The “Boring” Things That Actually Make Money - with Atlassian’s Rae Wang

18:49Yeah, that's interesting. Or even indeed for the people on the ground who are using the product who don't, who may not be privy to why their seniors have decided that the product needs to operate in a particular kind of way. I mean, even getting down to the level of like that suggests to me a myriad of instances of any given product and managing those instances that operate. You know, you might have someone with a very particular use case that requires certain features to be shut off or certain other things, but you don't want that to be true for most of your other customers. It must get exceedingly expensive to kind of maintain all of these different instances for these different types of customers.

19:24Are you finding that that's getting easier with certain kinds of engineering developments or harder over time?

19:31Rae Wang:So most of the customers fit into our typical commercial cloud, right? And then we give them the controls to say, okay, I want to shut down this feature. I don't want to allow shadow IT, right? You know, AI is a big one. Everybody wants the initial knob of, I want to be able to turn AI on and off on my terms. I want to be able to turn on and off of particular sections of data, particular sections of customer. I want to decide who can build agents. So those are the typical controls that you expect to go out with any enterprise software. So I think anytime you're building anything for the enterprise market, you kind of intuitively know those are the controls you have to be building.

20:04Rae Wang:Those are the ones that I think are a bit easier. And over the years, the computer science community have adapted the thinking, the architecture, so that these things, they're fairly well-known design patterns to build these things. So it's reasonable for the system. There's a reasonable user experience. You can do policy as code and configure as code and manage them with some certainty. Then we go to the kind of the long tail, right? Then the customers were like, but that's not enough for me. For example, I want my data to be in Japan. I want my data to be in Australia. So then we get into, okay, the commercial cloud with a little flavor.

20:34Rae Wang:So that's things like data residency, right? I want to live in commercial cloud, but I actually want specific deployment requirements. Then beyond that, you're getting to people who are like, I just cannot be in the environment with everybody else. Starting with people who are like, I want single tenancy. I don't want to share any computer networking with people, even though I'm on cloud. So we have things like isolated cloud for that. And then there's the U.S. government cloud, FatRamp Moderate and FatRamp Pi. Those cannot be mixed with the commercial stuff. They have to be their own instance.

21:01Rae Wang:And then towards the very end, you have customers, you know, who we're talking about, like the government three-letter agencies who are like, I kind of need my own region. I need nobody else interfere with me. I am an island. Yeah, I'm an island. I'm disconnected and all that. So those become special environments. Now, the difficult thing is when you have your general, when 99 % of customers, and then we also have to provide those special environments, which, by the way, all cloud providers pretty much do now, how do you ensure parity? Anytime you have an innovation, how do you make sure it flows consistently across all those environments?

21:33Rae Wang:Because often those kind of customers actually use three of the different environments. They have data that's a little more open, data that's really secretive, but they want to have the same experience and the same set of tooling for managing all of them. So how do you deal with that? I mean, surely it's also a case of saying like, we're confident in the 99 % are happy to see a new feature shipped and experiment with it. This customer set, we need to literally ask them, do you want it on or not? Like, is that the experience of kind of going out to those customers? Or do you put it in there and wait for the response of like, no, no, no, take it off.

22:04I don't want that.

22:05Rae Wang:So you don't want to diverge your engineering stack because then it gets really difficult to manage, right? So you do want to, the default is like everything should go to parity. However, you want to give them the control to turn it on and off on their terms with their time. And then you also have to turn it around and convince your engineering team because they're like, we shipped that feature. You're like, but here are the other three environments. You have to make it work. That's the part of the boring job, but it's not building new cool UI. But it kind of has to be done. So you have to regularly bring in the impact to those customers, right?

22:36Rae Wang:Yes, you're getting, you know, a kind of manufacturing to be able to use your JIRA board. But what about the government? What about healthcare? What about, you know, the banks with the highest regulations? Don't you want to impact that part of the world as well? You have to continue to motivate them so that they understand that work is not just boring, busy work I have to do, but they actually understand why an impact leads to. Let's talk a bit about that because when I was researching, one of the people I spoke to said that you were particularly persuasive at getting these kinds of messages through.

23:08And you've also alluded to like they're not necessarily easy messages to sell to the kinds of people who want to be able to say, I built the button, I did the thing. A lot of this sort of work is actually the stuff you don't see in the product, you don't experience even as a user, but it's essential to keeping the product running the way that customers expect it to. What have you learned about persuading engineers to do that kind of work?

23:31Rae Wang:So the first thing to convince them is that no customer is stupid or evil. So a lot of times when you say - Is it hard to get that through the line? It's something you have to tell people. Like when you're setting your own silo, right? If you're focused on, I'm just building a feature, I feel like everything else is an annoyance. And then it takes a little bit of, and you can't do that just when, you know, there's a conflict, you need them to do something. It has to carry out throughout the life cycle. You have to regularly feed them this information so that they understand, like, a customer who has a lot of regulation needs, it's not because they want to slow you down.

24:06Rae Wang:You kind of have to, you know, put themselves in the customer's shoes, right? Like, why is that important? Would we want our airfield controller to randomly ship a new feature without testing? you wouldn't want none of that as well. So you have to gradually help them understand the trade offs the customers are having. They're not intentionally trying to slow you down. They actually have real world needs. And then by helping them solve that needs, your product can have a much larger impact than just with the ones that can deploy right away. And that has to carry out consistently so that by the time they have to come to this trade off, it's already in their mind.

24:41Rae Wang:It's not the first time they hear it. And then once you kind of convince them to have empathy on the customers, then as you're asking them to do that additional work, because it's always a trade-off. I can ship this new cool feature, or I can make it work consistently in each other environments. And you have to help them go a step further, understand the impact with the new feature and the impact to these customers, the impact to both the customers and to our own business, our own bottom line as well, right? So then they understand I'm not forced to do that. They feel like they're part of the decision because it is better for the customers, it is better for the business, hence this is the right trade-off.

