Canva's first Chief Algorithms Officer on what the data misses - with Nirmal Govind

21:05Nirmal Govind:It's just very hard to predict. And so we have an internal system, you know, where we basically look at the metrics as some, you know, unusual launches. And, you know, there's people that will make bets. Like, I think this is the one that's going to win, right? And most often human intuition turned out to be wrong. people's preferences are sometimes just very hard to predict. And so, yeah, that's one I can think of. And sometimes, you know, you'll see patterns which are kind of interesting. And maybe in hindsight, it's like, yep, that makes sense. But an example was, we had a show where in the first episode, towards the end of the episode, there's a battle scene.

21:41Nirmal Govind:And the director decided in the second episode, they would actually start the episode with just a continuation of that battle scene, no opening credits, nothing, right? It just continues. And that thing in the second episode goes for about 20 minutes. Typically episodes, you know, that show are about 40 minutes or so. And at the 20 minute mark, you start seeing these opening credits show up. And we saw a whole drop off. People were like, is it finished? The episode's done. Like, let's exit, right? And so it's like, you know, in hindsight, you're like, yeah, that makes sense. You know, you put the credit somewhere.

22:15Nirmal Govind:And so that was an example of something that, you know, we looked at the data and we're like, huh, that looks interesting. Why are people dropping off right in the middle of an episode? And that's something that we fed back to the creative teams. Like, hey, we're seeing something like this. So watch our trade next time kind of thing. Yeah, that's interesting. So that feels to me like it might be naive, but I imagine that there was at times tension between the folks that you mentioned earlier who have like a strong intuition. They've built shows, they've sort of promoted new types of content, they're coming in, maybe pitching Netflix on like, I believe this is a show or a movie that has real mass appeal.

22:48Your team gets sort of wheeled in to say, well, actually, like based on the data, we found that actually that sort of, you know, content doesn't tend to do so well on Netflix. How did you sort of think about the role that you played in that creative process? It sounds like you've applied quite a lot of humility to that process, but like, can you tell us a little bit about how those experiences played out within a Netflix context?

23:12Nirmal Govind:Surprisingly, probably, the tension, I would say, was very minimal. I feel like at Netflix, just because it's a very data-driven company, and it starts right at the top. So Reed himself is very much a data person, and he looks at metrics pretty much every day and looks at them very, very closely. And so generally, I would say, across a company, you could sort of go to any team and say, hey, we've got some ideas, and here's things that we want to do with data and machine learning and models and so on and so forth, then you don't generally get any pushback, which is very unique in terms of a culture because a lot of places people can be a little bit more protective and so on and so forth.

23:50Nirmal Govind:But I would say Netflix that way was great, which is why I would say we went and built out algorithmic experiences and different kinds of models and so on and so forth in many different parts of the company, that otherwise, including studio, that otherwise you would never think to apply algorithms and data and machine learning and so on. So because of that, I think, you know, generally I would say it's not as much tension as you would, you know, imagine. I think there was definitely an understanding that the creatives know what they're doing and that's why we hire them and so on and so forth. And so, you know, if there is a question mark there, I think we definitely defer to judgment more than the data.

24:29Nirmal Govind:But I think the data was more to inform, right? And it's sort of like if you see something in terms of data that as a creative, you're like, huh, that's interesting. Then the hope is that they would dig in a little bit more with the team and say like, okay, why are we seeing that? You know, what's going on over here? What should I be thinking about? You know, once in a while, once in a while, we'll get, you know, some folks that come from the more traditional sort of studio network type of environment, because Netflix, even as a studio, operated very differently, I would say, from what I've heard.

25:00Nirmal Govind:I've never worked at another studio. Other studios in Hollywood operate, right? And so they're not necessarily looking at data as much. The information flow is a lot more controlled at other studios. And so sometimes once in a while, you'll get somebody that comes from that, hasn't quite figured out how Netflix operates or how the culture works and so on and so forth. And they might be like, oh, you know what? I need to look at this and approve this metric or whatever. And then that's when you have to intervene. And you sort of say like, no, no, no, that's kind of not the way we do it. Here's why we don't do it that way.

25:30Nirmal Govind:Here's the benefit of it. You sort of set context, right? And so once in a while, but I would say that was more rare. Yeah, interesting. So you obviously got a thread in your career of this sort of intersection of creativity because it wasn't just at Netflix. You then went on to Fable to do, you know, algorithmic design in and around the experience of books and then Mango. Maybe you could talk us through a little bit about your forays at this intersection of sort of content creation and experience of users and the application of data science. Why is that the part of your career that ties it all together?

