 Good morning. Let's see if this is working. Can you hear me guys? I have five amazing women with me this morning. My name is Yana. I will be the moderator this morning to talk about, or is it still morning? Actually, yes it is. We will be talking about data-driven innovation across different industries. So we assembled these ladies here who are working in different industries. We have Maggie, who is a head of data at Citi. We have Celine, who is a head of data science and innovation at AXA. Then we have Judeon, who is... I'm now leading the data analysis team in Shopee. And then we have Jen, who is working at Facebook, and she is... Jen, where do I have you? Partner at Marketing Science. And then we have Sigrid, who is a lead data scientist at Stock Exchange. What we will do today, when I hear the subject when Anne came to me, and she said, let's talk about data-driven innovation. And I said, oh, this is great. Data-driven is kind of everywhere. So let's share the experiences that you ladies are using the data in your day-to-day job. And maybe we should start about the data-related kind of things and procedures you run in your profession and in your job. So maybe introduce yourself a little bit more how you landed in this particular position and what it means for you data-driven and data. Maggie? Hello, everyone. So I work at Citibank for the Consumer Bank, and I look after all data technology across all our APAC and Amir countries. So in that technology, I basically provide all the technology that data scientists need, all the tools that the data scientists actually need. I look after... So I own the data lake, our real-time event hub. I build the real-time use cases that says what's the best next action for our customers at this point in time, what are the best offers to give our customers at this point in time, our risk analytics, all of the real-time and old times. I own both the old and the new technology. And how I got into data... Because Yana always wants me to talk about how I got into data. This is very interesting because Maggie used to be ahead of infrastructure. So how do you go from infrastructure to data? That's an interesting transition. Right. So basically I only started my job in July this year. Prior to that, I've always been managing infrastructure for the same consumer business. Last year sometime I decided actually I was really, really interested in data, and that was the area and the field that I wanted to get into. So the first thing I did was actually make sure I was relevant enough to say I want to get into that field. So I got involved in lots of external things including trying to get to know Yana and kneel in the back, most data professionals know. I ran without even being in data. With the Singapore Computer Society I ran a woman in data analytics conference where I then met lots more data professionals. I went and read things up and I ran and registered for a course in NUS. Having done all of that and feel like actually I can reasonably talk about why I'm passionate in getting into data, then the next thing I needed to do was convince my company that I was the best person for the job. This is interesting as well, right? How do you convince your company coming doing such a leap? Yeah, so I first had to actually convince my business stakeholders, in fact the guy who needs the business analytics, and he had only just joined the company a year ago coming from eBay. So I knew actually one of his frustrations joining Citibank was, he didn't know how to execute at the same speed as eBay at Citibank. So I actually went and put myself in front of him. I then told him that I know exactly all the issues you are having and I am the very person that can actually help you solve them. I said, I actually said to him, I'm not going to pretend I'm a data expert because I'm not, but because you are a data expert, why do you want me to be a data expert? That's my advice. Then I had to do the influencing because I had been managing infrastructure for the consumer business four years prior to that. I actually went to talk to all my previous customers who I had built a very good brand and reputation with and basically got their sponsorship, telling them that the same way that I succeeded in infrastructure, I was succeeded in data. Then my new stakeholder went and checked with his team peers and they all basically said, yes, maybe it's your person. That's fantastic. Thank you so much for sharing that. Now we have Celine and Celine as well as coming from slightly different background. She's coming from marketing and we know that marketing, it's all about data today and technologies. So it's not such a kind of strange change, but it's interesting coming from automotive into insurance industry. So what was this transition for you about? So my focus is, so we're running some of the same use cases as what you said internally, but for me data and I think the topic is data driven innovation across industry. There's the innovation part and there is the across industry that is very important. Today the business model are changing. So I come from the automotive industry and my previous job was, oh my God, autonomous vehicle is coming. How will we survive? And this is where all the car makers are moving into mobility. So a new type of business model, you know, with the arrival of Uber or Grab in the market. How could those guys be born? They were born because they had access to your real-time geolocalization data. And Uber created Uber and Grab created Grab. It's not Apple or Samsung or whoever. It's not the hardware manufacturer who is actually creating all the services around it. And so they were able to create Grab because Apple and Samsung and Google were sharing the geolocalization data, instant geolocalization data. So for me, I've been working on new business model and how the word is changing with the arrival of the sharing economy because you do have access to people data. And you are able now through platform to propose new type of services. We are