Latest / Digital Leaders Club By Ingeno / Jad Yaghi, CTO at Baselane | Engineering Leadership & Product Strategy
Transcript
- Ingeno: My name is Guillaume Dichano, I'm CTO at ⁓ in Geno. So ⁓ today we'll we have with us ⁓ Yad, who's gonna be talking about ⁓ yeah, bridging the communication communic communication gaps between ⁓ you know technical folks and let's say more business folks, I would say. So you correct me if I'm wrong, Yad, in a few minutes. ⁓ in our discussions with Yad, ⁓ we we realized that Jad is pretty good especially at at two things. at least in what we observe. ⁓ first is feeling fast and hard and moving on. So this allowed him to quickly prototype, experiment and iterate his way to solutions to painful problems on both tech and people's sides. And second, he knows what it takes to build large, cross disciplinary, effective and lean teams. ⁓ and through the combination of those two skills and the many more that he has, of course, he has been able to create Multiple tech startups such as Vero creating one of the first three D engines in a browser or Project Beagle a system in the tourism industry. He's currently building a next generation next generation bank, which is quite something, I would say. Quite an undertaking. ⁓ specializing in in serving ⁓ the needs of tenants and homeowners. Yeah, they seem to be a firm believer that technology, for the sake of technology, is kind of a waste. There has to be a purpose, a framing vision to guide its application. Thank you, Guillaume. Hi everybody. I I appreciate you making the time to come here and and and listen to my stories, I guess, and and my ⁓ experiences built over a lot of failures. ⁓ feel free to be curious, feel free to disagree. I love disagreements. That's ⁓ the friction is where where the real stuff happens. some more context about why I I feel a little qualified to talk about the interaction between engineering teams and and non-engineering teams. I have built multiple teams across over ⁓ sizes of over a hundred people across multiple startups and a couple of companies. My usual way of working since I was ⁓ really started my my career is I'll build a company for a couple of years, burn myself out, go in the industry, work for one or two years, get bored, go start another company, burn myself out. And I've been doing that for the last twenty-two years now. It's been it's been ⁓ about six or seven startups that have been funded, a lot of prototyping in between and ⁓ you know, various jobs between ⁓ coding and later on in engineering management and engineering leadership at at several several larger companies and and it's always been fun to work at the very edge where where technical technical capabil capability meets pretty much intersects and meets everything else. I have also raised over thirty million dollars ⁓ in investment funds for ⁓ various startups. So I've talked to a lot of investors, I've done a lot of pitches, I've been rejected many times and end of the day, I got folks to open their wallets for me. And I have shipped multiple dozens of products, the majority of them from designing them on You know, quick whiteboards to Figma to production to users' hands and all the iterations in between. I'm also probably the only Lebanese you'll meet who does not speak French. So I appreciate you listening me to me talk in English. My wife's a French teacher. She has tried multiple times, but it just does not take on me. So I apologize for not able being able to speak it there have this in French. I've also failed multiple times. ⁓ I've failed at three companies. big failure. I've had to, you know, after hiring dozens of people, have to shut things down and let them go. I've burned through millions of dollars for my investors, not on purpose, you know, back at the time they all seemed like great ideas. Unfortunately, a great idea is nothing if you don't execute and there's no market for it. And I've I've, as I mentioned, I've created tons of products that didn't work. and in hindsight, now, you know, I cringe thinking about it at how stupid my ideas were or or or the ideas I'm trying to execute on. The flip side here is I've learned a lot of lessons in doing that. And I make tons of mistakes and I will continue to make tons of mistakes. So what are we covering today? I am an engineer by training ⁓ and and and by experience. However, as of necessity from starting my first company, I have found myself having to also be the person who's fundraising, the person who's selling, the person who is figuring out what to build and what order to build, figuring out how to get our users until we're at a point where I can afford to hire people who are smarter and more experienced than me. But I have ⁓ for a lack of better word, been there and done it in in in in several industries. And ⁓ what I'd like to talk today is first I'd like to create some empathy for the engineering role and technical roles. And then we know we'll move on. Once we have built that empathy, we want to move on to talk about the friction that rises between the engineering and non engineering roles. I say non engineering roles, there are many, we'll cover a majority of them, I think. And then we're gonna talk about once we we have the a good understanding of the friction, I'll share some of the tools that I have used very successfully after finding them through lots of pain and effort to be able to to bring alignment into the multiple teams. As an engineer, it's very easy to stereotype. I promise I will try not to stereotype. It is very difficult. It's easy to put things into buckets and organize them. So let's start with a stereotype. This is me when I started my career. ⁓ I I was a typical nerd. I l I didn't care how I looked. I didn't care what I wore. ⁓ you know, in the morning I grabbed the first shirt that ⁓ that that that didn't smell stinky and I just got to work. And I I loved what I was doing back then. And back then I was working ⁓ building a a call center module that allowed a caller based on their region of the world. It would it connect them to a caller, ⁓ to a call center agent that ⁓ spoke their dialect. or or the closest to it based on which area they called from. In the Middle East, this was was pretty groundbreaking. So that meant somebody from Saudi Arabia would get an agent that so spoke an Arabic accent, a Saudi Arabian Arabic accent, et cetera. And the product was very simple. We had an admin panel. We had the warehouse call center, warehouse user who would would make sure all the all the goods are updated. We had the call center agent and we had some connections to telephony systems and we had a database for persistence. And we'd write our code We'd compile our code. If it compiled, it's working. And ⁓ you know, we'd ship it either through CDs and and you know, ⁓ I remember some days we'd we'd FedEx some ⁓ some USB and some dongles. And that's how we shipped code. It was very simple, very clear cut. You got your business logic right, you QA it little bit, you test multiple scenarios, and you were good to go. No overheads, no complexity. C was was what we mo you use mostly for for a lot of our logic. We used some ASP or PHP for some of the ⁓ you know, was really cutting edge building some browser stuff, but