25:14Rae Wang:Interesting. Interesting. So I'm interested then to kind of delve back into customers and you, you know, you've spoken about this a lot already, Ray, that Atlassian, obviously, and this would be true for, we already know it's true for Google and Microsoft, have all got large contracts with governments, some of them controversial because of the kind of expectations that are put on those kinds of companies, obviously controversial for some of the employees that work there. How do you think about the role that you play in representing the interests of those types of customers, sometimes where their asks beyond the kind of technical ask, the actual sort of what lies behind, what are they doing with the product, how are they using it, might not be straightforwardly, you know, the sort of thing that every engineer gets up every day to work on.

26:02Rae Wang:When I first joined Google Cloud, it was at the beginning of Google Cloud. Google was actually struggling with, are we a consumer company? Do we actually want to do enterprise? They seem awfully boring. Do we want to get into it? Right. You were early days trying to convince Google that there was a market there amongst business. When I first joined Google, I was told enterprise was a four-letter word there. You don't want to say that. And the mentality was, if we're smart, we build and they will come, right? So I actually went to, Google has hired a lot of famous ex-professors. So I went to one famous distributed systems professor who was a Google fellow at that point.

26:36Rae Wang:And I'm like, do we have, you know, I like ask questions. So I'm like, you know, if the company hates it that much, maybe we just don't do it. Would that be okay? You're like, no, no, no. Like, no matter how much joy you can bring to the small businesses and consumers, when you think about the things that really impact our lives at the very end, you want to make sure. We don't, we, our fires are like, oh, the government. But we really, it is our interest. We want the government to work well. We want financial services to work well. We want healthcare to work well. They might seem cumbersome. They don't move as fast, right?

27:05Rae Wang:But helping them is like how you land the biggest impact you can have with the world. He said they also have the largest scale. So if you want to evolve your service to the highest standard, they are the ones you have to tackle. So he told me, he's like, just be patient. They will turn around because we are smart people. We will understand this is how we leave the biggest impact to the world. And therefore, you kind of have to go there. And so, you know, that also does mean, and you've alluded to the sort of three-letter acronym government clients, that need technology to achieve a whole bunch of different goals.

27:34And sometimes that then lands you in hot water politically about what tools are being used to do what things. I'm sure you followed some of these controversies emerging in the last year or two. Do you have any take on those?

27:47Rae Wang:When I took on the JWCC project for Google, that's with the DOD, I had Europeans in my team quitting, saying, I don't believe in working with the U.S. government. I believe like, you know, if we are so afraid of them or so annoyed of them, we're like, therefore, we want to do nothing with you, then that's kind of like giving up, right? Because you also have no influence in that field. The best thing I can imagine we can do is to influence them in the way that's more kind of aligned with, you know, how we think about the world and how we want technology to work with the world. As you can see, like most regulators don't understand technology.

28:22Rae Wang:So there have been some really stupid regulations like, you know, mobile providers, you have to give me a backdoor to all of your users. Like that's technically not safe to do, right? So like leaving them alone is, it might be easier for us, but that's not the right thing to do for the world. It's messier to getting with them. But if we actually want to have a way of influencing how technology should be used in the government, of how the government should look at technology and regulate it, then it's better to get messy, to be there, to have a seat at a table, to work with them, right? In some cases, they might do things that is not the thing that you agree with the most.

28:58Rae Wang:But the more you understand them, the more you put yourself there, the more you can influence them, guide them, make sure you have your input in how technology is used correctly. I don't believe in just giving up because it's too hard. Interesting. And I mean, I'm sure you've been in lots of instances where, you know, you've had to say, that's not a thing. I mean, technically, we could do that. But it runs counter to the whole philosophy of our organization and how we operate. How do you navigate those conversations? Yeah, we learn from improv. It's never no. It's always yes and. Do you do improv?

29:32I've watched some improv trainings.

29:34Rae Wang:I'm like, that's a really good point. We should all use that. Well, if you're going with empathy, either with your customers or with engineers, your assumption is no idea is a stupid idea. No desire is a stupid desire. So when they desire something, you're like, that makes no sense. The best thing to do is to bring your curiosity. So you don't just shut them down and go, no, we're not doing that. And you're like, that's interesting. Tell me more. So anytime you hear something, you're like, that doesn't sit right with me. That's so offensive. But the right thing to say is always, that's interesting.

30:00Rae Wang:Tell me more. And then as you peel the onion, eventually you connect at a human level. You kind of understand where they're going. Everybody's trying to do the best they can, right? And a lot of times it's trade-offs, it's constraints, it's assumptions. Then you kind of understand, you know, you go to a point where you actually understand where they're coming from. And it's often not just, I want feature A, therefore you build on feature A. You peel the onion and go, why do you actually want it? What is it you're trying to achieve? Eventually, you get to the most basic need and their constraints, their struggles.

30:26Rae Wang:And often, with good product managers, you end up with a solution that's different from the initial feature aid. But that address their needs better and puts them in a much more aligned position with the rest of the technology stack. Interesting. Really great. So then I'd love to come back to sort of internal culture as well. So you sort of alluded to the Google challenge of convincing them that enterprise customers were worth going after, which you successfully did in the work that you did there. You've come into Atlassian. Atlassian obviously has this storied, you know, history of being like one of the true PLG companies, you know, just did this sort of penetrate at the level of the engineer, give them extraordinary tools, allow it to be very cheap and easy for them to adopt those tools.