26:05Nirmal Govind:Yeah, it's a good question. I mean, there's an earlier part of the career, which was, you know, more about operations and, you know. Well, you started in Semicon, right? Yes, that's right. So maybe actually let's go back to there. Like you started there well before it was even a thing that we talked about in such short terms. Why was that the place you started? And then why didn't you stay in the kind of hardware oriented part of the world? Yeah, so this is going back a long, long time. But I came to grad school in the US and my advisor was working on semiconductor manufacturing. This is like the late 90s and when manufacturing was still very much a thing in this country, I would say.

26:44Nirmal Govind:It still is. But he was working with a lot of semiconductor manufacturing companies, both in the US, but also internationally. And one of my first projects was actually working with Samsung in Korea and trying to figure out planning for their factories and scheduling for the factories and so on and so forth. So yeah, I would go to Korea, sit with my laptop with engineers there and like, you know, we're trying to run these models and see how they work. That sort of led to work at IBM and Intel and so on and so forth. But towards the end of my time at Intel the last year or so, I got very interested in healthcare and, you know, Intel had a digital health group.

27:20Nirmal Govind:So I started working a little bit with them and And they were kind of, the hospitals are asking the question, if you guys know how to run factories so well, can you actually help us run hospitals better, right? And so I worked with the team there and we traveled around the country, like talking to different hospital directors and so on and so forth. And it seemed like there was a lot of opportunity there. But then Intel decided to go like remote health and so on, which was not quite what I was interested in. And then sort of coincidentally, a founder reached out to me and he had a startup which was doing, you know, health tech, so healthcare, like scheduling for physicians.

27:56Nirmal Govind:And I was like, huh, this is kind of serendipitous, like, you know, similar timing. And yeah, sometimes the universe just works, you know, in interesting ways. It's the squid game of your own career. Yeah, exactly, right? And so, yeah, so I talked to him for, I don't know, six, seven months before I said, finally said yes. And, you know, I joined a small startup and that was a great experience in terms of like working more with end users, right? because we were working directly with the doctors. And the schedules that we're producing affected their work-life balance, really. And it was a big issue because Dr.

28:27Nirmal Govind:Bonnard is still a thing. Physician Bonnard is still a thing. And so we were trying to really help with that problem. But that's where I would say, you know, like working with actual humans and end users was, you know, very satisfying, very, like, interesting. And, you know, that's sort of when I started thinking about what else we could, you know, do. And Netflix had this very interesting role about machine learning in different parts of the business. And I was like, oh, they've got a lot of data. They directly impact consumers. And we were consumers of Netflix. We had Netflix accounts and watched and so on and so forth.

29:02Nirmal Govind:So that kind of got me curious in terms of what can we do with machine learning on the consumer side. But that's kind of how the transition happened was really just curiosity in terms of both machine learning but also in terms of consumer software and so on. But yeah, ended up spending a long time on Netflix. And I think one thing led to the other and Fable sort of came along. And that was an interesting problem because it was sort of started around COVID timeframe. And the idea was, how do we actually use technology to help with mental wellness? And reading, it turns out, if you read 30 minutes a day, it can actually increase your lifespan by a couple of years.

29:39Nirmal Govind:Yeah, there's some studies out of some academics that actually show that. And so that was the key insight that the founder had. And it was like, okay, we should actually do something to help people read more, build communities. So there was an online book club part of the app and so on and so forth. I was like, and I read. And so maybe not as much as I should, but definitely was trying to look at how do you reimagine the reading experience? And so it was definitely a creative plus modeling operational type of a world. And yeah, that's sort of how it went to Fable. Well, and Mango AI was, you know, I think similar thread, maybe to your point, which was we loved video.

30:17Nirmal Govind:And, you know, we did that for a long time and definitely enjoyed being at that edge of creativity and, you know, algorithms. And, yeah, the question was, you know, what kinds of things could be explored in that space that given the tools that we have now? Because when I left Netflix, it was, you know, 2022, early 2022, Chachi PT had not yet come out. And so all the things. Hard to imagine. It's hard to imagine, right? But all the things that we did at Netflix were pre-LLM. And so we were sort of, you know, thinking about given the new technology and especially the visual LLMs and all those models, what can we reimagine in terms of things that we did in the creative space that we could now do with like the latest technology?