going to an ownership word to a usage word. We don't want to own anything, especially the young generation. I used to work in car manufacturing for like eight years in an R&D center. I never owned a car. Really. I don't see the point. It's an object that is like very expensive, very like you need to maintain, you need to pay parking, you need to pay fuel. And then you have the chance of driving in the traffic when it's raining. Did you share this opinion while you were still working? Yes, I did. This is why I was in charge of mobility and new business model because I'm sure that a lot of people feel the same way I do. I do prefer to use a Grab or a DD in China for that matter rather than owning my home car and having the pleasure to spend time in traffic, wasting time, my hands on the wheel. That's why I was really good. Someone enjoys it. But let's talk about your transition from automotive and obviously R&D and very innovative industry from this perspective. Going to quite traditional insurance business. Very traditional insurance business. I think it's about curiosity because there was this thinking that my job was in monetizing data, so creating new business model with partner. I was working with DD in China and Alibaba Tongxun in sharing data to create new type of services for the car. And the industry thinks that 80% of your revenue of monetizing car data will come from insurance. They say, yeah, pay as you drive, pay how you drive. And I was not seeing the business model going on. I was like, how will it work? And all the car manufacturer will put all the data into big data dump and all the insurer will come and then pick it up. How is it working? So I wanted to understand insurance. What's about insurance? And I think that no one understands insurance, to be honest. There is lots of statistics and mathematics behind the insurance model. Yes, but today using very non-behavioral type of data. That's the thing. Actories have been using data and statistics for ages already but are not able today to use real-time instant data in order to do their pricing. And so I wanted to understand insurance and how it's working and what would be the value of the car data for an insurance company. And then I go into an insurance company and I'm like, okay, everybody's wrong. Everybody thinks that the insurance will be paid something but that's not how the word of tomorrow is going to work. And so now we've been focusing also on creating new type of product that are more insurance usage-based product but with our partners. For example, we did launch a product with Grab in 2016 called Pay As You Grab that was targeting the part-time driver that wanted to part-time drive for Grab and had an insurance, commercial insurance that they were only paying to the kilometer. So how did we do it? We partner with Grab. We exchange our data with Grab. And thanks to Grab data, we're able to do a product, a usage-based pricing dedicated for the driver without having to install a box, an app, or whatever. We're just already using the existing ecosystem and our partner in order to create new value proposition for the customer. Okay, let's stay in our finance industry and let's go to C-grade because so we do it a little bit more in a structured way. So you work at Stock Exchange and you've been there for two years and Stock Exchange, based on my opinion, because I don't know much about it, but it's all about data, all about technology, all about like fast real-time decisions. So what about your job? What do you do with data specifically? Alright, so when I joined two years ago, there was no data team. So it's actually a fairly recent initiative. We started by building the infrastructure so making sure that all the data is in a centralized location. And then six months later, I got hired. And so I started from scratch. I built the team, I decided the technology, I decided on the use cases. So the type of use cases, so as the data team, we serve the whole company. So we serve our securities and derivatives market, but we also serve operation, technology, and regulation. So we have really a very wide range of problems to solve. So the first thing is trying to understand how people trade on the platform. So if you understand their behavior on the platform, we are able to serve them better and recommend them better products. Then we also have a lot of things around text. So believe it or not, we have a lot of data in text. So all the corporate actions, so when a company declare a new dividend or a new board of directors, they have to send us the data and they usually send it as a PDF. And we have like 1,000 PDF, 1,000 pages PDF long, and we have all this data in there that we have to mine. But we also look at operations. So for example, latency is a really big issue for traders. They want to have their orders in the first, as fast as possible on the trading engine. So we make sure that there is no anomaly on the latency side. So it's really, really a very wide range of problems. And that's why I really like this position because really exposed to a wide range of different issues. Thank you so much. Let's go to Jen. Jen is in advertising and tech, Facebook. And she's a partner in marketing science. So how, and I know that we all know when we say Facebook and data, it's a little bit sensitive. But how is it from your position? I won't be speaking about all those topics today. We don't have enough time clearly. No, what is your, from your position, how do you work with data? Yeah, so we're working with clients, many brands that are represented here on the stage today. And I'm sure many of you in the room on a daily basis to assess advertising effectiveness. So proving that out in various forms on Facebook as a platform alone, Facebook properties, and kind of what we call a cross publisher, cross media as well. So as we look at the relationships of advertising across digital and TV and where we can assess that. For me, as Yana said, I'm a partner in what's called our marketing science department. But