31:08And over time, you'll get the pressure put on higher parts of an organization that mean eventually they're like, oh, that's right. I guess we maybe need to pay for a license for our engineers to use these kinds of amazing tools. Can you talk a little bit about where Atlassian was on this journey around the opportunity for enterprise and the kind of cultural tension? I mean, did it still exist when you joined Atlassian?

31:31Rae Wang:It still does. In fact, I just launched an enterprise training for APM org. And one of the things we talk about is that it's not product-led growth or sales-led growth. It's bringing the best of the two parts together. This is about understanding of customers, right? With the small, medium businesses or with, you know, the underground business teams, product like growth is very easy for them. They like a product, they go to a website, they buy it. But then you also understand there are customers who actually want a technology that can actually use the technology to benefit their work, to benefit the world.

32:00Rae Wang:They don't operate that way, right? They have, for example, very top-down kind of purchase processes, in which case they can't just go to a website and click, like, I want to buy this, right? They need you to have a sales team. They need you to have solution engineers. They They need you to have partners and that's how their world operates. And if your passion is about helping, you know, to land the most impact in the world, then you kind of understand them. You do what's necessary to help connect them into your world so that they can take your tools and make them really powerful. But it is definitely a challenge to the culture because like at the last time, they didn't used to have sales teams, right?

32:33Rae Wang:Now we not only have to have them. Quite proudly didn't as well. I mean, it's still considered in parts of tech to be great that you've got this like small GTM team. What a sign of the power and value of your product. that you don't need people to sell it. Right. And I mean, the way we need to make this work is to kind of continue to aim the product at the quality, the level of joy that a large portion of the customers, especially small, medium businesses, will continue to drive the flywheel of product growth. At the same time, then we also have to understand there's other half the world, right?

33:06Rae Wang:We have to meet them a little bit halfway. Maybe not extremely to where they are, but we kind of just, we can't expect the government to jump into how we sell to a small mom and pop shop, right? So we kind of have to meet them a little bit halfway to cater to what they need so we can gradually connect them into the best and newest technology as well. And then to the internally, it is a different way of operating. Now it used to be just let's collect usage metrics and see whatever is used or gets doing. Now we have to have the PM to, we have to train the PM to understand sales data and customer data, to mine those data to understand not only it's not only the usage from the initial launch customers, but it's how many seats do you have in the pipeline?

33:46Rae Wang:What is the kind of revenue ramp? You know, what are the large names you're blocking? Some of those logos matter as well. So it is the, that's, you know, what you used to rely on is not half of it. And then here's another half of the data points you bring. The one interesting thing I found is that it is also not clean partitions. They kind of iterate with each other. For example, some of our security products were like, we were just amazed. they've been selling through product growth. Even with the enterprise customer, it's where you don't actually have to actually sell them because the value is there.

34:16Rae Wang:And if you build a product to be easy to use enough, I think at least in the initial phase, they started selling themselves in the flywheel. And then we do understand typically what happens that gets to a certain stage of selling to large ones. Then you have to transition into a more kind of sales guided process. Even in a single product, you have to go through that back and forth again and again. Yeah, which I imagine is also helpful for your sales team to be able to say this is already being used by X number of customers. you know, you're not the first for us. We didn't just build for you, we've built for all.

34:42And, you know, actually, some of your competitors may well have bought the product through another lens is also quite helpful for kind of getting sales over the line. So I'm interested now to move a little bit to the world of AI. You've touched on this a couple of times in some of your responses, but AI has kind of exploded your world in a lot of respects. On the one hand, no shortage of amount of people talking about the risks of getting things wrong or the wrong deployments of AI. And you talked about a great example in Japan earlier. But also, you mentioned the huge amount of opportunity that this opens up for a lot of types of customers.

35:15Let's start with you're clearly really, really close to where enterprise minds are at. What's your current take on what's occupying their minds around this question of use of AI products within their tech stack? And then we can move, I think, to what the future of that looks like at Atlassian.

35:34Rae Wang:So pretty much all my enterprise customers know they have to use AI. They know that's how you compete in the market. They all know it's an existential crisis to them, right? When you think about banking, manufacturing, that is a new way to compete. Unless they do that, their business is not going to survive. So they all understand it. They all really want it. Now they have constraints. They can't just be like a student or a small company turning on right away. They still have to, even though AI is very exciting, their security and privacy and regulation responsibilities don't go away. So they're all in the process of trying to figure out how do I kind of, you know, get a value without breaking my other requirements, right?

36:12Rae Wang:And that is where a lot of, I think a lot of the AI companies are working with right now is we build the functionality first. Right now we all have to build a governance so that they can shine in the largest and most critical businesses of the world. So you don't have to convince the enterprise customers or even the governments. They all want AI. But now we're getting to kind of norming it a little bit. You know, hey, along with AI, you should be able to turn on and off. You should be able to have other logs and have an access management. Those features are flowing into not only in the analyzing products, but even the basic models, right?

36:45Rae Wang:In OpenAI and cloud, they are getting building. So I think the understanding of how enterprise users is maturing and enterprise customers are now going from the initially they all turn it off when they first got it, right? But then they started turning it on in little pilot environments. And now we're seeing some of the first wave of the large enterprises actually turning it on company-wide so that they can start getting the new tools. And the employees can request for those as well. The business team's like, I can't do my job. I can't meet my metrics unless you give me the latest tools. So they get pressure that way as well.

37:16Rae Wang:And I think that market is starting to mature. We're starting to give enterprises the things they need to turn it on. And what do you see as the kind of frontier for Atlassian in this space? I mean, you are very visible as a company, certainly in our market, as taking 20 years worth of technology that's had this phenomenal impact on other technology companies, but also enterprise, and thinking about what repositioning yourselves as an AI company actually looks like. I'm sure you're involved in those conversations to the extent you can talk about it. Where are we seeing the push-pull within Atlassian around that question?