30:58Nirmal Govind:And that was sort of the genesis of Mango AI. And yeah, it was still connecting the creativity, you know, sort of algorithms slash data thread for sure. And you mentioned earlier, you sort of dropped in that your co-founder was who you first hired at Netflix. When I chatted to him in advance of chatting to you, he sort of shared this interesting thing where he felt that the two of you explicitly or implicitly had always danced around this idea of like, it would be fun to found something at some time, you know. So I'm really interested in like, you know, by this point, you were a storied, you know, machine learning leader, you know, C-suite, all that sort of jazz.

31:35What gave you the reason to leave? Like, what was the reason to say, now is the time we're going to have a crack at founding a company?

31:45Nirmal Govind:Yeah, I mean, I've always been interested in startups because, you know, after Intel, I was like, let's go try a startup and went to the health tech startup and so on. And so I think there's definitely a thread there in terms of like the excitement of a startup, the excitement of, you know, doing something new. And it's always very impactful, you know, in terms of startup, right? And so that was definitely there. And yeah, even after, you know, I left Netflix and Vineet left Netflix, we stayed in touch. And once in a while, you know, we'd get together to play tennis and, you know, start talking about startups and so on and so forth.

32:18Nirmal Govind:So that was definitely there. I think for me, the thing that really helped was, you know, I've sort of over time, I suppose, as I'm getting older, adopted more of a regret minimization framework. So, you know, I think the idea is like, you don't want to sort of 10, 20 years down the line sort of feel like, oh, I should have probably done that, right? And like, that's kind of a lens that you can use. And we still have this leadership seminar at Netflix, they still run it, which is less about leadership training. It's more about how do you think of yourself as a leader in the world and how do you think of yourself as a human in the world and so on and so forth.

32:57Nirmal Govind:And you're reading very deep stuff like Leo Tolstoy and Dr. King's letter from the jail and all those kinds of things and having very deep conversations. And one of those topics, one of the questions they ask is, what would your 80-year-old self tell you now? And what would you tell your 19-year-old self now? And so we're all sort of mid to late stage career, if you will, at that point. And it's a very interesting reflection question. And so if you apply that, and that's sort of a lens I try to apply more, I'm like, I would definitely regret not having founded a company 30 years down the line.

33:35Nirmal Govind:And so that was definitely part of it. And then I think this confluence of the tools coming together and just the sort of time that we're in, felt like a good time to actually, you know, try something new. Both of us being a little bit more later stage in terms of like, you know, we're not fresh grads out of college trying to do something. I think definitely gave us a little bit of a cushion and, you know, financially as well as like we had contacts, right? So we could go and approach people for trying to do a B2B thing. We can get pilots and so on and so forth. So we had the network in some sense to go try something.

34:08Nirmal Govind:And both of us were very much like, let's go do it. Let's see what happens. And, you know, So it's either way we'll learn something and, you know, we'll have fun doing it together. And so, yeah, that's how it happened. Well, you've had an unusual journey since then because it was sort of, you know, sooner did you like create the website and make it public that you had acquisition opportunities coming at you. And obviously you've just joined the Canva universe as one of the sort of recent acquisitions into Canva and its AI push. I'd love to know a couple of things. The first one is like for someone who spent all of this time thinking of founding a company is a thing that I would love to have done.

34:45Did you feel like you got far enough in or was there any part of you that was like, oh, not yet? You know, it was so fast because people were so excited to think about the opportunity of bringing you into Canva. And I'm sure there were many other conversations that were happening in the background, too.

34:59Nirmal Govind:Yeah, it's, you know, definitely something that we still talk about. Was it too soon? Because what was it? It's 18 months ish, right? Yeah. by the time. And I think, you know, it was, it was a very interesting confluence again of things because we were, we started off working on some other problems. And, you know, I think one thing we were both very clear on because we did this at Netflix too, which is finding the right problem is like 60, 70 % of the, of the problem, right? Like off a startup. And like, and that helps you find product market fit and that helps you like scale and all of that kind of stuff.

35:33Nirmal Govind:And so we were very quick with, you know, going through multiple ideas. And so we found out that trying out things with actual customers is actually a great way to find out if, you know, you have a problem market fit or not. So we did a whole bunch of pilots in different areas and so on and so forth. And then we settled on this problem. Actually, say a bit more about that. Did they all know that you were pursuing sort of two or three different potential businesses simultaneously? I mean, we weren't doing it at the same time because it was just the two of us. So bandwidth wise, you know, it's very hard to do multiple things because, you know, we're sort of like meeting multiple times a day, meeting at night, like, you know, it's like an ongoing thing.

