do stuff. Let's buildstuff. So that's what was the whole idea. And frankly, when we started, so the
learning even though I work for not for a long period of time, but the learning
was just to give example, when we started foodpanda, we started online business
without having online payment.

[00:07:40] Sharad Gupta: So we started cash on delivery.So food were delivered. Orders came online, but payment happened offline. So my
learning was, A, your experience is overrated. In digital world experience
comes by doing, failing, relearning, not just by having [00:08:00] 20 years of past experience. Because there was nobodyto build the delivery platform.

[00:08:04] Sharad Gupta: There was no experience, pastexperience. So you've got to earn that experience, A. And B don't underestimate
that things look small today. And people laugh. I used to laugh at some time,
like you are getting 100 orders a day in Singapore. And why you... Think that
it's a good business, viable business.

[00:08:21] Sharad Gupta: And frankly, that, that was avalid question, analytically beautiful question. But we now know that Foodpanda
must be doing more than 50, 000 orders a day. So no Excel sheet, no spreadsheet
can predict that from 100 orders a day, you can reach up to 50, 000 orders a
day. You can see the compounding growth for a period.

[00:08:44] Sharad Gupta: So as long as there is a productmarket fit, You have a conviction and you have, again, innocence in you to
believe that what you believe is true things fall in place. And that's what
exactly happened. So we started we didn't worry about [00:09:00]got worried about that. Hey, let's wait for six more months to get an online
payment up.

[00:09:04] Sharad Gupta: We found ways of doing. Fast andwith the limited means and, but get going.

[00:09:11] Andrew Liew: So I'm just curious, how do youknow whether you're on the right track when, because now you're in a new
frontier, you are like your experience probably give you a way of structured or
analytical thinking, but you are like almost so blue ocean or so like in the
dark and you're trying to figure out, Hey, am I in the right direction? The
right track? How do figure that out? I think nothing speaks bigger than and
better than the product market fit. If you are seeing growth over period of
time, not daily growth, it's not a stock market.

[00:09:41] Sharad Gupta: And even stock market doesn'tgrow every day. So you, that validation of customer feedback, if you are really
helping somebody, if you are able to do some Somebody's sitting in the wrong
and saying that, Hey, thank God I don't have to spend one hour going for my
food. Then you know that there is a [00:10:00]genuine need or a product gap and you are fulfilling the genuine need.

[00:10:04] Sharad Gupta: The rest of the things willfollow. As I said, no unit economics was in place. You can't even think about,
and even the business plan looks daunting. The kind of. The break even would be
around 10, 000 orders and you are sitting at 100 orders a day. So technically
you must be foolish enough.

[00:10:23] Sharad Gupta: But I think it is a, we knowthat this kind of a market, this kind of a model works because it was working
in Germany. It was working already. Just see it has proven in US. So we know
that product market fit is there. It's a matter of time and just keep going.
No, it could be a different ball game and Maybe my experience in Tukitaki would
tell when you don't have a precedence and a product market fit known to you,
because we know in this case, we knew that product market fit exists.

[00:10:54] Sharad Gupta: It's just a matter of time. Andwe were seeing good sign except that unit economics were not in [00:11:00] play. Yeah. Hold

[00:11:01] Andrew Liew: on to that. So because you saythat yes, you saw product market fit. Because all these other markets, they
were there. It was a precedent and it was a matter of time. But the question,
we all know that in a startup cash burns, right?

[00:11:13] Andrew Liew: And how that trajectory, like,how are you so certain that, like you said, you will eventually reach to 50,
000 orders a day from 100. orders a day when on paper or on excel you guys were
projecting 10, 000 orders to break even. So my question is that time period and
that trajectory what are the tell tale signs that got you thinking we are on
the right track?

[00:11:37] Sharad Gupta: As I said I think are you thefirst question is always are you really solving a huge problem here or not and
if you're solving That's the first part. Then if you talk to 10 customers, and
customers say, yes, you solved, thank you, you know then you're on the right
track. But, if you go and say that, hey, I'm not going to use it, there's no
repeat, there's no benefit, why [00:12:00]should I go, I eat at the next door food court?

[00:12:03] Sharad Gupta: Why should I just order online?It looks weird to me. We had that example. But we knew that the niche is always
about people who are... running short of time. They have family and they really
can't go. So those sort of myths and those sort of segments were giving us a
right feedback. Our restaurants were very happy because that was opening a new
channel for them.

[00:12:28] Sharad Gupta: And so I would say these werethe delta sign apart from Yes. Market validation and proof from the other
market. Ah,

[00:12:37] Andrew Liew: okay. Now coming back, youmentioned about to Kentucky, so you, when that was a B two C a business to
consumer. And then eventually you move to trying b2b business to business and
somehow it enterprise sales.