37:52Rae Wang:The question of... Sort of how to evolve the kind of product experience, which has been actually largely about empowering teams. I mean, the kind of history of Atlassian is about collaboration and empowering teams through better tooling. In a world where we've got a lot more AI enablement, like quite literally more agents doing more things, you're thinking about what collaboration looks like, I imagine, with this much wider lens. Can you talk a little bit about sort of product in Atlassian from that standpoint? Yeah, so I think there are two concepts. we now have strong beliefs in, right? One is AI as a teammate.

38:26Rae Wang:So the future of the workforce is going to be half humans and half AI, half agents, right? So a lot of the traditional tools, you're going to have to start re-imagining them as in, you know, what is the experience if most of the things that's accessing them are no longer humans, they are agents. For example, your code repository, right? If the most code is written by agents, then how would you optimize the experience differently to, you know, facilitate it? And also between human and agents, It's like, what does the collaboration there look like? And then I've seen some really interesting questions we, as a society, haven't figured out.

38:58Rae Wang:It's like, for some of our regulated customers, every year to keep your regulation, your employees need to be trained in the latest security and privacy trends. Do AI agents need to get trained now if they're your employees? So I think, you know, the big question of how do we solve the human-agent collaboration, assuming that is the future, the model of how work gets done in the future, that's one big part of it. A lot of our tools is now taking account of like it's agent first, assuming agent human interacting, assuming a lot of agents working on evolving the experience that way. That's one part of it.

39:30Rae Wang:The second part of it is that we strongly believe the value is in the context. So one of the things we talk about is just having access to AI models is not your mode because everybody can buy tokens. So then what is the true mode? And the formula we talk about is acceleration equals intelligence times context. So the intelligence is your engine, but the context is your fuel. You can have a very fast engine. If you have no fuel to burn, you can't get anywhere. So context now more than ever is that differentiation is what RFI customers have as the golden goose, their mode, what their business depends on.

40:09Rae Wang:So we pay even more attention in how we curate, both build, curate, grow your context, as well as how to protect it. So, for example, one thing we introduced is called TimorGraph. So it's not only in the old time, it's like you have all this data, you can find them, you can search and index them. But now relationships matter more than ever. So instead of thinking them as just database schema data, we want to think about them as graph. everything you're working on, like how's your customer feedback linked to your new product idea, linked to your goals, linked to your talent distribution. And those things all have to be related for you to understand your business holistically, understand how to optimize for a business.

40:48Rae Wang:So we put a lot of effort into building that graph, but because it's so powerful, you also have to protect it, right? Make sure it's safe. It follows the permissions, only gets used in the right context. But that entire kind of the richness of the context, and it's secure, it's protected, we think is a lot of value in the SaaS business now. Bringing two themes together on that, though, I mean, you alluded earlier to data sovereignty is, you know, a thing that we weren't really focused on five years ago. And now it's like every conversation, political conversation, it's on the front page of the news, you know, the average Australian is reading about data sovereignty, which was not a thing.

41:24But also the power of context. And of course, the more, you know, to use your language, the more fuel you have, the further you can get with the intelligence that you might be able to apply. How do you see this tension playing out in the medium term between customers that are saying, well, actually, because of the value of data, we want to keep everything extremely locked down, you know, AirGap being the extreme example of that. But at the same time, we also know that the more access to more context that we can get, the more we can achieve with it. Like, do you see that creating a tension that will bring us back to a more globalized sort of data world?

41:59Or do you think actually we're going to continue to create more and more fortresses and find new and different ways to get that context from outside of our immediate organization that gives us an edge against our other competitors?

42:12Rae Wang:That's where kind of understanding your data, treating your, you know, having the nuances there, treating different data differently makes sense. And then from a, you know, SaaS company perspective, what we need to do is to give customers those guardrails so they can kind of, what you want them to do is to be able to describe this data needs this level of protection, that data, you know, to treat them differently and then have some help in figuring out how to get a maximum intelligence without sacrificing security. So for example, by default, our graph is for an entire company, but then we have customers where like, I have many clients, I don't want their data to intersect.

42:46Rae Wang:Like imagine if you're a consulting company, you have a hundred clients, right? You're like, I wouldn't want customer A to do a search query and find out data about customer B. That just wouldn't be right. It's very timely that you should mention that, given what's in the news at the moment about sometimes that going wrong by humans. Right. So then we are shipping a feature that's in early access right now by the men of those customers is to give them that next level of scoping. So it's called units. It's like, okay, we recognize you have an organization, but within that you can partition. It's called units.

43:16Rae Wang:And then we scope the search and the graph and all the AI capabilities, including all the third-party apps, to that scope. And then there's no interference between the scopes. So no junior consultant could accidentally learn something about a competitor that might get used. Interesting. And so within your unit, it is a fully connected, intelligent, collaborative sharing experience. However, at the same time, you can also guarantee there's no sharing between the scopes. So you kind of have to give them those guardrails and then they can define where they want to maximize sharing. It's always a tradeoff.

43:50Rae Wang:Where do I want to maximize sharing? Where do I want very clean partitioning so there's no leaking going on? Yeah, very, very interesting. So for early stage founders that might be thinking like my market is enterprise, like I'm not coming out with a consumer ready product. I'm not even attempting PLG. I'm going to go straight out sales like growth or enterprise customers. a lot of what you've described takes a long time. The architecture and the infrastructure to make it all hang together and work, even with AI support, takes time. How would you suggest they sequence and prioritize this experience of creating reliability and trust in their product so that they can move as quickly as they want to without sacrificing the exact things that you know customers really, really care about?