36:06Nirmal Govind:And so once we're in some particular idea or particular area, we're going full force on that, right? And I think it's just being disciplined about looking at the feedback, looking at the data and sort of, you know. I mean, it's kind of meta to have people with your backgrounds, like applying the like, when do we let go? When do we know enough information to let go? Did you find a tension between your intuition and the data there? Like, did you internalize that prior tension? I mean, sometimes, you know, it's sort of your, you really want this to work, you know, because like, hey, that's an area that we know very well and close to it and so on and so forth.

36:40But yeah, you have to look at it very, very closely with like a data lens, but also just like looking at the feedback and talking to people.

36:46Nirmal Govind:And, you know, we would have internal discussions, debates around like, is this really, you know, worth pursuing? Is this worth investing? I think that having done that and having done that in fact together, because you meet us on my team and Netflix do like many, many times, definitely gives you a shortcut, right? Because you don't have to like discuss for too long. You can sort of like, yep, doesn't make sense. Let's move on. Right. So that definitely helped. But yeah, once we found a problem, we were like, okay, this could be something, you know, this could be impactful. And we were getting that signal from our pilot customers too, for sure.

37:18Nirmal Govind:And we were looking at the data and we're like, yep, there's something here. But what happened was, it turns out Canva had acquired, you know, this company, Magic Brief, from here as well. Also a Blackbird company, yeah. That's right. And so I'd known about that because I was in touch with Cliff already before that. And, you know, he sort of mentioned, yeah, we're acquiring this company. But then at their event in October, they announced, you know, what they were actually doing. And we were like, oh, this is actually almost exactly, you know, similar to what we're actually doing, except that we were probably a little bit further ahead.

37:49Nirmal Govind:We were, you know, in the video space. We had figured out the learning loops, that kind of stuff. And so... This is sort of ad and marketing optimization. Yeah, this is video, you know, generative video for ads and using reinforcement learning loops to actually, like, automatically improve them and look at the performance and so on. So, yeah, I think it was sort of, like, very coincidental, very, you know, accidental, if you will. And, yeah, so then we were like, okay, I told Cliff at some point I'll give you a demo. And I was like, okay, I should probably give you a demo now. And then, you know, we need, coincidentally, we need sole boss from Roblox, Steph, who leads AI research at Canva, was also at Canva.

38:26Nirmal Govind:And so a lot of these things sort of like, you know, fell in place and yeah, we showed it to them. And it was tricky for us because we were about to raise money. The thing that I would say we were very good at is, you know, actually hiring and, you know, getting strong talent and so on and so forth. And, you know, done that in the past as well. And so we still talk about it. It's like, we actually kind of got to that fun part. We almost got to the fun part and really good stuff, and we kind of didn't get to do it. But yeah, there's also another aspect there, which is we were about to raise money, and we had some friendly VCs.

39:01Nirmal Govind:We were doing demos too, and one of the first demos actually, very well-known VC in the space, he had a very huge exit a few years ago and so on. I don't know, 20 minutes in, he was like, hey, guys, I want to fund you. Here's what we're doing tomorrow. Here's the deal. and I was like, oh, we're not ready for this. And then around the same time is when the Canva discussion started. And as you know, this person would have wanted to give you even more money even faster after you were like, thanks, but no thanks. Yeah, so I mean, he was a good sounding board, if you will, and he was like, hey, this could be a generational company.

39:36Nirmal Govind:You should really think hard about this and we'll help you find other investors. We can do this very quickly. So it was an interesting time because we were definitely at that point where we knew we could actually raise money and we knew we could go at this very hard. And on the other hand, you have Canva here. And so, yeah, it was an interesting time and tough decision. What nailed it for you? Why path A, not path B? I think, you know, a few things. I think Cliff was definitely very hard to say no to. It is a unique skill of his. Unique skill. He's a great founder, I think. But I mean, for us, you know, we're both very impact-driven people.