[00:12:51] Andrew Liew: When there was no precedent, therewas back then there was a lot of nuances. emerging needs for artificial
intelligence or machine learning application. [00:13:00]Tell us about that experience about putting your feet into weathering, whether
you're on rocks or wet water that you're on the right track, okay. So I think
it is a tricky question to be frank. There is always a belief part which comes
in play. But over there, I would say our understanding was always follow the
customer. Like when we started. working with banks like SCB or UOB or many
other large bank. We saw clearly that the current systems are not doing their

[00:13:32] Sharad Gupta: Current financial crime systems,fraud systems are not doing their job, meaning thereby on simplest level, there
is, there are too many false alert. People are spending. In, in checking those
errors and alerts, which is just complete waste of time. So the problem
statement was very clear. Customer was crying and they still cry because AI
adoption is still.

[00:13:58] Sharad Gupta: In early stage. [00:14:00] But, I think the, We there also what wewere very certain, yes, it's a sizable problem to solve. They have, there is
enough pain in the customer to demand. Now whether they will pay for that pain
and that's the trickier part in B2B and especially with large banks and how much
time they will take to decide.

[00:14:24] Sharad Gupta: And when the cash flows willcome is a second order problem, but it is very important and a lot of time in,
especially in Asia, companies get they are not able to survive because the
velocity, sales velocity in B2B, especially for large bank, large enterprise
sales is huge, but we were certain that the future doesn't belong to Pure play
AI companies like, for example, those horizontal platform, those who serve
general purpose machine learning, the future belong to native AI products,
which are purpose built to [00:15:00] solve aparticular problem, not a general.

[00:15:02] Sharad Gupta: And that has its own challenges,but we were very clear that the future belonged to vertical AI stack, not the
horizontal AI stack. And because when we serve our client, they were clear
that, hey, I have a large data science team. You sell me a generally solution,
I can get it built through them. It's not gonna work.

[00:15:22] Sharad Gupta: So give me a solution to my problem.My problem is I'm, there are only 0.1% of the back people who are protester and
I need to catch that. And current tool doesn't provide me a efficient way of
doing it. And in the process I spending I employ army of people in doing it.
Can your AI do that?

[00:15:43] Sharad Gupta: If you can solve my, thisapproval. vertical problem, then your AI is useful, otherwise it is not. Yeah,

[00:15:51] Andrew Liew: about that, because I'm alsothinking about like you mentioned about horizontal platform and vertical, and
so the interesting question about vertical is also [00:16:00]the guy, clients will also be thinking, okay why would I want to buy your
vertical stuff when I have my own in house team?

[00:16:08] Andrew Liew: It's like always the classicquestion of buy versus build. And how do you get around that I think the first
question is vertical versus horizontal is nothing but a specialization. If you
have a pain, if you have a pain in in, in let's suppose your spinal cord,
you're gonna go to specialists.

[00:16:28] Sharad Gupta: You are not going to go to GP tosay that, Hey, I have a pain, give me Cro Brufane. And even the generalist will
be very cagey about giving a solution over here. And that's exactly happened.
When you are... Dealing with horizontal platform, they all can say, I can build
you a beautiful stack machine learning model, but that's not the problem.

[00:16:53] Sharad Gupta: Problem is business. And how doyou solve a business problem? If you have a generalist hat, you don't have [00:17:00] a specialization in the function whichyou're dealing with. So that's where our decision and in none of our client
would say, and I've seen it, that when we present, and not only us as Tukutaki,
but all are in, in a community who so ever present the solution saying that,
hey, for this is...

[00:17:19] Sharad Gupta: This is the solution. And by theway, this is how AI is helping in it. But AI is not the main driver. Main
driver is how you're solving the problem. While if you go, you take the internal
approach, data science approach. The whole idea is that I can build beautiful
model. That's one problem. The second part is how do you productionize it at

[00:17:41] Sharad Gupta: for solving a particularproblem. Now the beast or the people whom you are dealing here are a financial
crime analyst. What is their specific problem and how do you solve that problem
with your software? It's not a machine learning problem. It's. more around what
is their day to day [00:18:00] pain area and isyour software solving those problems through machine learning?

[00:18:04] Sharad Gupta: And so it's almost saying thatokay, curry makes you or or the spices make. Your dish are tasty, but your
spices cannot be a dish. The dish has to, has so many other ingredients.

[00:18:21] Andrew Liew: Ah, so in other words, likeyou're actually focusing on what... That specific unit business leaders, what
was the specific problem that you can solve way better than they can do it with
their own generalized team?

[00:18:35] Andrew Liew: It's a very good question..