other than that, compile and chip. And life was very simple for me and many of my engineering colleagues around us. Understand the business logic, implement it, and you are in business. Of course, we come 20 years later, and this is a recent project I worked on. And I have to admit, I still don't understand what half these boxes, how they work. I have a good level understanding of what they do, but the complexity inherent of most products today on the market. That rose from a need of being able to deploy code faster, to be able to write code faster, to be able to integrate with lots of services out there on the market. The complexity has multiplied by multiple magnitudes of order. ⁓ this was a simple e-commerce ⁓ customer that we were we were kind of taking taking most of what they built and and and and refactoring and rebuilding it to make it more efficient, because the way they built it had inherent difficulty in shipping a lot of their product lines. So what happened, I I think. Because of a lot of these tools, a lot of complexity grows. You need to not just understand what are you coding, but you also need to understand how you're deploying it, how you're talking with with other components. And and and that's why the complexity grows. And today, your typical developer is no longer writing code only. We have I I'd I'd estimate most engineers I know write spend between twenty and thirty percent of their time writing code. The rest of the time is either planning, trying to understand the business nuances of what they're trying to do. You have all the various builds, you need to test your code, write your test scenarios, then you need to release it. You need to understand how you're deploying it. Deployment is is non-trivial these days in terms of managing and building deployment pipelines and writing your Terraform scripts to be able to figure out how you're your your your your you know you're you're pushing your code, where it's being pushed, how it's being deployed. You need to support multiple environments, multiple browsers, everything needs to work. And then you need to monitor your code in real time. It's no longer a CD that's shipped to a customer. And if it breaks, ⁓ you know, ship another CD and ⁓ you know what the environment it's shipping on. There's there's inherent complexity everywhere. And ⁓ yes, ⁓ Egal, I agree with you. And and and we'll get we'll get there in a minute. And just as a case point to to to highlight ⁓ Egal's point, you know, the last two years ago, I spent two years working mostly in G C P environments. All the projects I was working on were were in G C P environments and G C P clouds. I got very familiar with a lot of the tools. It took some time, but I got there. And now in the recent startup that I started a year ago now, ⁓ it's been a year and a month, we switched to AWS. We had $100,000 in AWS credits from one of the incubators we we were at. We decided, you know, why why why should we waste those credits? We couldn't get something similar from from Google Cloud. And it took me an unbelievable amount of time To figure out just the console and just where the tools are and creating I I took a 28 hour Udemy course and still I I don't understand half of what's going on half of the time. And that's just for the AWS. And we also moved into a node stack. And I, you know, I I was on a Java, ⁓ Java React stack, now I'm on a node React stack. And just learning the intricacies of Node, learning and reading and understanding what's going on, me and my team. understanding how to do different things, understanding how to optimize lots of things, there's tons of learning. So yes, no longer is an engineer writing code, but they're also taking care of the entire infrastructure and the QAing and the testing and the building, but we're also learning a tremendous amount. And it's so easy to forget. I bet now if I open a Google console, even though a G C P console, even though I've spent two years building on it, I I I I would bet a lot of money that I would I've already forgotten in just one year where everything is and how everything works. And that's the reality of our of our profession. And I I've after a lot of informal ⁓ polling between different colleagues and different peers, I and myself as well. We spend on average between ten and twenty hours a week learning, depending on the task we're doing, if it's rolled over to something new. And it's mostly reading documentation, Googling Stack Overflow, and tons of medium articles. That's why I get nervous when I see the interview guides of Facebook and Google. I would I know I'd never land a job there. Because there there's so much learning that needs to be done. It's it's hard. It's the days of becoming a master in your domain is no longer dependent on your knowledge. It's just dependent on how many hours you've spent in that domain, in my opinion. And most engineers and software developers these days, in my opinion, are no longer masters of coding. We are masters of learning and also forgetting very quickly because the moment you're not practicing, that knowledge is gone. So if if you are a non-engineer listening to this, I I or a non-developer listening to this, I want you to build the empathy of that engineer. When you ask them to create a task or create a business logic, create a feature, often the case is that that engineer needs to spend a lot of time learning things, learning how they connect. Even if they've done it last year, odds are they won't be able to replicate what they've done last year again. So whenever, whenever you embark on a new feature, a new product, keep in mind that you need to at least Allow the time for your engineers to learn ⁓ and and to build new skills or to refreshen their skills. And it this is a non-trivial task. And I don't see it going away anywhere soon. And this is why, in my opinion, engineers or or or development industry is one that is not scaling fast enough, is simply because the amount of learning that goes on and coding becomes an afterfact of learning, just an artifact of it, if you will. So now that we've we've hopefully built some empathy for the engineering role. Let's talk about the friction that rises between the development slash engineering roles and everybody else in the company. And again, it it starts with the complexity of modern software. You know, you have your your your current SLT SDLC theory, ⁓ you start with the design phase, you go validate it with users, and you really don't want stereotyping again, but in the majority of the time, you don't want your engineers talking to your users just because they're not trained for it and they're not equipped for it. So you need to hire someone who's doing that. And ⁓ often not, they haven't coded these people that you've hired to talk to users and and and build a design. So they don't understand the details of coding. And now they need to hand off whatever they're designed to actual coding and implementation. But coding no longer stops here. You need to build the code and test it. And testing is no longer trivial. There's tons of of of happy paths and tons of edge cases that you know is is unfair to expect your engineers to think of all the time. Yes, we have test driven development. And