44:38Rae Wang:So amazingly, I've now seen quite a number of AI startups that target enterprise first. What I got from them is that there's very little money from consumers for AI. They're not willing to pay for AI, right? The way you get money out of AI is by selling to the enterprise. And so a lot of them actually understand it. That's where the money is. So I've almost seen a larger proportion of AI startups that's targeting enterprise than I traditionally saw among startups. And one way to start, I see some of them already doing, and even Lazen in the early days did that, right, is to when you're early, when you're small, when you're just pushing your functionalities, you're not quite a big platform yet, you build an existing platform, right?

45:19Rae Wang:So we have, you know, Lazen has a marketplace. We have many startups that initially start building our marketplace and they can use our tool, our data management tool. So they don't have to worry about these things. It's like somebody else takes care of you. even ourselves, we initially, our cloud was initially built on AWS, right? So that saved us a lot of the initial work. We can use a lot of the proprietary service where they already have these controls, these scalability. Now, as you get bigger, let's say if you become successful, right down the road, you also will need to think about, okay, I don't, you know, I, for example, we went to multi-cloud, right?

45:53Rae Wang:So I imagine a startup when they become very successful, very large, they're like, okay, then I have to consider scale. I can delay that by being a single platform or single provider. Eventually, when I get to, I need to scale beyond that, then I can come up with a different strategy. But that buys you a little time and lets you prove out your product market first, right? So you can leverage another existing platform to satisfy those requirements, get a functionality that is proven, then allows you to solve some of these problems, scalability, reliability problems, or build your own controls gradually.

46:23Rae Wang:So you can tackle them at different times. You don't have to do them all at the same time. So sort of stand on the shoulders of giants, as it were. Yeah, interesting. And one of the things that's sort of struck me, and I think this is true outside of your landscape, but certainly true in yours, is AI does make it possible to write a policy, write me a policy, a data security policy or something in two minutes. Something that previously used to take a person time to do or do well, or you pay a consultant to do it well. In a world where you can use AI to help you on the surface appear to have all of the things that enterprise customers are looking for.

46:57Are you finding that where enterprise customers now push technologies on the trust point has shifted from like document for me that you have all of these things, which is now cheap to do, to like what does it actually look like? Do I believe you are a provider that will be able to understand the nuances of your SOC, you know, or whatever, you know, documentation?

47:18Rae Wang:Yeah. What we actually realized is that UI and functionalities are easier to do with the assistance of AI. The enterprise features are much harder. You can say, hey, build me a project manager tool. You can't say, and please make a fat ramp ready. You can't say, and get Gira to scale to 100 ,000 seats. And AI is not at a point where it can do that yet. So AI is very good at the app layer, right? But for the platform layer, for the enterprise requirements, that's where it hasn't quite gotten to there yet. I do hope. So there's a moat there, you think, at the moment. There's a bit of a moat there.

47:52Rae Wang:So I actually think if you vibe code apps, you can probably satisfy some small, medium business needs. But the larger customers, both the stuff they need is not as easily AI codable. And also, they will need a lot more proof. They will need a credibility to understand it, especially when we talk about regulation. Like it goes through the human process of getting certified. And we don't have a path for AI to go through that yet. So right now, that's actually the harder problem for AI to solve compared to the basic functionalities. Yeah. I want to get to sort of your perceptions on culture. So you were sort of famously at Microsoft sort of building collaboration software in a culture that was said to go through an era of the way it was described online is sort of punishing collaboration through stack ranking and other ways of sort of understanding who's, you know, achieving what in the organization.

48:43I'd love you to share a little bit about your experience being there at that time and what it taught you.

48:49Rae Wang:Yeah. Early 2000s, when I joined Microsoft, that was the time it was the well-known kind of really aggressive, almost cruel culture, right? I remember the first week I, you know, I just moved from Australia to Seattle. First week I got in team meeting, 20-something people. Other than me, there was one other woman there. And then the leader of the group just pointed at her and said, I don't want to listen to you. I have a stupid idea. And she cried right now. And I'm like, what did I, should I just pack up and move back? And that was not abnormal back then. The stories back then, there was a lot of swearing.

49:20Rae Wang:People threw chairs out of conference rooms when they were upset. People punched a hole in the wall when they were upset. They kicked the door and the door fell. Those were all people who didn't feel bad about it. They were like legends that people talked about. It was like, look at how great it is. Passion was a creative tool which is going to handle very aggressively. People were like, they were openly aggressive. There's a lot of swearing words and all that. And of course, some of the things Microsoft did back then was also, they believed to get the best results. It was a normal practice to tell three teams to do the same thing and let them internally compete.

49:52Rae Wang:They're like, there's going to be competition in the market anyhow. We might as well do it internally. Let three teams compete and see which one wins. The intention was okay, but the end result is you have internal teams were like, if you win, I will lose, right? So you're getting this like zero-sum game type of game. And that didn't help with collaboration. and I remember one of the early anecdotes from Microsoft was when we did performance review calibration. You want to understand which team members are more valuable. By the way they did it, it was almost too direct. It was like, let's imagine your team is in a lifeboat and it starts leaking.

50:24Rae Wang:Which one would you throw out first? There was the right intention. We want to get to some kind of understanding of who's valuable, but the way it was framed, I think that primed people in a way of thinking it's always a survival game. It's always me against you. But that was also the early 2000s. Both Microsoft and industry over time changed. As you said, exactly, if we're building collaboration software, if we're building software that help people to connect with each other, we're both at work and outside of work, then we can't always be living in the jungle because then our products are going to help people do that.

