6P Framework for Success as a CTO: The Future of Technical Leadership with Dropbox CTO Ali Dasdan

Thank you Priyanka. Thanks for having me.

Yeah, as you said, I'm the CTO. It's a very busy job. The company is a very efficient company. Lots of work. We like to move fast. As a CTO, you have to be on top of almost everything. There are people sides of things and product sides of things and operations and process and platform and executive roles, definitely strategy of the company, planning and budgeting and all that. So it's a balance across all of these, so that's how I spend my time.

I think two things. One is definitely coming to the US. I'm originally from Turkey. I came to the US for my PhD, so that was a pivotal decision. The next one was across my career, definitely lots of mentors, lots of guidance and recommendations and suggestions along the way. But one thing I distinctly remember was that I was doing my first internship in a company and I was still a PhD student. I was in the mood of writing more papers, spending graduate student life, and he asked me a question. His point was, what are you trying to perfect? What's your goal with this PhD? Then I realized I couldn't answer him properly. I just answered similarly. So his recommendation was, similar to what many people are saying now in the valley, perfect is the enemy of the good. He said basically you have done enough, talk to your advisor, finalize the goal, put yourself onto a schedule and just deliver your PhD thesis and move on. There's no need to perfect. If you want to do more things you can do after that. Even though it was something simple, it basically triggered and put things in a form that I time-bounded my PhD with the deliverables that I had to deliver. Two things came out of that: one is perfect is the enemy of the good because iterative improvement is what matters, and the second thing is once you clarify a target, it makes it easier to find a path to it and then you just execute that path. Along the way there are bumps and learnings. I don't think he knows that, but he actually influenced me greatly.

Yeah, I think that's a really good point. I think one thing is to realize that it's maybe just modesty, because sometimes what you think perfect is basically defines on a definition. Here are the requirements or here's the definition of what perfect means. But I think the modesty part is to realize that your set of requirements or definition is incomplete most probably because you are still lacking data, there are some unknown unknowns. As a result, it's not perfect actually, to be honest. In certain situations it might be, but once you realize that, you realize basically what matters is how fast I move towards a better state. Better and better and better state will ideally approach what you call perfect. But if you think that you were doing perfect before without realizing that your definition of perfect may not even be perfect, you were probably chasing the wrong target. I think that realization is important, and that's why you have that modesty saying that probably there are things that I do not know here. What makes sense is I make progress towards a goal and along the way I learn, I adjust, I improve things, and that is the far better approach for problem solving.

Yeah, I think one thing is I sometimes use this 6P framework that I've come up with. It says people, productivity, passion which I use for culture, then there's platform, products, and process. Six Ps. If you deep dive into the people side, how you lead your team, how you do succession planning and interviews and onboarding, promotions and performance management, etc., there are lots of details within each of these P titles. As a CTO, sometimes if you are growing as an engineering leader and you go from a good manager to senior manager to director, I have seen that engineering leaders do not realize they think that it's just a natural progression and they can just be a CTO. Being a CTO means you have to know across all these dimensions that I have just mentioned. It's not like you're just going to be a technical person and then somebody else is going to handle that. As a CTO you have to deal with that and you have to have knowledge and expertise and quick decision-making like we discussed. How do you problem solve across all of these? That balance is across all of that. It's not technical versus just the people management. Sometimes you need platform issues, sometimes you have to deal with a very difficult incident, sometimes your guidance is needed or communication or encouragement. You have to be on top of all of these as a CTO. It's a multi-dimensional job as a result of that, and you have to know enough from each one of these areas. It helps a lot if you have done these things before. The simplest thing is interview process. I have designed this so many times in so many different companies that it comes as second nature when I go to another company. Of course you adapt to that company's requirements. But if you say that I'm going to learn on the job across all of these, and these are just six dimensions but underneath there are so many other subcategories of things you need to do, that will take time and make your life more difficult. So I think it's a good idea to have that kind of multi-dimensional experience and learning as much as possible before you jump into something of this nature.