yes, there are theories. If you can write the code, you can write the test case. I've experimented with with a lot of them. You can pretend that you have 100% code coverage, but you know, it's it's it's it's easier said than done. So you hire a QA team. And the QA team, often, more often than not, are not developers and and they haven't coded and they haven't shipped products before. So you're giving them live code and you're telling them, hey, go test it and and test it against the design ⁓ That was two steps before you. And then once you've tested it, you need to figure out how you deploy it. And in this environment, you need you need ⁓ specialized DevOps. And if you're working in anything that's ⁓ you know, secure, ⁓ whether it's banking or or any kind of of of of handling money or or or even even PII today with GDPR and all that stuff, you need a security team. And now you have another, you know, one step removed or two steps removed at this point from your implementation. And then you need to maintain everything and you need to repeat the cycle over and over. So we've developed all these roles. You know, we've we've hired product managers and analysts to figure out the requirements and why we should be doing something. We've created roles that are dedicated for UI and UX designers. And now we see a complete, you know, separation of the UX and UI role. A couple of years ago, most people would put, I need a UX UI designer. Now it's almost heresy if you say I want to hire a UX UI designer. No, you hire a UX designer. You hire a UI designer, and then you have to figure out how they talk between each other. And then of course you have the engineering and your development teams, and you have your QA teams to figure out how the quality and you have your DevOps and security teams to make sure everything's running and it's secure. You have your support staff, your customer success and your customer support who are fielding the phone calls and making sure everything is running. You have your growth and sales teams, you know, you have your growth hackers looking at top of the funnel, bottom of the funnel, where people getting stuck. Send me all that information through segment, through customer I.O. through this, through that. And I want to look at the numbers and slice it and dice it. And of course you have your marketing and and and brand people. Okay, let's build the ads. Let's let's make sure we get the customers. And ultimately you have your stakeholders who are just really care about the bottom line for lack of of of a more polite word. They just want to make sure that the numbers add up and everything's good. So if you look at today's modern stack ⁓ of people, not of of of technologies. You have, you know, this this this tiny little engineer here in one little corner writing the code and and it's it's there's so much pressure coming in and so much pressure going out. And the engineer has to learn and the engineer has a lot of complexity, you know, surrounding the ability to code and and and deploy code. And every single one of these roles do not speak the same language. There's a lot of miscommunication. The product managers who have never shipped before will probably say, you know, what's taking the engineer so long? They're probably just Goofing off. I saw them reading some medium article. What's going on? I'm being a little sarcastic, but I have seen such conversations before. You know, many times I've been parachuted in roles where where where I'm talking to the head of my counterpart at product, and they would tell me, Hey, your engineers are very lazy. You know, we we and and those are words I've heard before. And you you can't get defensive because This person has no understanding of what it takes to code and the learning that it takes to code. And, you know, there's there's all sorts of of of friction because there is no common language, there's no common framework between a lot of these roles. Most companies ⁓ I I I have coached or mentored, you know, they tell me we need to hire QA. And when I ask them why do you need to hire a QA role, we need to test our code. But you know, they there's a lack of understanding from stakeholders what QA entails and what DevOps entails and and and and how they interact And this lack of communication and lack of of common language creates a lot of friction. And you get a tug of war. And all of these people pretend that they serve the user. But you know, if we're being honest, everybody just cares about their own sphere of influence and they care about what they're doing and and and very often the user gets lost. I can't tell you the amount of times I've read a backlog from a company that I'm helping or a team that I'm helping, and the backlog doesn't mention a user once. The stories don't don't even have a user. It's all it's all around, okay, the product manager wants us to do this or we need to have this task. And the user is lost. Where what am I building? And because people lose this focus, in my opinion, this is why a lot of products fail. I know that when I have lost the focus on the user, my products. So how do we get the user back in the center? How do we resolve this tug of war between the why, between the experience, between the how, between the quality, between if it's running or not, between helping our users, between making sure we're making the money. And end of the day making sure we're making a profit. And this is the bulk of my learnings. And in my opinion, in my experience, there are five major buckets that if you get right, I think you've nailed 80% or even 90% of forcing these teams to collaborate together. So what are these tools? What are these tools? And and and my my my ⁓ my experience has led me to these five things. The first is the four constraint models. We'll talk about it in a minute. The second is making sure you have a good roadmap. The third is breaking that roadmap into a backlog. The fourth is setting up good firewalls. And we'll talk about what those are. And the fifth is is something that terrifies me, the word MVP. And we'll talk about that in more depth. Let's start with the four constraints model. This is one of my favorites when I'm talking to non-developers. I have I always you know, when I'm interacting with with non-developers, they come to me with these amazing ideas. You know, one one ⁓ one example, me, you know. blows up my mind always. I had this ⁓ this very senior executive at a large multibillion dollar company come to me and say, Cad, we gotta build the next generation of ⁓ you know website that allows people to buy their homes faster and easier. And I was great. What's your budget? That's always my first word. Because more often than not, executives don't have a budget in mind. And they they need to understand the budget from you. So the the conversation has to start somewhere. And My my my my theory is this these four constraints have always has always helped me start the conversation with any new project. So always start with the cost. Put it up on the whiteboard, put a dot for the cost. If you don't know the cost, keep it empty, but put a dot for the cost. And then I ask them when do you want to launch? And again, most often executives don't have a good answer when they