Digital Transformation- Leveraging AI for Business Success

Deepna Devkar

Vice President, Data Science & Engineering at CNN

Learning Objectives

Please join the Vice President, Data Science & Engineering, Deepna Devkar in this Executive Interview where she will discuss the importance of leveraging data and AI in a time of digital transformation.

Key Takeaways:

  • In an evolving media landscape, what does the future of CNN look like?

  • How is your team driving the change towards CNN's digital transformation?

  • How are you leveraging data and AI to get closer to your users?

  • How do you justify and get buy-in for the investment in data teams? How do you measure ROI?

"I'm a huge proponent of building in-house data teams because ultimately you are going to be in the best place and be most incentivized to care about what the organization cares about."

Deepna Devkar

Vice President, Data Science & Engineering at CNN


Hello, everyone, and welcome to the Quartz Network. My name is Britt Erler, Executive correspondent here at Quartz. Thank you so much for joining us. I would like to welcome our guest speaker here with us today, Deepna Devkar, VP of data science and engineering at CNN, as we discuss the integration of AI and data within a digital transformation world. Welcome deep now, thank you for having me. Of course, thank you so much for joining us. And taking the time out of your hectic schedule. I am so thrilled to have you here and can’t wait to hear what CNN has been working on, especially with all of the shifts we are seeing in 2020. And all of the changes within this media landscape. Yes, this has been an unusual year, hasn’t it? It has been and for every market across the board, too. And the one thing with media is even with the shifts we’re seeing now it’s constantly changing, not just with what we’re seeing this year. Now, you are the VP of data science and engineering. And that title sounds very technical, very challenging in the world that we’re currently living in. Tell me more about your role with CNN. Absolutely. So as you mentioned, I’m VP of data science and engineering. And I co lead this team at CNN digital called Data intelligence, alongside my product partner. So we are a team that is relatively recently formed about a year and a half ago. And we are comprised of data scientists, machine learning engineers, and product managers, and specifically with the goal, to drive CNNs digital transformation as we become more user obsessed, and leverage data and machine learning to address several of the organizational challenges, which, as you mentioned, aren’t just bespoke to our organization, but just the whole media landscape, in general. So our goal is to always be human LED, and machine informed. So you know, we’re not here to automate journalism. So let’s not freak out about that yet. We’re not here to do that. We’re simply here to give access to data and build the right tools on top of it, so that we could empower our organization and various teams within the organization to use that information and provide scalable solutions that they can use to innovate further, right. So this is always going to be human lead, but machine informed. I love that. And it’s so important to really keep that human to human interaction, especially right now with everything that we’re seeing this year. And in regards to your team, how large of a team is it? How many of you are working towards that goal? That’s a great question. It’s always in flux, because this is a very hot field. So it’s a whole separate conversation about hiring and retaining talent in this area. But give or take, we’re about between 30 to 40 people right now. Very cool. And in regards to your role, specifically, as a leader, and also your teams, have you seen a lot of change this year alone? Or has it stayed pretty consistent with what you started off to do about a year and a half ago? Um, that’s a good question. So we are we have mostly straight and stayed true to our mission. And you know, CNN. So just just for clarity, I only work on the digital side of things on the on the linear TV glass side of things, although they’re very much in combined, right and even more combined now, so than ever. So cnn is mission is to inform, engage and empower the world. And you’re very true to that mission. So despite all of the changes, we are very true to that mission. What has changed, however, is the way in which we communicate this information with our audience, right? So it’s no surprise that the world has become more digital in the last decade or so and is continuing to become more and more digital. So, you know, so while our mission in how our content is delivered to the audience has changed, we are not, we’re not swaying away from what our product is, right. So there is a large diversity in how audiences like to consume news. It’s simply not standard anymore, like on the TV glass, even though it is still the majority. So like every other media company in the market right now, we are trying to directly connect with our consumers understand more about what it is that they care about, and then how to reach them in a way that is more accessible to their lifestyle. So while we are staying true to that mission, the future of CNN is inevitably going to have to be become more invested in a portfolio of digital products that drive direct to consumer relationships?