there are many different departments where I have colleagues who focus on data and data analysis. So whether it's day-to-day campaign management and assessment, it could be our business integrity team who's working with the platform and with our product teams to assess, you know, bad actors on our various platforms. It's actual operations analysts who are working with product to understand how you all are using our platforms so that we build better products for that very need. So there's many different roles that encompass data. So as Yana said before, kind of, you know, touching on everything. Me, the way I face my role, I actually am similar to Maggie, but a little different. My background is not in data analysis. I come from startup and large advertising agency background. And the way I... Well, the way my role came about was by way of a reorganization. Shortly after I joined Facebook. So I joined Facebook in an advertising technology role. And then, obviously, advertising technology is used to assess advertising effectiveness, right? So it only made sense that we re-ordered into marketing science. And with that, I was asked to move into this role. So my company made that request of me. And at that time, I could have panicked. But when you think about it, whether I was on the startup space before and building out models or on the advertising side before, always assessing my client's data, it wasn't that much of a stretch. Yes, were there things I had to upscale on? Yes, I took the initiative to upscale and kind of go on data camp and work with my colleagues, ask them for help, you know, okay, teach me how to merge these data sets most effectively. And you just have to find your way about it in that respect, I would say. If I were to give one piece of advice there, it would be don't let other people define you. You define what you want out of that yourself. So for me, I come at the role a little bit as a data consultant. I know what advertising agencies, what my clients need, the challenges they face in servicing their clients. And I try to come at it from that angle. And I want to say, Jen is our, I actually run community that is called She Loves Data. Do we have anyone here from She Loves Data? Yay! Yay! This is fantastic. So what we try to, we try to raise awareness about data and we try to say that ladies, data is everywhere in every position, no matter what you do, you don't need to be in tech, you don't need to be in STEM. Make sure, as you said, learn some basics around data because in that way you will be actually future-proofing your job. And Jen reached out to us like few months, one and a half years ago actually, and Facebook is our partner and helping us to reach out to the women and help them to take this education for free. And I know Jen actually brushed up her SQL skills to be able to be one of the instructors. So that was fantastic. Now, thank you, Jen. Let's go to a different industry. Industry that is as well as advertising and marketing full of data. And it's e-commerce. And I know you are one of us here who as well studied computer science, so it comes natural to you. What is your role and how are you guys using data? You have 50 people in your data science team, right? Yeah, yeah. Or data analytics team. Data analytics team. So my name is Dujuan, and so I'm now working in Shopee, leading a team of around 50 people of data analytics team. So what we do, what is the daily job in Shopee for data analytics team? Basically, we utilize our data to support both product and the business. We know like for a product data-driven lifecycle, it includes like design, development, and analysis. So for data analytics team, we participate a lot in the design and analysis stage. For example, right, since we built our in-house user tracking system, which is something very similar to Google Analytics. So we based on these user behavior data, we did a user journey analysis, and we found like if we simplify the user journeys to reach to our recommendation module, the outcome will become better. So we give these suggestions to our product team, and product team accept then they design the new features and then release. So we finally value the outcome. The outcome is like the order number increase around 5%. So this is what is the impact of data, and in addition, for e-commerce companies, we know for users, for people, somehow it's a little bit hard for us to find something on the platform, especially on some like ladies fashion dress. You see this piece of item is very nice, but I don't know how to describe it. So yeah, we also provide this suggestion to our product. So we have the image search function, which allows you to take a picture and then search our platform to find your nice piece of dress easily. So this is how our daily life looks like. In addition, from my opinion, another area that becomes more and more popular is about causal inference analysis. The reason is like what we are doing for the analysis right now, most of them are correlation. So correlation somehow is very different from causal. For example, so here is a question, right? Whether attending these coding girls' events can help me to find a better job? This is a question, right? Maybe we can do some studies based on the people who are attending these coding girls' events, and we found out it seems like it can help you to find a better job, but is it really scientific? It might because the ladies, the girls who are attending this event is very talented, it's very smart, so they can find a better job. So causal inference analysis is another very interesting topic, and I think if you want to join the data area, you can start to learn these causal inference theories, and I think the market will need more and more people who can conduct this analysis. This is fantastic. I think you jumped into actually a real example of how data-driven innovation can be run and how companies are doing it. Do you ladies have some example of a data-driven innovation? Let's just simplify it for me and for