50:54Rae Wang:And as you can see, the early social media products were kind of like that. They're like, do whatever you want. We won't censor anything. Go stalk whoever you want. Go defame whoever you want. And then very quickly within industry realize that's not right. And that doesn't help either the society, doesn't help our business either. So I think it was interesting, like the product ideas help the culture transition as well. Both we realize it's not a good way for people to work, and we realize our products can't do that either. I think 2010 and onwards, you know, when Microsoft got to the Satya era, but also when industry overall understood, like we actually have to protect people, that's how the tone changed, right?

51:31Rae Wang:So now, in a lot of places, the stack ranking is a little more downplayed. You encourage teams to collaborate more. In fact, I know, you know, after I left, I heard some of the Microsoft performance review sessions now talk about, like, you have to talk about how you collaborate with other teams. Without that, you don't get a good review. So there's quite a bit of turning in the culture. And I'm really happy that happened because otherwise, I don't know how we all could have survived. What was your survival technique? I mean, you were very successful through that era. Well, see, I went through being a poor immigrant child in my teenage years.

52:03Rae Wang:So I was always like, respect the opportunity. And if it's hard, just live with it and go through it. And initially, it was like that, right? Initially, I felt like, you know, I knew the least. I'm still learning. So I should just put up with whatever. And then, you know, as you go through the years, I started to understand you can't just put up with whatever, right? You've got to speak up. You've got to have an environment that works for you. At the same time, the industry, I think, also changed how people collaborate. So we arrived at a good spot together. Yeah, that's interesting. Because you then moved to Google, which sort of famously at that time was quite a different culture, right?

52:37Like it had this really interesting era of investing heavily in people's science. And, you know, you had a lot of people. Laszlo Bock was there, of course, around the same time that I think you were there. And a lot of this discussion about what it means to be a Googler and all that sort of stuff. Was that a very big cultural shock for you or was behind the scenes it actually not as different as it seemed on the outside?

52:57Rae Wang:It was a very big difference. In fact, all the Microsoft people who joined Google were warned, don't bring the same attitude. You won't survive here. Cleanse yourself before you arrive. Yeah, people are actually very nice to each other. It was a very nice environment. But after a while, you also realize every style has its pros and cons, right? So at Google, people were very nice. But as a result, decisions were extremely hard. You can't decide on things. It's very hard to point out something was wrong. Like everything became a committee. You write very long pages. Whereas Microsoft, it could get really sad sometimes.

53:30Rae Wang:But at the same time, if a decision was made, you know exactly how to get a decision made. If it was made and it was top-down enforced, it was very cleanly executed. So you're like, then you get into, you know, how do we do kind of system engineering on the organization, on engineering culture to bring the best of both worlds. And that's going to be the difficult thing to solve. Okay. So talk to me about what do you, like you now are in a position at Atlassian to shape some of all of those things, bringing all of those experiences together. what's the rave version of that at Atlassian? So at Atlassian...

54:00Are you throwing things out the window occasionally just to keep people on their toes?

54:03Rae Wang:I haven't seen anything being thrown yet. So Atlassian culturally is actually a little more like Google. By default, people are very nice. I think it comes... You and the product you build are very similar, right? So it comes from building collaboration software. The default is very nice. And also being an Australian company, people appreciate that niceness, that collaboration. Haven't had any swear words being thrown at me. none of the early Microsoft experience, it is a much more pleasant environment. Now, then the thing you ask is how do we ensure there's still velocity, right? We can still have the hard conversation, making the hard decisions while we're doing that.

54:35Rae Wang:And I'm one of the more vocal people in the company. I go around and ask stupid questions. But sometimes you feel like you just have to nudge it a little bit. So then, you know, we do talk about like, how do we simplify the processes, right? Lazian is very writing-based. We write a lot of pages. But a lot of times you feel good about writing stuff and gathering data, but that doesn't push you to ask the hard questions. So how do we balance the overwhelming information? We want to give people a way to express themselves and still how do we make the hard decisions? So that is the thing we continue to evolve, continue to iterate out.

55:04Rae Wang:And I think that's also why Atalazian in the last few years hired some people from the industry, from different companies to bring, you know, cross-pollination is always good, right? To bring different ways of doing things. And we want to lead industry, not only in our culture, but in our products, in how people collaborate. So we have to be very open-minded on what are the different ways to collaborate, what have we tried, what are the proven evidence that we have seen, and how do we synthesize them into a model that works for us, but we can also put into our product. I mean, you've been described as stoic and direct.

55:35You've even described yourself in this interview as direct. And someone told me that you were particularly good at holding your own with heavy weights. You've been in very founder-led DNA companies. What would your advice be for others that are going into those sorts of environments to find the navigable path on asking the hard questions while also sort of respecting the culture and the way that it operates.

55:57Rae Wang:I borrow a concept from distributed system from storage. It talks about eventual consistency. Usually people are like Buddhism or whatever. But so you have multiple copies of your data, eventually they become consistent. But then I'm like, the universe is eventually consistent. I believe eventually the right - That's a controversial statement at this very point in time, seeing where the world is, right? It is often said you're going to have collateral damage while you're waiting for that eventual consistency. But if you see, like even looking through history, the pendulum will swing. The wrong thing will not always be true.

56:30Rae Wang:It will always swing back, right? Although there's price to be paying in the process. But I just always have the belief that the right thing will always eventually be the thing we do. There's eventual consistency. So then you want to have your strong conviction and you want to continue to be vocal about it. But you also want to, it is important to pick a battle. You don't want to die the first battle. You never get a fight for the eventual thing. But you kind of have your belief. You believe in that. And then you look at, given the constraints, like where can I move a little bit first? You never give up the hope, but you're also practical about like where can we push a little more?