Absolutely. And and as you mentioned, it is an ever evolving landscape. And as far as the future of CNN, you guys have kind of a five year plan set in place. Are you preparing for any shifts that you’re expecting to see next year? That’s a great question. And while I can’t speak about direct shifts that may come like within a year, our five year plan, as I, as I mentioned, is to invest heavily on our digital property. Because you know, when you think of CNN, you think of what you see on on on TV, right, but we have other digital products like, we have a mobile app. And, you know, as as consumers become more and more digital oriented, and you know, try to consume news on their commute, or, well, now there’s no commute, right? so valuable. So that’s one example, you know, about a shift in which we were thinking, you know, lots of consumer from their phones, but maybe it’s going to change in the coming year. So that’s why I can’t speak directly about exactly how it will evolve. But inevitably, it is going to be more digital focused, whilst keeping our TV apps and sorry, our TV programs as is. Absolutely. And I think a lot of companies are kind of moving in the same direction as well. A lot of them are innovating. They are using technology to make their operations more efficient to make their products more appealing to their customers. How are you and your team using AI and data to get closer with your users and your customers? Yeah, that’s a great question. So as I mentioned, part of this digital transformation, right, in order to drive the digital transformation, we have to engage and connect with our users more. So one of our biggest mandates is to increase engagement across our user base so that they’re not just coming to our homepage to read what the headlines are about, and then leaving us but actually engage them inspire them to read something more beyond that, right. So one mandate that my team has is to personalize the user experience, right? So when when Britt let’s say comes to Or is consuming news on her app on the CNN app, the whole experience is personalized to what you like, right? So it’s, it’s it’s quite a challenge, that it’s not just a run of the mill, any recommendation algorithm that you can use? You know, like on Netflix or Amazon, specifically, I mean, with news, the added challenge is that it’s it’s such a moving target, right, like what you consume this morning is going to get stale very quickly. Right. So recency is an issue that we have to deal with, we have to make sure that while personalizing the experience for the users, we’re still keeping them up to date on what is breaking news, for example. The other thing is to balance it with novelty. Right. So this concept of novelty, we want to make sure that you know, we’re not just using your profile and your previous history of what you care about and what you’ve read about to inform the entirety of what you would consume and be knowledgeable about right. For example, right now, it’s typical, to come to, to just see what is happening in the Trump world, right? What is the latest headline there? What is happening in the COVID? world? What is the latest headline there, but if we just used that information to be able to train our use that data to train our model, then the only thing we would be recommending is more Trump content and more COVID content. And we would definitely leave our users very depressed by recommending just that, right. So it’s the idea of also introducing novelty into our algorithm. So how do we inspire you to read content that you that is related to your interest? But perhaps even something new, right? How do you stumble upon something that you would like to learn about? So we’re doing all of that in a machine driven way, but again, staying true to human lead? And then all of that information, of course, can feed back into our editorial content as well. Right. So we use that information to inform our editorial, hey, this is what the users are actually caring about and help them empower them to figure out new avenues of content that they could write about, as well, aside from just breaking news. So right, and in terms of this data that you’re collecting, does it primarily come from click throughs? From surveys, you know, what’s your primary source?

That’s a great question. We’re actually doing it from both angles. So not specifically my team. But there are other teams within CNN that conduct the surveys as well to directly connect with the users and say, What is it that you would like to see you know, so we have a set of core users as we call them, which are heavy consumers, they come to CNN multiple times throughout the day. So those are the users to really go after and ask like, how could we improve this experience? On the other hand, we call them cold users, which are users that come on our site every so often, but not not repetitively, right? And we want to know, what is it that’s missing? Like, why don’t you use that as like your daily, good news for news. So there’s that side of work that’s happening, but specifically for my team, we’re using click through rate as an information as okay, because that sort of, you know, survey questions are sometimes subjective. Or if they’re not anonymous, then people don’t want to answer them was probably one of those

survey, I don’t know.