everyone here. Data-driven innovation is something that we use data to create some results that help innovation in the business, right? Just let's simplify it like that. But how do you measure that? Do you have some examples of projects that your organizations put up out there based on data, data-powered innovation, and then you were able to measure right away the positive impact on the business? Anyone? It's there. So yeah, the way we work with the data is that every time we're scoping a project with the business, we try to actually assess what would be the impact. And then we're running a small-scale pilot where we are actually seeing the impact of the model before deploying it. So we are tracking all the time what is the business impact in order to also transform the business. But because also the team might be a bit tired about evaluating all the time or quantifying all the time the impact, we're working on a what-if tool that will be able to automatically tell us what would be the impact of such a machine learning model being deployed in their organization. Any examples, one of you who can actually say what the impact was? Yeah, I mean, we've been rolling out since we've had real-time life data coming in that every moment. We know what's happening. It's raining out there. You are sitting in Vivo City. You love Chinese food. It's personalization about how we can actually give our customers better offers of what they really need. Since rolling out a couple of the real-time models, we rolled out one where we actually make the merchant offers that are most appropriate at that point in time. And then we rolled out another model where we offer financing when we know our customers just bought a large purchase. And actually, would you like to share with us your interest financing options? And actually, those two, within a couple of months, and we always measure them, it is very important to us in the banking industry to measure what we're getting in return. So just a few months in a couple of markets because consumer has 17, so we tend to pick which markets we go with. Two million additional sales just in a couple of months. Fantastic. And I think one of the things about innovation is there's so much technology that's going coming out. The difficulties in the technology, technology is there, is actually finding the right products and the right customer needs is about actually being customer obsessed. That's actually going to be able to and actually developing those products that actually does add value to your customers. And as well maybe collect all the data that is useful because we collect so much data and it's all out there. But how do we define what type of data is useful for these type of innovation projects? Sigrid, you wanted to add something. Yeah, I think when you are not directly customer facing, it's harder to measure the impact of what you're doing. And so the way I approach it, for someone inside the company, I build a dashboard, I build an analytics, I want to know whether they're using it. So I don't just build my model and that's it. So I go back to my customer, my internal customer and ask them how it has helped them. I also try to find ways to measure how often they have logged into the dashboard or how often they have run the analysis. And I think that's a little bit tricky for us because really we serve internal people and then they serve the customer. So it's a little bit more diluted. So I think that's a little bit of an intermediary measurement to see how the adoption is there. That makes sense. That's actually very true. How do you measure is different for different industries and businesses. Jen, can you share a little bit about how you actually drive the innovation so it gets to everyone within the organization. And I mean, because often what we see, we have a group of data scientists, group of data analysts, they are very aware that we change the adoption process of the entire organization to work with the data in a smart way. Because what I've seen with many of my clients have been working with data for 20 years is that it's easy to put the innovative products out there, but the adoption process of the organization to change the processes and mindset of people is the hardest part. And I know you guys are probably, or I have the impression, you're good at that. There are some key things to change. For us, it's not only the analysis involved, because our analysis is informing our auction every day, so that's just a common practice there. But it's actually using that analysis to educate and educate not only our clients, but the industry and the advertising industry and even broader than that. We're having data-led conversations that are talking about advertising effectiveness and why are we looking at apples-to-apples comparison of ads on Facebook, a video ad on Facebook versus video TVC, right? It's not apples-to-apples. And so, yes, we need to have benchmarks in the industry, but we're looking to kind of educate and bring about, well, this is the way people are using our platform, for example. Think about it. How many of you scroll rapidly like this, your Facebook newsfeed? You're a thumb scroller. It's either thumb or index, usually. Yeah. If we find any other fingers, I think you're anomaly. But, yeah, it's using those insights, like I said, and not only informs advertising effectiveness and how marketers can use that data, but also informs our product teams. How you all are using, that's how we develop products. It's funny when you hear in the trays or just consumer publications sometimes, oh, why did Facebook roll this out? It's all you are behaving. We're following those trends, right? Maybe you're the anomaly, you're the outsider that's not on that trend, but that's kind of a very common way. Makes a definite sense. Because we have not enough time, but can we continue for five more minutes because I have two very important questions based on my opinion. Innovation and diversity. Because we are all women here and majority of