Hunter Harris: Hey everybody and welcome to Build and Break Through I'm your host, Hunter Harris of Hunter Software Consulting. With me today is Joel Tosi. He is the author of Signals and Levers. You might remember the other co author of Signals and Levers, Elizabeth Hendrickson, who came on earlier. And he's also the co founder of Dojo and co. Joel, do you want to introduce yourself a little bit? Joel Tosi: Hey, I'm Joel, been in the tech space for geeze a few decades now. various places, low high late or low latency, high throughput trading systems, Red Hat, blah blah blah, all that kind of fun stuff. Hunter Harris: Yeah, you y places like Red Hat that no one's ever heard of, right? Joel Tosi: It was Red Hat before it was IBM man. I was there for the good days. Hunter Harris: Yeah, man. I I I don't know if you had a chance to look at my LinkedIn, but I had my stint at IBM too, and that's a whole another beast. came Joel Tosi: okay. Yeah. Hunter Harris: in through an acquisition as well, but that's neither here nor there. so Joel Tosi: Yeah. Hunter Harris: I guess kicking us off, the main thing that I wanted to talk about today is high functioning product delivery teams. So I I mean, you've clearly seen a ton all kinds of levels of scale. What are some of the through lines that you've seen? Joel Tosi: So when I think about like high functioning product teams, I I I don't even think about like teams that necessarily ship the fastest. Because I think a lot of teams get kind of stuck in that space, like we just gotta Hunter Harris: Quite. Joel Tosi: crank out more stuff. The the the teams that I think about that I think about like high functioning, I mean are the teams that are constantly thinking about are they doing the right thing? Like ca the constant focus. One of the the best groups I was with students I'm consulting with this is probably 2012, 2013, and they brought me in and they said, We need some help kind of thinking about product stuff. And I sat with them for like a day or two and they go, What do you think? You know, how would you help us? And I go, I think you guys are already pretty awesome. And they go, But we're not shipping fast enough. I go, Yeah, but how do you know you should ship faster? And they were asking Hunter Harris: Mm. Joel Tosi: each other these wonderful things around like, if we did it this way, would the experience be better? And like in 2013, 2014, when they when this when I was with this team. They were like part of like the quote unquote R and D group and they could actually go downstairs. This is at a big box retailer, and they would actually go downstairs into the store a couple times a week and try things with customers. And it was never about like trying to get them hooked. It was the it was the good observation, right? Like, did this solve their problem faster? Did they get confused? And they would come back and they would say, They got confused here, let's fix it this way. And then they would try it again. And when I think about high functioning product teams, man, it's it reminds me of that, like solving the actual problem in a way that is intuitive for your users. Hunter Harris: Absolutely agree. And I mean, I love the call out to the user or to the customer there, because I think Joel Tosi: Yeah. Hunter Harris: that through line there is absolutely vital. So you reminded me significantly with that of something that I refer to as integration. I don't know. I mean, we obviously have continuous integration, which we'll probably get to, but one thing that I found really, really interesting between the high-function teams that I've worked with and the less high-functioning teams is integration, that is to say, silos. l many, many organizations want to protect their engineering organizations from the rest of the business. Joel Tosi: Yeah. Hunter Harris: Put them in an ivory tower, put them in a silo, say, hey, you need this heads-down time, don't let anyone bother you. Just go and build, build, build, build, build. Versus, for example, I when I was at SalesLoft for example, I had the opportunity to shadow salespeople and sit in on sales calls. And man, what I could bring back to the team from these calls was just game changing because I Joel Tosi: Yeah. Hunter Harris: knew exactly like we were building sales software, so we were our first customer. So first of all, dog fooding. And then second of all, being able to like view the results of that dog fooding firsthand and be able to advocate for those people and how they were actually using the application. Which Joel Tosi: Yeah. Hunter Harris: man, places aren't doing Joel Tosi: Yeah. Hunter Harris: either of those things enough. Joel Tosi: It's