Yes. One is people side, people come first, number one priority, the biggest asset of any company. Second thing is productivity. I tell my team all the time that speed is the only advantage of any company. How fast you move matters a lot. Productivity is part of that. There are so many details across that. I have a LinkedIn article actually I have written about why productivity is also multi-dimensional. Third one is culture, but I am finding a word that starts with P so I call it passion. If you're not enthusiastic and passionate about what you are doing, it affects the culture. That's why I think it's a good word for that. The fourth one is platform. My definition of platform is it helps you to move faster and it helps you to do things easier, not only for you but also third parties. Otherwise it's not a good platform. It's not about the code quality or how the code looks like. It's about does it make you move faster? That's the platform side. Fifth one is the products definitely, the products that you are building or features that you are building on top of the platform. And last one is everything that's associated with process. How do you do monitoring? How do you do incident management? How do you do alerting? All the things around that. These are the six Ps that I try to use. It's very easy to explain and also very easy if you say what you have done as a CTO, then starting with these titles is a good idea. I recently heard somebody added performance, they had four Ps but they were including performance as one of those. I think everybody has their own definition of these, but I thought these are generic enough to cover most of the things that at least I'm doing, and they also cover the scope and the breadth of the work that a CTO would need to do.

I think CTO in my opinion is a position that you have to deserve. Which means you have to go through training and failing and learning and delivering and all that. Then I think you come to a position like that, probably it's applicable to many other C positions. As a result, that learning has to come with you. The technical depth that you should have is you should definitely have a very strong idea about where the future is, what's going on in industry, the developments, because it's going to affect your overall company strategy as you're contributing to it as well as your technical strategy. But also you have to be, in my opinion, a very good architect. You have to know the cloud, you have to know front end technologies, back end technologies, data science, assuming these are going to be in your teams. Operations side of it. That technical depth across all of these you learn by doing. For many of these there's no textbook that you can just go read and learn. If you are in a company that's a fast growing company, learning many things on the job is going to be not good for you and not good for the company. You have to have these kinds of backgrounds before coming to a role like that. Yes, there could be changes with respect to the company and industry, so there will be some new things, but basically that depth has to be there in any discussion. You should be able to see issues very fast because you're going to make lots of trade-offs, lots of decision-making. Almost every day you are dealing with all of these. To be able to do all of these with quick decision-making with limited amount of data, it's going to rely on what you know, pattern matching and things like that. That depth is required to do a good job, and it's across the board.

Yeah, very good points. I think before coming to the CTO role, what's going to be helpful across all of these is you have to know how to communicate. It's a good idea to know how to write well because I believe writing is thinking. It's very important to have that skill. You have to make your pitch or your communication, you have to influence teams and make them enthusiastic. You can do this at any level before even coming to the CTO level, but when you're at the C level, stakes are higher. You're going to go to a customer meeting, customers sometimes scream at your face. How do you deal with that? Sometimes there will be tension within the company, especially if the companies are having growth challenges or other challenges. How do you deal with that? How do you keep your cool? It's very important to have that one team mentality. In the end, all of these stakeholders within the company are parts of the same engine. The engine is not going to work if the CTO side is only good but product side is not there or sales side is not there. You have to realize that and have this kind of partnership with the other team members. Hopefully they're going to follow the same kind of mentality and be a partner to you. Through the action and what you're saying, it's going to come out. Many of these skills should be learned before you come to that level, but of course you won't have a chance to know how to interact with big state customers or board members before becoming CTO. That's okay, you can also learn that. Even when you go over there, in the end they are human beings, so there are some rules associated with each one of these about what to say and what not to say. Also being a really good observer of what's going on around you. At the C level, sometimes you might get a mentor that can actually help you navigate some of these, including some of the board members might act like that or an external one. There will definitely be some learning on the job, but many of the fundamentals should be coming before that.

Yeah, I agree. At that level, let's say you go to an executive offsite and you're discussing the company's future. You cannot just say, most of the time those discussions won't come to technical details. If you are not well-rounded like you said, it's going to be very difficult to contribute to that overall vision and strategy. In the end, you are a problem solver, you are part of that team, you have big stakes with the company. You are a leader that company relies on, and your team and others. To be able to contribute to all of these, you have to be part of that team, part of that activity. That well-roundedness will come as a result of that.