want to launch. But that's okay. This is a conversation. This is this is this is a negotiation between you and I. Let's figure that out. So we'll just put another constraint, launch time. And then I ask them, how good do you want their quality to be? Are you okay if five million users come onto your platform and three million leave because the quality is not, you know, a hundred percent and you keep two million? No, that's not acceptable. We need all five million. Great, hundred percent quality. Let's keep on. And then the final constraint is the number of features. So how many features do you want to ship? Do you want to ship one? Do you want to ship two? Do you want to ship 10? Because the features is going to impact the cost, is going to impact the time. And the quality is also going to impact the time. So all of these things are interdependent on each other. And here's the fun part. I then tell these executives, you can only control for two. Tell me which two you do you want to fix. And they tell me, well, my budget is $10 million. That's unmovable. Great. We have one corner time. When do you want to ship? I want to ship in a year. Great. We have two things that we can control. Now you can no longer control quality and features at this point. And it this is where this is where this tool is very important. That's alignment. Because when executives or or really anybody comes to you, even if it's a UI designer and they come to you and hey, this is what the new component should look like. The conversation, in my opinion, should always be based in cost time quality features. So you're telling me this one feature we need to build. Great. What when when should we get it built? And we need to now involve the product manager and or the stakeholder who's championing this project to figure out when it needs to ship. And we need to consult, you know, our architect, we need to consult a couple of senior folks to figure out Can we ship these features in that timeframe? And if the answer is no, this becomes a negotiation. Okay, so we either extend the time to ship or we reduce the number of features we're shipping. But having these constraints up on the board turns a one-sided conversation, you know, from executive or from decision maker, build this into, okay, let's build this, but it's not gonna work. And here's why. And let's try to figure out a compromise. And I found more often than not that my role is a mediator. My role is not to say no, no, no, no, no. My role is to say that won't work, but here's what can. Now, these features, these 10 features, we can't ship them with this budget and that that time, this quality. Why don't we reduce it to three features? And now the conversation becomes much more robust. And you start understanding, they start understanding why you can't ship them on time, why you can't ship them at this cost. And it then becomes their decision. Should we increase the budget to 20 million to ship these 10 features? Or should we just, you know, accept shipping these three features and and see if we can get more budget after we ship them or not? But the four constraints create create a framework for conversation. And that is the most important step in alignment is that conversation needs to happen and you need to dig up all the things that are that are that are, you know, the boundaries of of what you need to build. And you need to help you you you your job as engineer and and facilitator is to is to help these folks understand the complexity of building something and and and keep it constrained and cost, time, features, quality. Don't talk about, ⁓ it's gonna take me ⁓ you know, three days to build ⁓ a new database, another four weeks to integrate with this API, another four days to build a calculation engine, and then we need to add the web. You've lost your audience at that point. Keep it at time, cost, feature set, quality. If you believe you do the random math in your head and you think these 10 features, you know, all the technical complexity won't work, go back with three features and which ones they are and try to keep the conversation, in my opinion, I found keeping the conversation at this level of these four things helps. And something interesting I've observed is if you fix cost and time, you get an agile model, right? So an agile philosophy, to my understanding at least, is you have a backlog of things, you have a set team, you have a set sprint cycle, and then, you know, you're compromising always on features and quality. And if you fix time and features, you end up with a waterfall model. ⁓ because you know, we must ship on that date. And these are the features we're shipping. And then you have your your good old waterfall and all the spikes that happen towards the end and the rushes that happen towards the end and the awareness that you're gonna fail that happens towards the end. The worst thing is to promise you can fix more than two. In in in all not just my experience, but in everything I've read and everything I've appreciated, I I I've interacted with, you know, I have never seen or heard. or or or even even had a hint that such some something like this is possible. You must fix two and then worry about the rest. Finally, a word of warning here. Beware the mythical man month. If you haven't read the Man Month, the Mythical Man Month, it's a great book, I think from the 70s or the 80s. It is it is one of the one of the best reads I've I've had in my life around software development. It still holds true after all this time. Basically It says the summary of the book or one of the chapters of the book is that you cannot add people towards the end of the project and expect it to go better. So if you're building this project and you know you're running out of time, you know, you can't suddenly increase cost and and and and and ⁓ you know get more time or or lower the time. You because you need to factor in the education, you need to factor in the the overhead, you need to factor in the learning that that new team needs to come in and pick up. And they're not learning from reading documentation, they're learning from your team. So that's why you have the slowdown effect of adding new people to the project. So just keep that Man Monk policy in mind. So once we've had a conversation around what we're doing and we have a realistic conversation, one rooted in reality, not pessimistic, not optimistic. Hopefully you're threading a very thin line there. Experience helps. If you don't have the experience, there are tons of people around you in your network who have that experience. Talk to them. I very rarely start a project without finding someone who's spent at least a year or two immersed in that tech or in that framework or in that industry to to to understand what what what you know mistakes they've had, what failures they've had, what lessons they have for me. And ⁓ you know, once once you figure it out and you have alignment from all those roles that we mentioned before, you have alignment on what you're doing in a general sense. Okay, we're building a new bank. Great. What are the features for whom and all that? Now it's time to build a roadmap. And this again is where I see a lot of companies fail. I often mentor and coach a lot of