wild, dude. And I think what's crazy is it's not even just for like new feature functionality. Like I had I had a team that I was working with, again, a lovely team, love them to death. They had this issue with like dirty rights to the database, right? And they and Hunter Harris: Mm-hmm. Joel Tosi: it been going on for years and no one could figure out why it was happening, but it was happening all the time in production. And one day I just go like, Can you guys can you guys call somebody? Like, can you just talk to a user and see kind of what's happening? And they go, they pull something up and they look in the database and they say, yeah, this dirty right just happened at this one location. I go, can we call them? And they go, I I don't know. I mean we have the phone number, but we never I don't know if we can if we're allowed to. And I was like, just try it and see what happens. Yeah, exactly. We called and then we found out what was happening. And like the and the issue, as soon as we heard what it was actually happening, it was like in essence two screens open at the same time. And it's like, whoa, what? Why'd you have two screens open at the same time? Well, I started it, but my secretary needed to finish it. So she opened it in a different screen. It's like, never thought anybody would do that. It's like, now the problem and it was trivial to fix after that. But if you didn't actually ask somebody, you get all these problems and people are upset because I did all this work and it it it's gone now. It's just like talk to integrate the feedback, man. Hunter Harris: Absolutely. I love that one too, because I was about to interject a little bit with man, what is the name of that product that I love? story something. I can't believe I'm blanking on this because I talk to people so much about Joel Tosi: Ha ha ha. Hunter Harris: this particular product. But I mean it it's one of several products. It hooks in your UI and it traces the user journey through the UI. Joel Tosi: Like pendulum kind of thing? Hunter Harris: I don't know Pendo. Joel Tosi: Okay. Hunter Harris: I mean like heap is in that category, but there's a sp Particular one that I can't remember the name of right now. But but the point being that like I usually recommend these pieces of software for exactly this reason where bugs occur, you can go back, you can view a playback of their session, see what was actually happening, see what like weird thing was happening. But this is a great example of like even tools like that would not catch it, right? Because you'd have a session over here, a session Joel Tosi: Right, because it's a different system. Hunter Harris: over there, no way to like share that particular content. So it goes back Joel Tosi: Yeah. Hunter Harris: to like there's absolutely no substitute for actually You know, talking to your end user because no matter how good your spec suite is, no matter how good your r like R and D process, no your ARDs, your P R D's, whatever, none of Joel Tosi: Yeah. Hunter Harris: it is gonna survive contact with a customer 'cause they're gonna do a bunch of stuff that you never thought they would ever do, right? Joel Tosi: Yeah. One of one of my other favorite stories in the whole customer space, like and this is the same thing like talking to your customers without giving too much information away, you know, imagine you Hunter Harris: Course. Joel Tosi: had you have internet because we're talking over the internet right now. So you have internet provider. And so I'm guessing if you're like me, when your internet provider goes down, if you if you have to call, you don't wanna like you just want your internet back, right? You don't wanna you don't wanna and so the the This organization that was providing internet service decided that was the right time to try and sell people HBO Max. Like people are calling. Hunter Harris: Lord no. Joel Tosi: Yeah. So so there were two things happening here. First of all, when we talk to the operations center, they're like, please don't do this. Like people are so mad when they call. Don't don't try and get us to upsell them. Hunter Harris: Please don't. Joel Tosi: Yeah. And then the funny thing was, they had already like the operations center where people would call in already had the promotions on. Like available to them. But the kick in the pants was the developers and the designers and the UX team worked on 21 inch monitors at the time, a couple of years ago. The operations center had 14 inch monitors. So they never saw a third of all the information. And so when people are asking, why aren't we selling enough HBO Max? We gotta push it harder, it's because your people can't even see it, man. And so