I think it comes down to some of the things that we just already talked about. Decision-making with uncertainty without too much data in front of you. How do you make iterative improvements? This Amazon thing that people say, one-way door versus two-way doors. There are different ways of problem solving or thinking or decision-making. You have to know those skills. How do you do from a target backward kind of reasoning or forward? How do you realize at what point you have to say maybe let's just slow down over here because we need to check X, Y, and Z, or there's too much uncertainty or risk over there and we got to take a little more time? It's impossible to make all of these perfect or correct, but you have to know at least to a large extent where you are. As a leader, we are expected to know that most of your decisions should be correct, otherwise you're going to be in trouble and the company's going to be in trouble. Through that experience and the learning and your leadership and your skills, it has to be there. If I am good technically but at the level of that company decision I make lots of mistakes, then you don't deserve what you are doing. Also, through the stakeholders and the other team members, with an open discussion, everybody feels that it's a place where you can actually be very open about almost every aspect of what's going on with the company. That also helps because it triggers things. In those discussions, sometimes one opinion or point may trigger something else. It's an iterative process by itself. You also have to realize what are the good things that you know, what are the things that are more certain versus other areas that need extra attention. Maybe you have to do a POC for something, or you go test something, or in certain cases you may not even have expertise within the entire team and you need to bring some external expert because of their experience or they have done it before. You got to know all of these are variables or parameters of that decision-making or making progress. Once you have that awareness, I think you make good progress.

Yeah, I was lucky because in the 90s even during my masters I was doing neural networks and part of data science optimization. That was a good background. Even what Google was calling AutoML, in the 90s we even have a paper on how do you create the optimal neural network architectures based on the problem. After that, throughout different companies, I kept that interest as well as research work with respect to data science. In some companies we had to do data science, you and I worked on real-time advertising, lots of data science and machine learning. That was part of the learning. Many people are not maybe aware that machine learning as a technology has been in industry at full scale. Entire advertising or search e-commerce runs with machine learning, and this has been going on for decades even before gen AI came. AI or machine learning was not a revelation by itself saying that there's this new thing. It was part of the life. Now generative AI of course changed the picture drastically. Personally, I deep dived into that and learned that. One thing with generative AI is it almost feels like the first time the worldwide web came, because you can try yourself. N equals one, one person myself, and you could see the benefit for yourself right away. It's very easy to see that there is something over there. I couldn't get this with bitcoin and other things, but web was like that, mobile was like that, and gen AI was like that. You can see the benefit to yourself. It doesn't take much beyond that to convince yourself. With respect to the business side, in the ZoomInfo case, we are a big data company. We believe we have the most comprehensive and accurate data in the industry. We are a market leader. Machine learning AI has been part of what we have been doing for many years now. With gen AI, we jumped at it immediately. The immediate first applications were about summarization, extraction of structured data from call recordings and descriptions of companies. Last year we released a couple of features and products on that. The biggest thing that we did last year was what we call Copilot. It's a copilot for sales and marketing personas to help them stay on top of what's going on in the customer space for them, all the signals and what they should be paying attention to. It works so well not only for our own sales and marketing people but for our customers too. It's already selling really well. Seeing that benefit right away for the case that we are solving did not require too much to convince ourselves that this is something we should rely on and deep dive into. It is part of almost everything that we are doing, and more things are coming. We are converting the entire ZoomInfo into a platform we call GTM Intelligence Platform. It has to have data, first party data, third party data, as well as lots of machine learning, not only generative AI but predictive AI, to find that combined solution. I think it's going to change the game not only in this area but other areas too. It's all over the place. We also definitely have deployed it for our developers, GitHub Copilot and things like that, and lots of internal tools we have created using gen AI, all the applications that are actually automating many tasks that we have and saving so much time and improving productivity.