startups specifically in in in in incubators around Toronto. And, you know, a lot of them come to the table and say, Hey, Jad, we have a great team, we have a great product, but we're always stretched. We're always losing engineers. I don't know why they quit. We try to create a great culture. We have a ping pong table and and and we have free lunch and stuff like that, but we see the loose talent. And we're still churning and and ⁓ we're still not not pushing code fast enough. And my first question always to them is show me your roadmap. And ⁓ you know, I'd say maybe between 60 and 80 percent of the time, over half the time, they don't have a roadmap. Well, how do you expect your engineers to know what they're working on? How do you expect your team, your entire company, forget the engineers to know what you're building and working on? End of the day, you need to have some kind of vision of where you're going. And this is what the roadmap does. And the roadmap is very simple. It doesn't have to be complicated. It is what are we building? When are we going to ship these things based on the four constraints model we talked about? And for whom are we building these things? Just just keep that in mind. And this is not set in stone. Revisit it as many times as you want, change it as many times as you want, as things, new information comes out. We're not politicians here. We can flip-flop, no problem. But understand the what, when, and for whom and keep it up there, eye up on the wall. This is our roadmap. This is what we're building. This is why we're building it. And here's how it's going to roll out. And the key in my experience to good roadmapping, they're not hard. Just establish the goals with the owners of that roadmap. So whether the product team, the engineering team, the, you know, your design team, just tell them here's what we're building. Here's the order in which we're trying to build it in. Very high level. Engage your team. Get them to understand what you're doing and why you're doing it. And something really cool comes out of this. It forces you to prioritize because your team is going to come to you and say, ⁓ these three things, which order are we building them in? And you can't say we're building them in the same order. That doesn't work. So it forces you to prioritize, which is a good thing. We're going to talk about that a little more in a minute. And then it forces you to also visualize everything. Once you visualize a roadmap, it's easy to share. And now folks can look out and say, okay, one, here's what we're building. I get it. And I can contribute to that. And I know what we're going to do there. Q2, here's what we're gonna try to build. And you change that and you refine that as you go along. And that's all it takes to build a great roadmap. And of course, keep iterating on it and changing it. Once you have a great roadmap, now it's time to start rolling up your sleeves and build a great backlog. Most companies I know that have built backlogs or most teams that I've observed that built backlogs just build backlogs. You know, you have a product manager who looks at the feature set and says, okay, this is my feature set. Let's break it down into stories. They break it down into stories and before they know it, they're running a waterfall model without knowing what waterfall is. I'm not I have nothing against waterfall. I've built successful products using waterfall because I got lucky maybe. I failed a lot with it to I'm not trying to to to bash it or bash agile. I, you know, these are frameworks, you use the to right tool, the right situation. But what I have a criticism against is building something when you don't understand the fundamentals of why you're building it, because then you're gonna get yourself into trouble. You're using the wrong tool for the wrong job or you're building it in the wrong Way. So the backlog, in my opinion, must be derived from the roadmap. You look at your roadmap, you look at the priority of things, and then you break it down. Why do we break it down? We break it down into small chunks because people can't can't spend their entire, you know, Q1 building one thing. It doesn't, it doesn't make sense spending three months building something. You know, very few people, I don't think anybody, honestly, has the mental capacity to look at something. Hold that information for three months and just go code for three months. And also you need to iterate and test and all that cool stuff. So the backlog needs to follow from the roadmap that followed from the constraint models. And now we still have everybody involved that needs to be involved. We have the QA team looking at the backlog saying, I know what I need to test. We have the DevOps folks looking at the backlog saying, I know what I need to deploy and I know what needs to come tomorrow. And because these are smart folks that you brought into your organization, they can tell you, hey, if we're actually going to hit this roadmap and build this backlog. These are the three things we need. Now suddenly, because of this visibility, what you want to create is people coming to you telling you what you need to have. And that's beautiful because nobody, no executive, no engineering leader knows it all. You know, that that that I don't buy that. My role as an engineering leader is to make things visible for people and then help them integrate all the tools and all the things they need to be able to execute what we need to execute on. That's my role is to make things visible, to make them Real. So going back to the backlog, take that roadmap, break it up into a bunch of stories or or epics or whatever you want to call them. And the purpose of them is that they're discrete, testable chunks. So what makes a good story? I tried looking for a picture of Mother Goose. I couldn't. It's amazing how how there's really no good pictures of Mother Goose on the internet that are free to use. Anyway, what makes a good story? Single owner. There must be somebody accountable for that story. The moment you have multiple owners on a story, you know, the thread gets lost. You can't have four storytellers getting up in a in a in a crowded place and all of them saying the story at the same time. It doesn't work. So how do you expect that to work in a development backlog? Single owner, focused on the user, small enough to be completed in a couple of days so you can go. ⁓ and and an engineer can go take that story, break it down into tasks and work on it, whether alone on a team. But also the stories need to be large enough to be end to end testable. You know, having a button and saying, I want a button here, that's not a story because you can't test what a button does. I want a button. When I click on it, it creates a reservation. Now that's end-to-end testable. Now, as a user, I can go in, I can click that button, I can see that it creates a reservation or not. So the stories should have zero technical value. Like it shouldn't. That happens later. The team should take task breakdown and break it down, but we won't go there. But it should be end-to-end testable. And this, in my opinion, is what so many stories lack. Make it end-to-end testable. Test it from a user, from a business point of view, from