again, sometimes Hunter Harris: Mm-hmm. Joel Tosi: you just gotta go to where people are at and see what's going on. Hunter Harris: Yeah, and I mean that issue is especially exacerbated now given the fact that we have not only like a plethora of monitor sizes, we have Android and iOS and Pixel and Samsung and d device device device device device, iPads, Joel Tosi: Yeah. Hunter Harris: whatever tablets. Joel Tosi: Yeah. Hunter Harris: I know back in the day I was using Litmus to like for email marketing to try and at least render marketing emails across all these devices, but man, there's no substitute for just figuring out what people are actually using and making sure that it's actually working for them. Joel Tosi: Yeah, and it's trivial, right? It's like you show up you talk to I mean it's not it's not revolutionary. Hunter Harris: Yeah, exactly. So I I guess though, like well, let's trace this back to kind of like some of the cultural angles. I do see a lot of companies struggling with this. What do you think are like the core beliefs, practices, and cultures that lead to a breakdown in this sort of like integration piece to where engineering is walled off from access to the customer? Joel Tosi: You know, that's a good question, man. I I wonder if it isn't the idea of, you know, the just the sheer cost. You know, engineering historically has been a it's been a high cost, right? You know, to to build software has never been a cheap thing. So the whole idea is it's almost Tayloristic right? Like as long as somebody else can decide and figure out all the messy details, and then we just kind of give you the details, the expensive stuff, mm, you just do it. So that way you don't have to worry about the noise. But I think the challenge is like there's There's so much information in the perceived noise, like your noise isn't necessarily my noise. Hunter Harris: Can you define Tayloristic there? Joel Tosi: yeah, the whole s like this almost the separation of strategy and execution, right? It's just if somebody else can figure out the right thing to do and I can just it just tell you exactly what to do, I can give it to you and you can just, you know, clickety clack. Yeah, Taylor Taylorism from like the early nineteen hundreds. Yeah, yeah. Good call, man. Hunter Harris: Yeah. It's a clickety clack, I love it. That's one of my terms too. Just Joel Tosi: Yeah. Hunter Harris: just clickety clack at it and it it'll be gone. I I mean, haven't Joel Tosi: It'll be a solved problem. Hunter Harris: we repeatedly relearned that you can't really sever strategy from execution though, right? Like Joel Tosi: It's wild, isn't it? It it feels like we're we're we we we've been learning the same lessons over and over again for over a hundred years. I Hunter Harris: Yeah. I mean it feels like about every five years we have to relearn that lesson. maybe Joel Tosi: Yeah. Hunter Harris: bef before things sped up it was maybe every ten years, but it feels like it's at least every Joel Tosi: Yeah, I say, I I I yeah, I think that's a good point, man. Because now, especially like with the prevalence of AI, it's just like we can build things faster, let's just kinda just go faster. So now it's like, Well, hold on. Now like with like we like with AI, if if you're amplifying the wrong thing and you're doing the wrong thing faster, you just Hunter Harris: Exactly. Joel Tosi: you just burn more dollars. That's all you do. Hunter Harris: Yeah, it's such an interesting conclusion to me. So there's this school of thought that with AI, execution has g become so cheap you can compensate for strategy with just more execution, right? Which I mean, obviously I agree, but there's this counterpoint there that I've been thinking very heavily about in terms of opportunity cost, right? So if execution is increasingly cheap, that means opportunity cost wise, the cost of your wrong strategy is increasing, not decreasing, right? Joel Tosi: Yeah, very much so. Yeah. Hunter Harris: So a as a result, like, it's not that execution pays for strategy. It's actually the opposite. The increase in execution means that strategy is more important, not less. And we can see that borne out across the market, right? Like the companies that actually invest in direction actually achieve high velocity versus the ones that don't. Well, I mean That's always been the case, right? The market is littered Joel Tosi: Yeah. Hunter Harris: with corpses of in basically infinity startups that built a thing in the wrong direction rapidly, but then found out no one to pay any money for it. Joel Tosi: Yeah. That it's an interesting call out too, man, because I I think there's two interesting facets there. Like one is I forget who