57:04Rae Wang:What information do we now have, right, that can help us to make a little more progress in what fronts? You fight a battle on many fronts and you build allies. So there's the, what do they say? Your vision is stubborn, but your execution is flexible, right? That's how you get yourself not frustrated. That's also how you know you're not forgetting about what you're trying to get to eventually. And then the other thing I think about is that, I mean, what is the point of life, right? In the end, it's like it is to understand ourselves better, perfect our own character, and do that for our teams as well.

57:34Rae Wang:So not just ourselves, but also the community, the team as well. But then you're like, everything that stands in my way is an opportunity for us to understand ourselves a little better, how we work a little better, and to grow our skill set in how we kind of work towards the eventual goal we have. Then you don't get frustrated. And the worst thing is you're so passionate about something and you get frustrated. You leave it. Then you never get to contribute to it, right? So if you can understand different ways in your organization, gain value from these experiences, then you don't get frustrated and you keep making your movement gradually towards that long-term goal.

58:05It's interesting. I don't know whether you intended it or not, but I see such a beautiful coming together of your earlier statements around the curiosity you bring to seemingly irrational things that your customers ask. It's sort of that self-curiosity of like, why am I finding this so hard? Why is this so hard? Like stick with it. You know, the impact is there if you do and apply a sort of constant curiosity and pragmatism. That's really cool. I did want to move through to a statement that someone told me. They said to me when I said I was interviewing you, they said, oh, she's a woman in tech, like actual tech.

58:38And they made a point of saying that. And I thought, isn't that interesting at this point in time? And there's a lot of discussion about all manner of challenges that exist around the absence of women, certainly in certain engineering roles. You came up through the engineering profession and have now fashioned this career at the intersection of sort of engineering and product and go to market. I'm really interested in your reflections. I'm sure a thousand people over your career have said to you, you must be so rare, you know, a rare woman in a senior role in technology. I'm actually interested in how you react to that framing because in my experience speaking to senior women, there's either those that sort of eschew it and say, like, don't think of me as a woman.

59:18I am my job and I am my role. And if I happen to be a role model for others, then fine, but it's not a thing that I sort of step into. And then you have other women who have taken it on as a very distinct part of their identity. I'm interested in where you've landed on that question.

59:33Rae Wang:I definitely went through a phase where I'm like, no, I got here by my own merit. Don't think of me as just a woman because there are people who will say, oh, maybe you got there because we needed a diversity quota. That does come up sometimes. And I talk to a lot of - Have you experienced it directly? Yes. And I've talked to a lot of other women who deliberately want to downplay being a woman because of that. Everybody's like, I want to be known I got here by my own merit. Yeah. Right. I kind of came out of that phase. And now I don't hide away from that. I still got here by my own merit. We can go toe-to-toe debate technical issues.

1:00:08Rae Wang:I have that confidence. I can go as far as you want to go. Is it not true that you also studied at the same time as Mike Cannon-Brooks? I wasn't in the friend circle. I had the wrong friends. Mike told me he didn't go to classes. I'm like, oh, that's what I did wrong. I studied in the front row, did all my homework. That's why I didn't meet the right people. But I went through computer engineering. I was very proud of being an engineer, right? doing the hard maths and physics and all that. But now I don't hide away from that. I actually want to say, yes, like, you know, women can be like me because I realized it actually hasn't gotten better from 20-something years ago when I was in a university, right?

1:00:40Rae Wang:It's gotten worse. The percentage of women who get interested, who get into computer engineering, and also the percentage of women who persist through the first few years of early career years, it's not gotten any better. And I realized when I talked to some of them, to them, I didn't realize. I always felt like I'm the small chicken and I'm just like, you know, still learning and getting all this help from a mentor. Now I realize I finally got to a point where the more junior women who are just starting, like just me being there saying, I'm also a woman. I went through that. So you can as well.

1:01:10Rae Wang:That's how you do that. It gives them so much value, right? I didn't realize that when it dawned on me, then I'm like, oh, then I'm going to always say that because by being here, by myself being here and say those things, I can encourage so many junior women, make them feel like there's actually a path for them. Some of them have come to me and go, like, I was thinking about giving up. Then, you know, I heard you talk. And I heard, you know, you came to our team. Like, you know, you had a one-on-one with me. And then I was really encouraged. So I'm going to persist. And now I realize it's not just about building products.

1:01:39Rae Wang:Like, building the next generation of women leaders is actually important. And then the other thing I connected from that was it is not just about women. I would say, like, a lot of people in the industry feel like I'm a little different from the rest of them. Maybe a race. Maybe a gender. Sometimes maybe, like, I don't come from a technical background. Sometimes it's like I didn't work in these large technical companies. Sometimes because I changed discipline. I had people, you know, in my team who were like, used to be in our strategic business ops and they changed to product management. They're always like, but I didn't grow up as a product manager.

1:02:07Rae Wang:So a lot of people feel like I didn't have the same pedigree as the mainstream. Yeah. Can I still succeed? And I think rather than talking about just being a woman, it is about the, as an industry, we need to absorb and encourage people with different backgrounds, with different strengths. And that's the only way we can build the right products. because in the end, our products are used by people from all walks of life. That's right. That's right. You know, in many respects, the sort of diversity angle is cleanest where you just say, like, mimic your customer base in the case of being a company or if you're a government, you know, like, who are the citizenry?

1:02:38I mean, you'll do a better job of delivering to their needs if you understand their circumstances. You did say earlier that things haven't necessarily gotten any better. I'm sure you've observed through your career a thousand different ways, but even just sticking on the women in STEM question, let alone the kind of broader diversity question, I'm sure you've seen a thousand attempts at various well-meaning programs that just haven't really delivered the goods. Do you have any left field ideas you'd want someone listening to this to pick up and say, hey, maybe I can give that a crack?