a business value point. And Because there's business value here, the backlog must be strictly prioritized. And what I mean by strictly, you can't have task one and task one having the same priority. There must be task one and task two. Why? Because the team can only work on one story at any moment in time. Yes, you tell me, okay, there's a work in progress limit and a Kanban model. Great. You know, nobody's blocked. But when a developer sits on their keyboard and starts coding, they're coding against one store. So you must tell me in which order of priority do you want me to complete these stories? Because if I do not have that order of priority, then it's, you know, all hell breaks loose and you have all this friction all over again. Because the product manager just put the backlog up there. They don't know what the team's capacity is to build. There are ways to estimate, but but you know, longer story. They don't know where the team's gonna be roadblocked on. So if the team just takes haphazard stories, you've created fix friction between the product manager, product owner or the product manager and the team. The stakeholders don't know what's going to get done when. Long story short, strictly prioritize your backlog at all times and have strict conversation around strictly prioritizing it. Is this more important than this? Yes. Is this more important than this? No. Let's flip them. Is this good now? Yes. You can switch priority whenever you want outside of the sprint, but strict prioritization. Now you've got yourself a nice back. And the team grabs it, sprints, and then you test iterate and shit. The next tool. We now move away and abstract ourselves out of the the you know, the minutiae of of story and and and code development. And this this next one is near and dear to my heart. And it is organizational firewalls. So I don't know how many of you have been in conversations with the executives when every day it's a different thing or product managers, I want this feature. And then they talk to a customer, no, I want that feature. No, I want this feature again. No, actually I changed my mind. I want that feature. And that's totally okay. There's nothing wrong with that. That flip-flopping, that chaos, that up and down, totally cool. And it should be allowed 100%. But what should not be allowed is that up and down to be exposed to your team. You, in my book, you should never, ever, ever, ever let your team be exposed to the volatility of decision making that went into creating the backlog. Tell them what the roadmap is. Tell them if you want that we've changed the roadmap. No problem. Here's why we've changed it. But as you get to closer and closer discrete chunks and as you hit the sprint, that's it. Don't touch the team. Don't tell them we change stuff. If you must change stuff, like 90% of the time, just swallow a bitter pill and wait two weeks until you change it. And if it is urgent and and product breaking, that's okay. The team will understand, but don't do it often. Organizational firewalls are really simple. This is just saying that here's the sprint or here's the can band board or here's whatever you want. Mr. and Mrs. Engineer, go do it in the corner. Be happy, I nobody's gonna interrupt you. Why? Because you want to give your dev team the time to learn and to implement and to execute and to code and to build everything cool. And if you keep exposing them to that flip-flop, they won't be able to find the time because you know task switching is is is a real thing. Anyway, philosophically, I isolate the the engineers, we said that. Philosophically and and and from a from a historic point of view, the scrum master role arose in agile philosophy and agile and lean implementation. The the scrum master role rose to create that firewall. So the scrum master, when it first started, the role first started out and I first started experimenting with it, the scrum master was like, okay, you want some changes to the backlog? No problem. Talk to me. Leave the team alone. The scrum master's role was a firewall. This got lost in my opinion. I think we forgot why the scrum master role became something bigger. It's now a full-time job. I don't want to insult anybody, but I really do not understand the need for a full-time scrum master in most teams. And in anyway, but but the scrum master role is traditionally what's been the firewall. Technical lead in many companies is the role. The role is called the technical lead, but you need a point person that's isolating the team or the VP engineering or the engineering director, whatever you want to call it, but have somebody dedicated to protecting the team and protecting the team's time. And one really important thing to note: this firewall role is not full-time. This is I'd say it's around five percent of the role of of of a person's time. Just you know, that that's that's why the these crumb mass anyway, five percent of the time of a person's time is is literally what, a couple of hours a week, just pushing back on on time with the team, pushing back on the need to flip flop things, helping people understand what's going on in the backlog. But it's not by any chance a by any means a full time role. It's something you can step up tomorrow and start playing in your company. And if you get failure points anywhere along this role. Or along the line, just have a retro. And I in my in my experience, retros, retroactive meetings are vastly underutilized. A retro is basically getting the whole team together, whatever, whoever needs to be together. It doesn't need to be the engineering team. It could be the development team, it could be the, you know, there's friction between the developers and product managers. So get them in a room. Have them talk. And a retro needs to be full of friction. If you're having retros that's all, you know, fist bumps and back slaps. And nobody's crying after it, you probably did not have a productive retro. Why? Because you did not identify where the friction points are. You did not have those cathartic, you know, come to Jesus moments with the team. And and and it's it's hard to identify where you what you need to do to improve stuff. So, you know, I like to go into retros and when I facilitate them, more often than not, I know what the issues are. And I if the team doesn't bring them up, I'll start hinting at ways to bring them up. But I want I want a full honest conversation between the product managers and the engineering team. I want the product managers to say, engineering team, you know, I feel you're lazy. And I want the engineering team to say, why do you feel like that? And if they don't say it, I'll ask them, why do you feel that? And then we bring out all the all the points that the product manager is talking about. We let the engineering team bring out all the things that the product managers don't see. And end of the session, we either have better understanding or we come up with a process that will allow better understanding. Like the product manager invited to dailies or something or whatever. But there's always something that can be done better. Finally, the MVP as a culture. This is my pet peeve when I talk to new companies or or or inexperienced product owners or managers or