This is this you know, anonymizing all user nation, but using click through rate and their online user behavior as a proxy to understanding what it is that they might care about. Absolutely. And I think something really important you mentioned, and there’s it’s not only your team that uses this strategy, but also your editorial team, as far as the other markets that we serve your courts, you know, not just finance, HR, it supply chain procurement, is this a strategy that you think other departments of their markets can really use across the board? Absolutely. I mean, data is ubiquitous, right? You can’t, you can’t ignore it anymore. So I mean, one of the questions that I get asked is, you know, how do you get buying for collecting this kind of data, because all of that cost money, like, a lot less than what it used to cost maybe 10 or 20 years ago, because now you know, every every company is collecting more and more data, but it still is, you know, and staffing for people, not just for data collection, but then somebody has to make, draw insights out of that data and make data driven business decisions, you know, so it is a big investment. But I think it’s an investment that we can’t forego for too long, right? Because this data is the new currency, and we have to adapt. So definitely, this has use cases all across the business. Absolutely. And for a company, maybe that’s smaller, or that isn’t as innovative and kind of in the know, as CNN, what do you think is really the first step for a company for a team that’s looking to begin their digital transformation journey? Yeah, that’s a good question. You know, because I prefer, aside from CNN, you know, my previous experience with was with a smaller company, and as I mentioned, yeah, you know, machine learning engineers aren’t cheap data, scientists aren’t cheap product, people aren’t cheap. So it is a heavy upfront cost to build a team, right. So one, one way that smaller companies can possibly do this is by consulting, you know, there are lots of consulting firms that offer these things on a contractual basis. And that way, you don’t have this heavy upfront cost, you know, without really measuring what the ROI would be, you know, so a low risk way would be to perhaps do a contract with one of the data science consulting companies, or even companies like, you know, Looker, or Tableau that gives you access to visualizing data very quickly, without having to learn how to code or without having to hire people that can code. So those are all low risk methods to start with. But you know, I, I’m a huge proponent of building in house data teams, because ultimately, you are going to be in the best place, you’re going to be most incentivized to care about what the organization cares about. Right. So doing just one off contract agreement isn’t going to take you too far. All right. Yeah, of course. And as far as your team specifically, how do you keep them up to date with the constant change with making sure that they’re on the right track that the strategies align, you know, across all the departments of the company, especially right now, in 2020? With all the shifts, we’re seeing, everybody’s roles are changing, they’re expanding? How have you kept your team on track?

That’s a good question. It’s actually not been all that challenging to manage both teams. Yeah. And I’m very happy about that. Because we were pretty distributed. Anyway, we have our team has people who are in Seattle, DC, and London, Atlanta, New York, LA. And so we were pretty distributed to begin with. And that’s part of the reason why the adaptation wasn’t as difficult. But I would say that it’s actually made it better in a lot of ways. If If you can put something positive to all of the all of the sad things that are happening in the world and for us. Yeah, but the positive has been, you know, people don’t have commute anymore, so they’re more productive. And so I think that’s been really nice. And as far as keeping them up to speed Then connected across the different business organizations and even across the the field of data science and machine learning. I think people are more protective of their time within within the business organization. So we started doing a lot of pre read documents, as opposed to doing, you know, meetings and more meetings, we’re not just tethered to the computer the whole time, right? Or tethered on zoom or WebEx. And that’s, that’s been quite helpful. And then as far as staying up to date in the field, I think, honestly, you know, before, it’s hard, you know, a lot of these conferences in the data world are extremely expensive, right. So you can’t actually set out the budget for the full team to go travel to the west coast or internationally, even to participate in these conferences, to stay up to date. But now, every single of those conferences have happened virtually, you know, so while there’s registration fees, but travel isn’t so expensive anymore, because you’re just logging in from the computer. So that’s actually been quite nice, too. It is. And it’s something that he records, obviously, all of our events before this were in person, that was kind of what we specialized in, and we really had to pivot and make them virtual this year. But I think even once things go back to normal, or as normal as they can be, we’re gonna still keep a virtual summit and virtual aspect to our shows, because it’s been so effective. And people are realizing they can work just as efficiently at home, it’s cost effective. They’re not having to travel be away from their families. So I completely agree with you. And in terms of the efficiency of your team, have you noticed any change with it going virtually? Or have you realized they’re working just as well from home? No, it’s actually been even better working from home. I mean, I was on on maternity leave when the whole world was adjusting to it. So I actually had the luxury of coming back to it when everybody was like, had adapted. And then I was the only one who had to adapt. You had a double sided? Yeah. But, you know, my team, they basically there was a dip, you know, like, everybody, everybody saw it, because we were just, you know, and all of this happened, we were just trying to figure out what the hell to do with it, you know, what does it mean, for the future? Should we stay? Should we not stay? You know, team members, we had kids had to figure all of that out. And so there was definitely a dip in the beginning. But I think people adapted quickly. And I think it’s actually kind of nice, you know, we’ve gotten to know each other. Well, we’ve gotten to know each other’s pets and kids well, because they just pop up on me all the time. I mean, as far as you know, my team is concerned where we’re gogogo at this point, exactly it and I’ve heard that, you know, across the board as well, I think it’s something that will actually stick, even even through next year and years to come. And then obviously, in terms of being a leader in this time, a lot of changes on your end how you manage a team, what are some of the biggest challenges that you’ve seen right now in your role? That’s a good question. I think for leaders in general, I think, the old ways, and by old, I mean, not just this year, but you know, lately, you know, when you are managing millennials, for example, there’s just a different I’m a millennial, myself, but