said this, but like, you know, it's it it takes so much time to gain a new customer but less to kind of lose And so like if you're just kinda cranking things at it's a certain point in time, like your opportunity costs People are going to stop paying attention when it's like, don't care, don't need it, don't want it, it lets me like just leave me alone. So you have that going on. So the opportunity cost is going up faster, kind of like you're like you were saying. And there's also this other super interesting space where it's like, I think a lot of organizations have learning problems in a multitude of ways. Like one is just kind of like we Hunter Harris: Multitude. Joel Tosi: we don't learn kind of like how we need to like we don't learn about ourselves, but then there's also this this lacking feedback loop of like Are we building the right thing? And then that what do we learn from people using our systems? And so if you're Hunter Harris: Mm-hmm. Joel Tosi: just executing faster, but you're not itch and you're you're not learning from what you're executing, or you can't learn as fast as you're executing, then go then going faster is i it's another kind of call it a bottleneck or problem there, there. You know, you're going faster than you should Hunter Harris: Mm-hmm. Joel Tosi: be going. Hunter Harris: Yeah, like if you're going faster than you have the ability to evaluate the results of, I I think I butchered that sentence, but I think you understand what I mean. Like Joel Tosi: Yeah. Hunter Harris: it's kind of a moot point, right? How do you know if what you built succeeded or not? How do you Joel Tosi: Yeah. Hunter Harris: attribute it to anything? How do you say if this particular exercise had ROI as a result? How do you know if you need to do more or less of it? And in the absence of that, in terms of back to things like engineering being extremely costly, you see things like your product surface area growing exponentially versus your headcount growing, let's say, linearly or not at all, or decreasing, which leads to just not only outrunning your own capabilities, but then further outrunning your capabilities as your organization changes in a way that like your your product surface is so much larger than your engineering. Which in turn drives like it it's a self-sustaining cycle, I think probably technically a vicious cycle in this case, right? Because Joel Tosi: Yeah. Yeah yeah. Hunter Harris: you have more product surface, which means you can't keep up with it, which means you have to delegate more to AI, which means you get more product surface, which means you further can't keep track of it, which means you have to delegate even more to AI, and the flywheel turns. Joel Tosi: Yeah, I mean there yeah, I mean there's a bunch of things new things there, like is as you have more initiatives in flight, it's even harder to discern signal from noise. Right. You don't if if you're Hunter Harris: Exactly. Joel Tosi: if you're pushing out one feature, you can A B test it. If you push out fifteen features simultaneously and you A B test you have this weird overlap and you get into calculus, and then it's just kinda like it gets crazy. Hunter Harris: And if you A B test it appropriately you have what, two to the fifteenth power like feature sets Joel Tosi: Exactly. Hunter Harris: and you probably don't have that many customers. I i Joel Tosi: Yeah. Hunter Harris: if you're aggressively A B testing, you probably don't have two to the fifteenth power customers. So I'm pretty confident. Joel Tosi: Yeah. And then and then you get into like this weird thing that's yeah, like Elizabeth and I even talk about like process theater and change theater. I think a lot of organizations Hunter Harris: Mm. Joel Tosi: get into like product theater. Like, I don't know if it's making a difference, but we got a lot more out there. I don't know, let's get s and like so that you start like since you can't measure and you don't even know how to learn if you're doing the right thing, you just kind of play the game. Like we have OKRs and KPIs, but we don't ever look at Come on, man. Hunter Harris: Yeah. yeah. Especially LKRs and KPIs, 'cause they can be so impactful if you do them right. Such a waste to Joel Tosi: Yeah. Hunter Harris: see them when it's just absolutely theater. I mean to that particular effect, one th kind of anti pattern that I see pretty frequently is we don't know what we're doing, so let's change all let's change our branding kit, change our colors, rebrand, Joel Tosi: Yeah. Hunter Harris: restyle everything, because that that's sure to make a difference, question mark. but It's so sad to see like a an application that just has like core onboarding