1:03:06Rae Wang:So what I see didn't work was if you hold a different bar, right? So for a while, everybody was trying to hire diverse workforce. And then some of them lower their bar to hire that. And that would never work. You hire those people in. They feel like they don't get there by their own merit. They're discouraged. and then they don't measure up to the rest of the people and everybody else, we can't trust them to do the important work. That never works, right? The real way to make them work is to, it is fair to give people who are a little more disadvantaged where the minority is a little more help so they meet the same bar.

1:03:34Rae Wang:That works. That's sustainable, right? But at what point do you do that, right? Because some people would say, oh, you could do that after you hire them, but you still run the risk that people in their teams distrust whether or not they were, you know, it's like, oh, you were hired on a lower bar, but we're helping you now to get there. or is it a sort of before the hiring? I think even before the hiring. You have to make people feel it is, it has to be fair, right? Everybody is hired, has to meet a buyer, has to be expected to perform the same. That's how you set them up for success. So when I was at Google, I was on the PM hiring committee and one of the things we discussed and we started doing was to say, if we see a candidate coming from a non-mainstream background, we would have office hours and offer them interview prep, right?

1:04:14Rae Wang:And that is fair. That is helping them to meet up because that's one skill they, it's not just they use it once. It's one skill they actually learn. They actually make themselves better. You know, a more mature engineer are more likely to succeed in the workforce. And that's also why we have, you know, we have mentoring programs. We encourage women through like the Women in Tech Award here. In the US, there's the Grace Harper Conference. I do believe in providing these supporting forums so that you help the people who have a little less, who are not, who don't enjoy the same benefits as the rest of them to have a chance of catching up.

1:04:46Rae Wang:One other thing I learned was in U.S., the colleges, the universities, they've been experimenting how to teach computer science. What they learned initially, that was my experience. I didn't program before I got into university. It was, I loved math. My parents wanted me to do medicine, so we compromised on computer engineering. It's like you have to do something that helps you find a job. I'm like, I like engineering. But my first year was very traumatic because a lot of my male colleagues program for years. They all had computers at home. I was a poor immigrant kid, never learned programming.

1:05:14Rae Wang:and I thought about giving up because they just seemed so much smarter than me, right? What the U.S. colleges learned was that they experimented. Some of them had like, I think it was like a golden path and a green path, but they separated people who have had prior experience from people who haven't. They just taught them differently because, you know, where's the different starting point. What they realized is after a year, there's no difference. Really? So it really is about what you can provide for people to catch up, right? Once they've caught up, then they're given equal opportunity. Then they don't perform any less.

1:05:43Rae Wang:So I think that really is the proven pattern. So instead of thinking about how we lower and buy, we should never do, but we should put more of our thinking into like, and it doesn't even take that much. It takes a little bit of intention, a little bit of effort from the community. We can do a lot in helping people to catch up. Once they catch up their lunch, then they're setting good paths. Amazing. So, Ray, final question. What's something you've changed your mind about? What's something I've changed my mind about? So one recent one. I'm a very practical person, right? I like to build stuff that I know will succeed, that will win customers, that is useful to the market.

1:06:19Rae Wang:Sometimes, especially in these founder-led companies, they invest in a very crazy idea like that would never work. So I was always bothered by, I'm like, why are we putting so much money and so much people on things that would never work? That's just a stupid idea. Why we could do all these things that are actually useful, actually helping. Like, I love helping customers, right, solve their pains. But we're delaying solving their pain because we're putting our research into doing this, like, really stupid thing. like sending a balloon into the air to downcast signals. So then one thing I learned recently, which I think now makes so much sense, is that sometimes you do products because you know they help customers right away.

1:06:53Rae Wang:But as an organization, you need to continue to learn. How do you learn? You don't learn by doing things you know will succeed. Sometimes you know something is, the chance of it succeeding is very low. But your organization needs to learn to be able to tackle these things. And often what they develop through that process, both the technology and the mindset and the product strategies, they can take back into their mainstream products. So it is necessary for an organization to always ring fancy or 20 % to do the crazy things. Even some of the things you're like, that just make absolutely no sense.

1:07:25Rae Wang:Now I stand behind. I'm like, yes, it is important. What was the watershed moment on that? Because you've been in this universe for a long time, these founders with these sort of seemingly harebrained ideas. Why was it just now that it sort of occurred to you that there is some value in that? So the AI wave, I think, spoke a lot to that, right? Like everybody's building AI features, including Atlassian. Some of the AI features, I'm like, that makes a lot of sense. I'm like, that will never work. And I think partly because now I'm at a level where I actually own some of the resource distribution, some of the investment decisions.

1:07:54Rae Wang:So it is more relevant to me than when I was just like building my own product on the side. And then when people explained to me, and when I actually saw in real practice some of those products that didn't succeed and we didn't expect them to succeed. However, the things we build, the muscles we build through them, the technology we build them, heavily benefited the other products that were being used in mainstream. Now I get it. And from now on, and also, you know, once I get it, I'm like, okay, how do we make sure it works well? Right. So now I also understand, you might know the book, The Innovator's Dilemma.

1:08:27Rae Wang:It is very easy for us to overbias on things that we know is the, you know, the main business for the company. It's almost like we have to intentionally create a little partition. You have to intentionally reinforce the resource so we get to do the speculative stuff. Otherwise, it would never happen. Now I'm a full supporter. Now I'm thinking about how do I make sure not only do it at a company level, but even within my team, every organization needs to regularly exercise that muscle, right? So how do I always sponsor my team to make sure myself, my own time as well, we do things that we know will fail?

1:08:54Rae Wang:So that's important to a company, to a product, to our organization. Ray, thank you so much for being on Wild Hearts. Thank you. It's been really enjoyable.

1:09:07Thank 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