40:12Nirmal Govind:and we looked at the trajectory and we were like, yep, we could build this out, but it's going to take us some time to actually have the kind of large-scale impact that we want to have. And Canva was right there in terms of like very aligned, in terms of what they wanted to do. They already have sort of a massive user base and trying to expand more in B2B. And so we're like, impact-wise, this would be much higher impact much sooner. And so that was definitely a factor. I think the other thing was definitely the team. We talked to, you know, everybody was great. And I think the mission of the company, you know, definitely appealed, right?

40:49Nirmal Govind:You probably have the two-step plan and how you think about, you know, what to do in terms of like good for the world, right? That definitely appealed to us as well. And so we're like, overall, it seemed like, you know, a good thing to do. So yeah, we felt good, I think, doing it. And yeah, so here we are. So maybe let's just quickly get into the nuts and bolts of the M &A. For listeners who have never been through one, and obviously you've got an N of one, so we'll take that as it. Is there anything that was surprising about the experience of the acquisition that others could learn from? Yeah, I think, you know, I've been through a couple acquisitions, not as a founder, but seen, you know, how the process plays out.

41:30Nirmal Govind:I think the one that really was surprising for me was, you know, we had a great legal team, fantastic lawyers and well-known in the Valley and so on. but despite all of that the amount of time that you as a founder have to actually spend in the minute details of these things and you know figure out when you should actually not let the lawyers go further on this thing and you know figure out with Canva you know like hey do we really want to go with this because both sides you know the legal sides are trying to protect their customers or their clients right and so at some point you're like this is probably not worth actually spending a lot more time on let's actually just figure out what, you know, compromises.

42:09Nirmal Govind:That kind of thing and just like being very close to it, I think a need to stay very close to it was a little bit surprising to me. But yeah, that's really what. Because of course, what's not often talked about, and lawyers can sometimes forget this, is like, if it all goes to plan, you will then become colleagues. So it's not straightforward that, you know, you might be adversarial on a particular legal point at this particular point in time, but like the relationship matters too. And your experience and feeling about continuing to work together is a big part of it as well, I imagine. Yeah, absolutely.

42:36Nirmal Govind:And for us, we were like, hey, it's not a very old company. We don't have a lot of investors, nobody else to ask approvals of and so on and so forth. So it should be very quick. And honestly, in the grand scheme of things, it was actually very quick. And Canva was great to work with. I think the team was amazing. And it was a very, very smooth process compared to other processes I have seen and heard of. And so I think overall it was very, very smooth. But yeah, that was probably one surprising thing to me. It's like, despite thinking of it as a potentially easy deal, you know, it was still, you still have to pay a lot of time and attention.

43:13Yeah, professional services only get you so far. Interesting. So let's talk a little bit about your plans at Canva. So you mentioned this earlier, you and Vinita were interested in the scale opportunity, you know, quarter of a billion users. help educate me on how do you think about algorithmic design in a Canva universe? You've done it before across multiple business lines within Netflix and in other roles. You have this opportunity to sort of also apply a lot of that to content creation experience for users. Canva has this really diverse user base. You've got everything from like kids in school all the way to, as you say, B2B or creatives for whom Canva is a tool they use to express themselves.

43:54That's such a huge universe of user needs to try to think about. How might someone like you look at that opportunity and think about how to prioritize where you spend your time?

44:05Nirmal Govind:Yeah, prioritization is always a great question, right? I think ultimately it comes down to like figuring out the biggest impact area to go after first. And, you know, I'm definitely data driven. So, you know, look at the data and see like what kind of numbers are we seeing across different parts of business? What's the metric you would go after? Is it like non-repeat users? Like are you going after like, well, that's a signal that they didn't get the aha moment in the product they were looking for? Like what's the canary in the mind shaft for you around where the impact might lie? I mean, looking at simple metrics, right, like activation rates or, you know, retention rates in different parts of the business, looking at, you know, turn rates.

44:46Nirmal Govind:I think those are sort of things that you start looking at. And then you've got to go, you know, levels deeper to figure out, like, is the problem really, you know, an algorithmic problem or is it, you know, further out and it's not something you can control. And so sort of start digging in and then asking a lot of questions. And, you know, once you start doing that, you'll sort of get a sense for where the opportunity might lie. In the case of Canva, I think there's definitely opportunities that I can see in terms of, like, where we can go after first. And the user side of it versus the B2B, I think the consumer side of it is definitely a great focus area to begin with because that affects millions of people and very, very, very impactful for sure.