If you go to our trust center on the web, we have lots of guidance that we honor. Security and privacy are at the top of that. One thing is if you are deploying a third party solution, it doesn't have to be gen AI, it applies to many other areas. What kind of data you are sharing, either customer data or your own data, all of these are very heavily monitored and there is guidance around that and processes. It's very difficult to bypass any of these, and gen AI is the same. I have a recent arXiv paper on how we deployed GitHub Copilot, and you can see over there that even for that, we asked the developers to go through all that AI guidelines and privacy guidelines just to make sure they are not making those mistakes. Part is the guidelines side, part is how do you ensure this automatically using different tools. We have deployed security solutions for any gen AI technology that actually scans and looks at the data, what gets shared and what doesn't get shared. We want to make sure that any of the data is not going to be used for any third party LLM vendors for their training. That is ensured by agreement as well as by our monitoring. You have to be careful about that because that uneasiness is there. The same thing applies even decades ago with respect to predictive AI. Are you using my data to train something to help my competition? It is not something new. With gen AI, you just have to be more careful because the models are far smarter. We are on top of all these, and there's probably so much more coming.

I think definitely AI. How well it works is mindboggling. Through different studies, people have shown that probably there will be some more changes to the architecture. This may not solve all the problems of intelligence, but it doesn't have to at this point. With respect to what gen AI is providing, it's already solving lots of problems, and I will definitely see more and more applications being written on that. Whether you call them agents or not, lots of automation will be coming out of those. Even if it stays at the same state, lots of things will be done with the current technology. I think many applications might get rewritten from scratch using gen AI. Definitely some developments, not only security and privacy, but to ensure that robustness is there so that there is some consistency in the quality of the answers we are getting. Too much randomness will make it a little bit difficult. Perfecting the current technologies is one thing that's going to be going on. But AI and all the things that we are talking about is probably one of the biggest revolutions that we are seeing, and it's super exciting for me personally as well as for the team and the companies that I work at.

Yeah, I agree. That's why it's a revolution. It's going to change lots of things we do.

One is your own ways of managing stress whether you're a CTO or not. There are some things that you can do. Definitely a good family environment, having some fun, a little lighthearted approach, just don't take things too seriously in regular life. There are a lot of political things or whatever, lots of bad things are also happening in the world, so you have to take everything with a grain of salt and have balance from that aspect. From the CTO side, one is just realize that lots of people rely on you, they trust you, their careers depend on you, same thing with respect to companies. I'm not saying this to add more stress, but realizing that you can do it and people rely on you to actually deliver on that is something that you should always keep in the back of your mind. You cannot just get frozen and have a problem in decision-making or moving fast. Same thing with if you take your kid to the street, it's your responsibility to make sure your kid is protected from the traffic issues. When you're a CTO or a parent, it's a similar feeling. The second thing is something that has always stayed with me: being curious and learning and seeing everything with a fresh eye. Sometimes a first principles approach to trying to solve problems extracts fun from whatever you are doing. If you just have an approach to things as same old, same old, yet another boring thing, then it adds up even though these are small things. What I try to do is even if you look at the hiring process, there are lots of interesting things you can do. There's actually research on how you select the best people, how you get good assessment, what does interviewing mean, why you are asking certain questions. All of these actually create more wiring in your brain and I think it also keeps you happier. Having that curious mindset and having fun while doing things, having a good network, and seeing yourself as a student through your team members and learning from them, having a really good environment to go have fun. It requires all of these. There's no single answer, but all of these add up across the board and hopefully it's going to manage your stress. But stress is part of the job in your real life as well as at work. You cannot avoid that, you just have to learn how to manage it as much as possible.

I think staying curious would be my advice. That's a key to get you in a way. The point is you just have to always improve. How do you keep improving if you have a closed mindset? I think that curiosity helps a lot. Having this kind of attitude adds up. You may not end up being a CTO or CPO or CRO, but you can definitely improve in whatever you are by actually staying curious. I think it's one of the key things that you can act on.

Thanks for having me, Priyanka. I admire what you have done so far. You are being curious and I see that you are delving into more and more areas and you've been running this podcast. Good luck with everything that you are doing and thank you for the opportunity.