issues and then they invest months in like changing colors and moving buttons around instead of trying to do basic design enhancements that are everyone knows is like, hey, your onboarding thing involves you filling out a seven part form and involves forty two clicks. Yeah, license key license keys Joel Tosi: Right. Like copying license keys and yeah. Hunter Harris: is a great example. Like all this sort of stuff where it's like This is not a user-friendly product. There's a reason why people bounce off of it. Instead of changing the colors, what if we tried to fix that? So d Joel Tosi: Yeah. And so often, like i in in our space, we'll talk about technical debt a lot, but I think a lot of organizations, a lot of software products actually get product debt. Like they have so much kind of like product stuff that's just kind of cruft hanging around that makes it harder to for the customers to navigate and then harder to build and man. Hunter Harris: That's one thing that's interesting too, because I feel like that particular category gets lumped in under technical debt somewhat unfairly, where for example, I've worked with a number of places with feature flags, for example, and they'll have a long tail of like, okay, we have 572 features, a hundred are fully rolled out, a hundred have never been rolled out, Joel Tosi: Yeah. Hunter Harris: and all the remainder are in some interim state of partial rollout. So Joel Tosi: Yep. Hunter Harris: To to the point of like your fractal complexity and to the X power like Joel Tosi: Yeah. Hunter Harris: QA sets. It it becomes just a maintainability nightmare of where it's more expensive for engineering, it's more expensive for QA. And no one wants to pay for all that. But everyone's Joel Tosi: No. Hunter Harris: too scared to go through and do things like, hey, these hundred are have never been rolled out, let's tear it out. But what if we need it later? You're not gonna need it later. You haven't needed it for like four years. Chill. Joel Tosi: Yeah. Yeah. Hunter Harris: Or and similarly it's a hundred percent rolled out. Well like what if we remove it and it breaks? Yeah, it's it's been Joel Tosi: Remove Hunter Harris: in there for four years. Chill, get rid of it. Like Joel Tosi: Yeah. Hunter Harris: or in this case don't get rid of it, like get rid of the feature flag. but I and what's wild about that too is especially like with AI and or just with a basic engineer, that's such a category of like low risk change, right? It's heavily, heavily mechanical to take the equiv I mean, I certainly hope that you're using a feature flag system that is the equivalent of an if else. Branch, right? If you're Joel Tosi: Yeah. Hunter Harris: not, you're doing something terribly wrong. But in that case, it's just turn the if-else into not if-else, and you're done. Joel Tosi: True. Yeah, exactly. It gets wild when you start seeing nested feature flags, right? Then it's like, well you could have had an if else, Hunter Harris: Lord. Joel Tosi: but now you have seventeen c possible permutations. It's like, wow, that's now I'm scared. Hunter Harris: Yeah, I've seen that in production and man that really jacks up the QA burden too. Joel Tosi: Yeah. Hunter Harris: And it has ramifications too in terms of CI C D like now you have have an exponentially large test suite to support that, which in turn means your CI suite runs longer, which means you can't complete work as fast, which means you can't deploy work as fast. And it's just Joel Tosi: Yeah. Hunter Harris: this huge invisible drag on the whole organization that looks like free if to a A naive outsider. Joel Tosi: Yeah, and it's not free. And like to you and to your point, it it adds to the complexity, it adds to the build times, but it also like, let's be honest, I mean, testing there are false positives all the time. And so as soon as you start increasing Hunter Harris: All the time. Joel Tosi: permutations, now some now you get false positives, and maybe it is just the wrong time. And now you have somebody has to do the debugging and the analysis and try to recreate it and see if it was. And so, yeah, the just the the the this this product debt. It it it I I think it's unfair to lump up with technical debt, although sometimes it's a technical reason for it. But sometimes there's just a bunch of these ideas that we wanted to see if they would work and then we never said they didn't work. So we just left them around. Hunter Harris: Right. Yeah, and man, there's some core there's some really interesting core entrepreneurship stuff in there too, continuing talking about feature five and product, because there's I one of the persistent themes I've seen as well is you