45:27One of the things I found fascinating about Canva is, you know, there are very few companies, I think it's now listed as like the third most used consumer AI platform in the world, which is fascinating because they were not, you know, quote, an AI company. but they have since become more so and through acquisitions and their own internal development of that kind of part of the business have really sort of refashioned themselves successfully into this new AI driven world. But I think one of the things that I've always found fascinating about how the founders talk about it is there always seems to be quite a sensitivity to the underlying customer base that has tended to be design oriented or creative oriented people for whom AI is not a universal good.

46:10They see in some respects, it's threatening to the underlying job or it's certainly threatening to the craft sometimes of what they love to do. How do you sort of think this might play out, particularly in a kind of like collapsing cost of AI generated content puts even more pressure on that tension? What's it felt like being in Canva, maybe contrasting to a Netflix type environment where that vestige and that history wasn't quite there in the same way?

46:38Nirmal Govind:Yeah, I think Canva's mission is empower the world to design. And I think the reason they've become so big is because they really helped the average person actually create something good. And without having to use complex tools and so on and so forth. And so I think the mission is still very much the same. It's just that you've got different ways of achieving it, which is you've got AI that can actually help supercharge what you're doing and so on and so forth. So from that standpoint, I don't see a huge difference in terms of what Canvas is trying to do for the world, if you will. I think there are definitely people that don't want to use AI, right?

47:14Nirmal Govind:And they're scared of it. And I was just reading a study from yesterday, something to the effect of, you know, Americans love AI and they use it, but they're also scared about the societal implications, right? Yeah, I think Australia is highest on that. We are some of the highest per capita users and also most skeptical. Yeah. And so that is literally a conundrum that's hard to solve. And I feel like people want more certainty in terms of where AI is going and how this is going to play out in terms of many, many different aspects of economy and society and so on and so forth. I don't think there is a clear answer.

47:51Nirmal Govind:And anyone who says with any amount of certainty, I think doesn't really know. And so that's sort of the problem here. And, you know, at school, parents are worried. It's like, hey, when my kids graduate, you know, they're going to have jobs and so on and so forth. So I think the macro picture is very unclear and it's very uncertain times that we're in. And there's obviously talk about AGI and, you know, is AGI going to come next year and four years? And what's it going to mean? I think a lot of question marks, right? And so I feel like if we provide value to the user and these users come here to achieve a certain task then it really amounts to like how do we do that really well and if users are getting used to you know prompting in the chat box and getting um you know whatever they want done very quickly then you know we're happy to use that type of technology and figure out okay how do we actually help them get more done quicker so that's the way i see it is sort of like you know the same mission but it's basically different tools and tools will definitely evolve over time.

48:53Nirmal Govind:And, you know, we're using the best tools, right, to help the user. Yeah, it's a good point, because I think, you know, for a lot of people, actually, including a lot of Canva users, myself included, like I don't necessarily have the design eye that would make for the best kind of content. And so more AI generated support on that is only of an advantage to someone like me. There's a smaller subset of users for whom that might feel a bit more threatening to the very art of what it is that they do. But for most people, that's going to be all upside. Yeah. And I do think, you know, creative execution, I think AI is getting better at, which is, you know, you can sort of say what you want and it's kind of doing what you want and so on and so forth.

49:31Nirmal Govind:And we saw that in Nago AI too, right? Because we were doing general video. But I think the creative inspiration, creative spark, you know, that's the real question in my mind is like, is it going to be able to like create new music? Is it going to be able to create new shows that, you know? Yeah, what's your hypothesis? You sat in and around Gen AI for a while. What's your hypothesis and where it goes? I feel like, you know, it's going to be very hard to replace human creativity. Is it going to really be able to come up with that type of show or movie that, you know, a director would have thought about, which, you know, hasn't existed in the past, right?

50:05Nirmal Govind:Because AI is, of course, learning a lot from the prior, you know, data. and so I just feel like it's going to be probably the best AI versus the best human, I would say the best human would probably be better. Best AI versus worst human output, probably the AI could be better. So there's going to be a spectrum and I just don't feel like, and maybe this is a visual thinking, but human creativity is going to go out of vogue anytime soon. AGI and all of it will come, but it may give us more time because we can sort of get rid of mundane things that we do and, you know, those things open up more time for you to think more creatively, right?