women out there. Does it play any role on innovation, the diversity? You have 50 people in your team. You have 200 plus people in your team. How does it look like, actually, in your industries? Do you have enough women? Do you have enough diversity? Not only gender diversity, but all the types of diversity. And does it have impact on innovation? For my 50 people, most of them are girls, right? Mostly half of them are women. Half of women. Because we are doing data analysis, you must be very careful. Because sometimes maybe not fair to the men, but sometimes girls are more careful generating the numbers. That's a good summary. I think it's very interesting for a team to contain a certain portion of ladies inside the team. Because the way of thinking is very different. Especially in e-commerce company. Why? Because women like shopping. They can provide more creative ideas about how to design these e-commerce products, how to utilize the data to provide suggestions. I think especially in e-commerce industry, I'm not quite sure about other industries. We want ladies to join us to do fancy data analysis stuff. This is a call out to you guys. We are hiring. Let's do this promo because I know Maggie is hiring as well. I'm going to shamelessly say I am always hiring. Hiring women in data. I've always been very, very passionate about women in technology. I actually run this women's network with the computer society on women in technology. Since coming into data, I've become more and more passionate about women in technology and women in data. It's more than just actually having the women's mindset when we are designing for customers. And it's very, very important for us to actually create in the data community enough women with the empathy and the attention to details and the sense of community to actually make sure that we don't go down the route of multiplying biases. And especially when people now are talking about in data a lot of machine learning. And again, machine learning becomes very unpredictable sometimes, unintended results. It's very, very important to have more women just taken over my team, not enough women. Please come to me if you are interested in joining data. This is the ethical approach to machine learning and intelligence is a big topic right now. So I think what you said plays a super important role. Jenny, you wanted to add on? Yeah, I will. I think we at Facebook come from a little bit of a juxtaposition with it because Yana mentioned when I reached out to her over a year and a half ago it was on the basis of I attended our annual conference in Silicon Valley and panel like this up on stage of our senior leadership in marketing science globally. And we all pointed out the fact even the head of the organization that only one of seven of those people on stage was a female. And so I decided to kind of take that away and put it into action when I came back here and so hence reached out to Yana and we we've been involved in many CSR initiatives together in that way upskilling, educating, partnering together and we're very fortunate at Facebook on top of that we actually have what we call a diverse slate initiative. So it doesn't mean that we're going to hire you because you are a female and you get the edge that way but at least the interview process starts with that diversity right? So we have some rules around that in our interview loops if you have four in an interview loop at least one are depending on the role there's certain criteria 25 to 50 percent should be of equal gender representation. I also just came from an event this week where we have women at Facebook a women leadership day and we not this year but in the past we've had Cheryl we've had lots of senior representation come out and we're very lucky to work at a company that drives that at the same time and we know in tech that that's a real challenge and we know that we're still underrepresented and to me that's where the long tail comes in because you can't just say go out and hire women on the spot right? So long tail initiatives like Yana and others are involved in like this thank you coding girls thank you very much really matter. So I just want to add something because like Elina said earlier that we are 60 something women in AXA and that's true but most of the women are today doing a job that will disappear tomorrow they're in operation they're like sending claims or you know like more traditional kind of career path and yeah and today most of those women job will be automated tomorrow with the RPA coming along with the data science and stuff so this is why for me we'll launch the super user initiative in the company to be able to up skill the people that had no data skill whatsoever so they can be empowered by the data and maybe tomorrow become data analysts because otherwise those women would potentially lose their job in the next five to ten years so we need to take action now to prepare the women whoever they are or whatever the background to the word of tomorrow so let's close this discussion by you ladies saying one sentence okay because the time is ticking what would you advise to the audience here one take away for them to get your jobs to get data related jobs because you are the leaders in the industries across industries so what would you advise them stay curious the field is moving so fast and find that passion curiosity ladies I would say be bold everyone can learn about that be bold I would say number itself means nothing just combine the numbers with business with use cases be impactful and actually bridge the data and the business right the technology and the business be the bridge be the translator of that I'll build on Maggie's a little bit develop a growth mindset it's about resilience and not necessarily endurance resilience and last but not least to me would be get mentors but get both men and women so you will have both sides of the picture so have a diverse mentors thank you so much ladies it's been pleasure thank you can we give them a round of applause