whatever you want to call it. Because usually the cycle of product development is proof of concept, prototype it, build the MVP, and then harden and scale. And I'll tell you why this upsets me a lot. MVP is supposed to be the minimally viable product. Let's think about that. It means what is the minimum amount of effort that it is going to take me to build a product? Why should this be a step? So are you telling me by extension, the moment I'm done building my minimally viable product, I can start bloating and adding stuff that doesn't make sense? No. And that's why it it bothers me. As an organization, I always try to build a MVP as a culture where, you know, it's proof of concept, prototype, and then harden and scale. At all points of these, at the proof of concept, I should be doing the minimum amount of work necessary to prove the concept. Same at the prototype stage. Same especially at the hardened scale stage. What is the least amount of work that I need to do to scale this? So one case in point, I had this, I had this really nice company I was helping the other month. And the the CEO and the CTO could not see eye to eye. The CTO was saying, Hey, we've built our MVP, we've proved our concept, we've built it on What was it? They built it on Ruby. And the CTO wanted to refactor this entire product. They spent two years building the front end to react. Now the CEO didn't know how to how to say no to that. They they knew in their gut, you know, she knew in her gut that she should say no, but she didn't have any arguments why because the CTO was really persuasive. And if you think just from a technical point of view, it made a lot of sense. But you know, when I asked the CTO why do you need to refactor besides the coolness and besides whatever you need to do? Ultimately it came down to some efficiencies. Well, okay, man, you're stuck with Ruby. How can, you know, go for a week and come back and tell me how you can optimize your current stack. And the CTO came back with a list of four things that they can do in Ruby that did not involve an entire refactor. And and everybody was not happy, but you know, it was it was better than an entire refactor of the front end. The point is MVP should always be pervasive, you know, should be a mental model always in everything your teams built. What is the minimal thing that I need to do? And in every conversation, everybody should have that. And once you're allowed, you do that, you see a lot of alignment. So the product manager comes to the engineer and says, Hey, why are you refactoring that? It's not on the product backlog. It's not minimally viable. Why are we doing it? So you give everybody the right tools to have these conversations to continue reinforcing a line. So basically, MVP is a journey, not a destination. It's it needs to be something you do always, not something you do once. So in concluding all of this, there's a lot of friction. We get that. Hopefully we've built some empathy if you're not an engineer. If you are, maybe you felt validated, whatever. But basically, these five things have I always bring out to any company I build or any project I do, put the constraints up on the board, talk them through. Build alignment from that alignment, build your roadmap. From that roadmap, build a backlog. From that backlog, make sure that the teams are firewalled and executing it. And at all points, in every single one of these conversations and these steps, make MVP, make everything minimally viable as a culture. I'd almost say have a culture of laziness. Like what is the minimal thing that you need to do to get the biggest bang for your buck? And if you're not having that conversation, there's waste. And if there's waste, somebody else is eating your lunch alive. So, you know, keep those things in mind. At least that's what's helped me a lot in my career. Anyway, we're at the end. I hope I gave you some good insights or at least some some some room to think new thoughts. Maybe you disagree with me. I'd love to have conversations on the side. ⁓ whatever whatever is needed to help you continue building the best teams you can and building the best products you can. And I'm always happy to chat. All right, thanks a lot. ⁓ yeah, I have a few questions I received in private that I can ⁓ share with you. And I don't know, Vincent, if you ⁓ received any ⁓ also on your side. So one of them is you said that the start that engineers have become master learners. Is it an option to reduce this burden by staying longer on the same stack and and foregoing some steps in the delivery process, etc. Or do you think ⁓ proportion of learning will stay no matter what? ⁓ So go ahead. Yeah. I I have I have I need to wear two hats to answer this and I I don't think I hope I'm gonna make sense. The first hat is the hat of ⁓ CT And and my my my answer is very clear cut. Yeah, stick with the same stack. It will it will it will help you. It will it will always help you as long as and the danger here is you don't get complacent. You know, you don't think, Okay, I've mastered node, I don't need to read any more documentation. So y you need to always find the ways to challenge yourself. Now if I put my engineering and coding hat on, I'd say the answer is definitively no. Because at least from my point of view, my personal point of view, I will probably shoot my brains out if I work on null for two years straight. I need something new, right? But that's my personality. I know that's common among a lot of developers, but maybe maybe not everybody. So I think the the the the compromise here is as long as you're learning and you don't get arrogant with your knowledge, I think the stack doesn't matter and the learning doesn't matter. It's it's what suits you. Are you trying to push yourself to always become a better developer and engineer? Or are you trying to push yourself to always become a better leader and an engineering leader? If you're trying to go into the engineering leader route, the more you know, the better. If you're trying to be a super specialist in one stack, then you know the the more the deeper you know the better. It depends on on where your goal is. Right. I I hope that balances it out the answer. Cool. Thanks. A question that that we received twice, so I'm gonna answer it. It's not something that It's it's more on on on our part, ⁓ yeah. So it's ⁓ can we have your slides or can we have a recording from this? It's really ⁓ there's a lot of value here I wanna share with my team. So the answer is yes. The recording is gonna be ⁓ shared to ⁓ to everyone in the in the CEO club. And and ⁓ yeah, yeah, if you wanna share your slides as well we can we can do that ⁓ also. Yeah. So that your your teams can benefit from this. Happy, happy to do so. All right. ⁓ Vince, did you have any other questions on your side that you received? No, that's it. ⁓ I think ⁓ yeah, no the the only comment was ⁓ from Major and Dev. About the recording? Yep. Okay. So any last minute questions, anyone on this? It's ⁓ it's a very broad subject, I would say. ⁓ maybe one that comes to my mind, maybe in the meantime if someone has other questions. How do you do you have another other examples of on how to to introduce or that ⁓ V P culture at all? you know, at all