50:47Nirmal Govind:And then you might get better ideas and, you know, newer things happening and so on and so forth. So overall, hopefully it'll be positive where it needs to be and, you know, still keeping the human creative spark, right, alive. Is there a way that you use AI that you think reflects that in your daily life, like for you personally? I mean, I'm a little bit of an efficiency geek. And so that doesn't surprise me. Yeah. So, you know, it's sort of like, hey, my Slack, I would love to have that automated, right? So I've got, you know, a cloud code routine, which basically goes and looks at all of that stuff and prioritizes it in a certain way and gives me all the links to it.

51:23Nirmal Govind:So I can go do, do, do, do, and then, you know, I'm done with my Slack because I also have a time difference from the US. And of course, you know, when I wake up in the morning, Australia's done a lot of stuff. And so I'm catching up on a lot of things in the morning. So it's just a very useful thing. And And yeah, user recording and all those kinds of things. But yeah, I think it's definitely helped with becoming more efficient with your time, for sure. So actually, I'm interested in the art and science of your own job. A lot of our listeners are founding a company or early stage founders or have small teams, don't necessarily have people in-house with expertise levels like yours, but do have access to these sorts of AI tools to help them do some form of rich data analysis on user behavior or preferences or retention, whatever.

52:06I'm interested what your advice would be for them on how to deploy these tools to best effect.

52:12Nirmal Govind:Yeah, I think the data analysis piece specifically, it's getting there, I think is the best way to put it. And I do think it'll actually probably start getting very good. But so far, I would say like, you know, data MCP servers and so on and so forth, they're not quite able to figure out, okay, here's the right set of, you know, tables to go after, use the right set of things and so on and so forth. What I'd suggest is really figuring out the questions, right? Because usually the problem is identifying the right question to ask off the data. But once you do that, some sort of an intuitive sense of like, you know, it should be around this, I think is useful.

52:47Nirmal Govind:And to develop that, I think, you know, if you're not able to do it yourself or the in-house team, et cetera, I think talking to people is probably what I would advise more than relying on the AI, because I've seen cases where it's like 3x off. And I'm like, I can tell that it's 3x off, right? And then, you know, have an actual data scientist actually go and, like, dig in and come back with the actual result. And I wasn't able to tell it's 3x off, just to correct myself. The answer turned out to be 3x off. What prompted you to ask someone to, like, do the work? It just didn't seem, I mean, I saw the numbers and the distribution.

53:18Nirmal Govind:I'm like, hmm, if it's true, it seems kind of, you know, low for the opportunity that I was trying to go after. And I'm like, this probably needs to be verified. And so then, you know, went and like had somebody like look at it. So a lot of it is intuition, right? It's just like you've seen numbers in the past and, you know, see different places and so on and so forth. So you kind of have that like stiffness, if you will, of like, does this sound right or not? But that's the thing I'm getting at, which is if you don't have people who can do that in house, you can't do it yourself. Then I think talking to people is probably a better way to sort of build that intuition.

53:48Nirmal Govind:And then as you start looking at it more, then, you know, you will start building that intuition. And then you can sort of use AI tools a little bit more because now you can, you know, test it a little bit yourself. You can sense check the output a bit better. Yeah, exactly. Interesting. So I want to move to learning a little bit about, you alluded earlier that there was almost a loss for you underneath at the very point you might be building out your own team. At Mango AI, you had this sort of great opportunity and ultimately decided to go down the M &A path. It tells me that you do enjoy the art of team creation and culture and you obviously did that at Netflix in your team, I think your team ended up being 100 people or something to that effect.

54:28So I'm interested to hear your reflections on, you know, you did work at Netflix during a time where its own approach to culture and talent, you know, became storied, right? Like, you know, various parts of the decks became like decks that every new founder read as a way of thinking about, you know, its own value definition process became a thing that other people copied. As did a few of the ways in which it thought about talent, including the Keeper test. I'd love your reflections on what you saw internally while all of that sort of was going on, while Netflix was engaging in this sort of broader ecosystem debate about what is culture and what does good look like.

55:06Nirmal Govind:Yeah, it's, you know, the culture, I would say, definitely was very unique. And it had many aspects, which I would say was unique for even Silicon Valley at the time, right, which is why a lot of people look at it and like copy it and so on and so forth. Well, famously, it did sort of like no policies on things, right? It was like trust people. If they want to take leave, they'll take leave. Yeah, freedom and responsibility was sort of the short version of our culture deck. And that's evolved, by the way, quite a bit. So now there's a culture memo and you've got new things and so on and so forth.