a different way of managing them, you know, it cannot be so top down. And in fact, it’s not going to serve you well, if it’s going to be to top down, right, because, especially in my field in the data field, you know, I, I alone cannot be an expert on every single machine learning algorithm that’s out there, or, you know, be be on top of the media domain, and what is the latest thing that, you know, New York Times is doing, for example, or some other companies doing. And so I think you have to rely more and more heavily on your team, to be the experts. So what you can best do as a leader is hire a team that is smarter than you and you know, more more not generalize, but more specific in their training, and then hire a bunch of diverse people, right, not just in training, and not just in backgrounds, but also in personal lives, you know, and so, I think if you let them if you empower them with the right tools and appropriate culture, then I think that, that that will put you in a great situation to lead them despite, you know, remote or not. I couldn’t agree with you more. And and something that’s something I’m going through currently, right now, one of the small positives of COVID was that we had the opportunity to expand our content team here at quartz and we brought on a number of new team members, all from diverse backgrounds, different stages of their life, different stages of their career, but it’s helps our team Excel and really innovate because they have so many new ideas, so many different ways of looking at things. They’re bringing in. from all different jobs that they’ve had, and that’s not something, you know, we’ve had in this department before, so I couldn’t agree with you more really bring on that diverse team. Yeah, and you know, it’s a, it’s a millennial thing. But especially so in the data world, like, we are just not the kind of people who will retire at this company or be there for a long time. Right. So that’s the kind of employee you’re looking for, you’re already setting yourself up for failure. So, you know, in fact, there’s a lot of advantage and bringing people as you mentioned, you know, from diverse backgrounds, including their different company experiences, you know, we’re trying to hire people actually, not from media backgrounds, you know, because part of the transformation is having someone who comes from a different background, and not just from the same world that we’ve seen, right, so totally agree. Absolutely. And as far as any final words that you have just how cnn is driving the change towards digital transformation, any advice you may have for leaders that are in a similar role? I think I think embrace change is the best advice, you know, if you try to lead the pack into some with with an old mentality, that’s not going to take you too far. Right? It is, this is a moving target, and you have to adapt to it. I mean, honestly, all of the things that we had talked about last year, are again, we’re we’re reconsidering where where will 2020 leave us in 2020? You know, so it’s a constant shift. We constantly revise our okrs we revisit them every month to see if this is still something that that still makes sense. So I think embracing change, being opportunistic, and, you know, keeping keeping your eyes on the consumer is definitely the way to go. Absolutely. Well, deep. Thank you so much. you’ve provided our audience and myself with some incredible insight today. And thank you to everyone who has joined us and tuned in. Please do not forget to ask any questions that you have for Deepa, any comments in the discussion forum below. Please be safe, be healthy and enjoy the rest of the summit. Thanks so much for having me.

Thank you.

Get full Q/N Access

Sign up to Q/N with a few details to watch this presentation.

  • Hidden
  • Hidden