stages. It's not just one delivery that's hey, that's my MVP. No, you need to think in in terms of ⁓ VP all the times. And some projects we we we often use the ⁓ the acronym YAGNI, so you ain't gonna need it. So just do the minimum thing that you need. And if you're you think it it might be great to do something, well just make sure you know you have to do it. That it might be great to do something. So how generally how do you make that culture happen? I I I think you you you already you know, introduce a great way to make it happen is is as long as you have the thinking, it doesn't matter if it's MVP or not. The the point is always do the least amount of work that you need to do to get something done. And I'm not saying work less. This is very this is very important to make this distinction. I'm saying work less for every task because if you work less for every task, you have more time to do more now or to optimize or to learn. And the the The way to build this culture, in my opinion, is really simple. It it starts with a lot of reinforcement and it needs to come from it needs to come from the leader of the team or or or the executives or or or you know somewhere that has influence in the company and and you need to reinforce always. Like someone comes with a great idea. That's great. Let's listen to it. But do we need this? And the most important thing is to challenge executives, to challenge CEOs, to challenge, you know, the executives of the company who come to me with feature requests. Like that's great. I love this feature. I get it. But let's look at our backlog. We have 12 other things there. And here's a cool tip. Here's a cool trick I use at least. You know, if I I get an executive or CEO who tells me, Chad, we must have this feature. Great. I love it. Here's the 12 other features we have in our backlog. Where do you want me to put it? You want me to put it on the top, the bottom? So forcing going back to the prioritization discussion of anything you bill is is helpful. So that's why having a good backlog that's prioritized works. Because if you need to insert something in it, you immediately see at a glance, does it fit or not? Is it more important to build this new thing than than something else? I think that's for executives. For for for development teams, it's just always reinforcing, like you said, you know, do you need it or not? And if you don't need it, don't build it. There's no need to build a new webhook if nobody's going to use it and you're not geared up to to to get your customers to use it. If you want to build it as a prototype and testing, go ahead. But to build the all the support and infrastructure for that, it just does not make sense unless you have everything lined up. So as long as it ties into the roadmap and everyone's aligned, then I think that's how you build start building and and you push back anything that doesn't fit the roadmap. That's where you start building such a culture. Great. Thank you. Another question I just ⁓ Christy is how do you how do you take learning into account in your four constraints that you had, for example, and time and cost and quality, I guess, also? Yeah, I mean it the learning is is in all the poor aspects, I guess. So how do you take that into account and how do you educate non technical people about it, about the reality that ⁓ if we try something new we need to learn, either i ⁓ because we need to learn about the user behaviors before we get too too far in this in this kind of feature, we need understand how the user's gonna do it, for example. Or if we try that new technology, we need to learn about how to use it correctly, or we we might we might have, you know, system downtime because we made some big mistakes. So how do you educate and take that into account when you when you are accountable and accountable in front of of your stakeholders? It's a very interesting question. I I my my immediate gut reaction is if the stakeholder is non technical, I wouldn't even bother explaining it honestly. It's it's going to frustrate it's going to frustrate everybody and because they'll come up with great because if they have never learned how to code, it's really hard to teach them how others are going to code. And it's just gonna take away from all the conversations that we do need to have. So again, the do I need to have this conversation kind of, you know, instead of coding it, do I need to have the conversation? More often than not, I I unless somebody's really curious and interested, I just won't have it. It it just factors in. The way it factors in is really a lot of mistakes I've done, a lot of failures I've done. But the biggest way to factor that in is talk to somebody who's a master in such a domain. So if you want to move, if you want to implement something new in Node, or if you want to implement a new API. Find someone who's done it. And and and you will very often, not very often, I have yet to have tried to do something new and not been able to find at least one or two people who have done it. And talk to them about, okay, so I come from a I come from a, you know, ⁓ a Java background. How long do you think it would take me to become really good at Note? And they can tell you, well, I've never come from Java, but I've come from C and moving to Note, it took me so much, and this is what I need to do. And high level, you can start getting the intricacies. And they can tell you, actually, I know my colleague. I come from Java and to know it. And you can start having richer experiences. So my opinion is talk to people who have done it. And that's the best way you can learn, other than making the mistake yourself. And once you have a good idea, you can iterate on it. And one thing I want to go back, you you mentioned something very important. You said accountable to the people you're working with. And I I want to make clear that when we put a roadmap or when we have the four constraints, we talk with the knowledge that we have today. Nothing is fixed. And it shouldn't be. If you're working in a culture where, ⁓ but last week or last two years you told me you're going to deliver this. Why did you not deliver this? And they don't give you the room to learn, then I I'd say something's wrong with that culture. So if if if if today you you you come up with, you know, the you have the four constraints, you come up with a with a model and then you realize that more learning is needed or more whatever is needed, and you come back and you you reiterate and you need to push some timelines. That's when it becomes okay, should we drop some features or should we extend the timeline? And that's where you go back to the four constraints. I have new information. We we might need to ship with lower quality or ship with less features or or d delay shipping. So summary, find and ask people, work with what information you have and update it as new information comes in. All right. Great answer, yeah. Thank you very much. We're at ⁓ one o'clock here, eastern time. So I guess we're gonna be respectful of everyone's time. So thanks for that. Yad again. Thanks everyone for being here and we see you in a month or so. depending on if you want to attend the next one. So thanks very much. Perfect. Thank you, everybody. Bye.