55:36Nirmal Govind:But yeah, I think the Keeper test was definitely one of the mechanisms. Can you describe for the audience what that is if they're not familiar? Sure, yeah. The Keeper test is this idea that if somebody on your team came to you and said that they were leaving, would you fight hard to keep them? it's a self-reflection question really that you ask yourself and if the answer is no then you should let them go regardless of whether they've come to ask you it definitely ensured a high performance culture that's for sure yeah it's just a hard thing for a leader to actually like do that you know on a regular basis and there's obviously the Netflix side of it which is you want higher impact and so on and so forth there's a human side of it which is like you know how do you do this compassionately and so on and so forth and because Netflix was growing at a very very fast pace when I was there.

56:23Nirmal Govind:It was, you know, I joined probably about 1 ,500 people or so and left when it was like 10, 11 ,000 or so. And our, you know, membership was growing. It was less than 50 million. And I joined with like 300 or something now. And when I left, it was 200 something. And that rate of change and growth, it's a hard thing to balance in terms of like the need of the team, the need for the company and, you know, your own sort of like need to sort of find the time to reflect and figure out you're making the right decision. And these are obviously not easy decisions. So you want to take your time and truly think through it.

56:54Nirmal Govind:And sometimes you're just, you know, challenged for time too. Yeah. So the data scientist in you, what do you think is your secret sauce? Like, you know, there's a lot of people who apply, you know, various forms of like, oh, culture is this part of what we need to do. And we invest in that in order to create an experience for people, but it's also in hiring and it's also in how we do performance management or think about all of those things. What was a sort of, do you have a philosophy that was just you? You're like, I've learned this over my career that I would approach team creation and team building in this particular way.

57:26Nirmal Govind:I mean, it's all about the people, right, that you hire. And I think especially for a startup, that initial team matters a lot. And so obviously you have the founders, but then, you know, the first few people that you hire have to be like really, really solid and you'll sort of set the direction for, you know, culture for the rest. And really, I think things like TAM and, you know, product market fit and all of that kind of stuff. If you have the right team in place, you can sort of figure that out, right? Because then you're having those hard discussions. You're trying to, you know, like what we discussed earlier, you know, this isn't worth investing in.

57:59Nirmal Govind:Like, you know, let's go figure out what else and so on and so forth. So I think that having the right team is very important. And I think it was Jim Collins in his book, get the right people on the bus first and then figure out where to drive it. right so that's kind of the idea is that you get the right people in then a lot of the other stuff just falls in place you still have to obviously pay attention to it and you have to be very deliberate and you know think of all those things um but yeah having the right people on board i think is the key so namar my final question for you what's something you've changed your mind about doing interviews say more no i mean we've chatted about this but yeah i think it was just uh you haven't done a whole lot um yes you have a very low profile yeah yeah i had to do a lot more taking.

58:42Yeah.

58:43Nirmal Govind:So I think, you know, it's just been very internally focused, obviously, in terms of the role and so on and so forth. And I tend to be more of a private person, I suppose. And so that's probably the reason for that. But no, I think, you know, in general, probably in terms of changes on the personal side, I would say it's been more about being more present and like in the moment and, you know, not thinking too long term because you never know what's going to happen. And so, you know, the impromptu trip or, you know, dinner or whatever it is, right? Like being a lot more open to that kind of stuff.

59:18And did something change that led to that?

59:20Nirmal Govind:I don't know, maybe I got older. I'm not sure. Passage of time. Passage of time, right? And so, yeah, so I think that kind of stuff probably on the personal side, because it's also a little counter to what you do at work, because at work, you're definitely thinking more longer term. It's like, okay, what do you need to do so that, you know, you're, this is going to last and, you know, how do you think about strategy and that kind of stuff. And, you know, the personal side is a little bit more like, let's just figure out, you know, today and tomorrow and the next week and so on and so forth. So maybe that's a change, I would say, a little bit more over time.

59:52Fantastic. Thank you so much for joining us on Wild Hearts.

59:54Nirmal Govind:Yeah, thank you for having me. And yeah, I really enjoyed the chat.

1:00:01Thank you so much for joining us for another episode of Wild Hearts. If you want to learn more from other ambitious people building, designing and creating the world that we all want to live in, then please hit the subscribe and follow button. This podcast is produced by Camilla Herring from Blackbird. Our marketing genius is Laura Cofford and our editor is Andy Jones from Colour and Sound Creative. Thank you all so much for listening and I'll talk to you next week.