have a particular customer, they may be in many cases like a lower value customer who's just hanging on by their fingernails for whatever reason, but they're like the one person that one particular feature has rolled out to. Joel Tosi: Yeah. Hunter Harris: And the organization just kept tr keeps that particular set of features alive for a low value customer just because they don't wanna lose one customer when in Joel Tosi: Yeah. Hunter Harris: actuality the dollar amount that they're investing in maintaining this aspect of the system is usually, in many cases, like an order of magnitude larger than the value of a contract, right? Joel Tosi: Yeah. I I I have a good friend, Jeff Patton. If you haven't talked through my introductions, love to introduce him to you. But like Je Jeff's a brilliant kind of guy in the in the product space, at least for me. And I remember talking to Jeff, this this is probably a conversation from pre-COVID. So you're talking like six or seven years ago. And I was I was at a client and the client kept on saying, We're we are customer centric, we are customer centric, we are customer centric. And I was talking to Jeff and I go, Jeff, help me understand this difference between customer centric and product centric. And he did this, Jeff's brilliant. And he said, in most places where he goes, when they talk about being customer-centric, it's kind of like you're saying, we don't ever wanna make the customer unhappy. If even if it's the one customer that's barely paying anything and they ask for a feature, we're gonna put it in. And he goes, That's not that's not product ownership. That's reacting to like everybody else and letting somebody else dictate where you wanna go. Product ownership Hunter Harris: Absolutely. Joel Tosi: is about having a vision and a strategy and talking about the essence of the problem. And sometimes that means you don't do what your customer asks because it's not the right thing for your product. And he goes, that's hard. And he goes, and that's that's when I hear you talking about like startups. it's for me, it's it's that separation, Hunter Harris: Mm-hmm. Joel Tosi: like the separation between customer oriented and product oriented. Like, you know, do you own your product? Hunter Harris: I mean I would actually in terms of like customer orientation, if you really want to say that you're customer oriented, I think you should be obsessed with your customer. And that means that y Joel Tosi: Yes. Hunter Harris: to the point of this, like you should care about the customer getting what they really need, not necessarily what they want. And in many cases, Joel Tosi: Yeah, yeah. Hunter Harris: that means firing them as a customer, right? Like if Joel Tosi: Yeah. Hunter Harris: they i if for example, like I talk a lot about opinionation and the value of opinionation in your products. Joel Tosi: Yeah. Hunter Harris: Because it's so much easier to build a business around like an opinionated workflow and saying, hey, for example, we do a ton of sales, we know how successful sales look. If you use our product, you will use this particular workflow. We will not make it customizable for you, because that's bad for you, that's bad for us. And Joel Tosi: Yeah. Hunter Harris: if you don't want to do that, you shouldn't be a customer. Right? Like that aspect of niching down is incredibly undervalued in terms of Not only like successful go to market, but also building an effective product team and an effective engineering team that can execute on that product vision, right? And so I I think that a lot of these companies l i if you have the attitude of like, we're customer-centric, therefore we'll do whatever the customer tells us to do, that's not being customer-centric at all whatsoever, in my opinion. Joel Tosi: It's not yeah. Hunter Harris: It's it's really just Honestly, kind of being cowardly with your sales, right? It's we w we will sacrifice our engineering budget to land a sale at any cost. And there may be a sick a scale early on in a company's lifetime where that can be a valuable choice. Like if you have no sales and you need to make one sale. But Joel Tosi: Yeah. Hunter Harris: if you're trying to maintain a customer base, that can be lethal to your own. Joel Tosi: Yeah. You get too much, too much spread, too many permutations, it becomes a maintenance nightmare, you kind of lose signal, you start reacting, and that becomes you get in those silly disagreements. Well, why would you do it for that customer and that not that customer? I had one client that told me, well, we have funding, so why wouldn't we build this thing? And it's like, just because you have the money to do it doesn't mean it's a good idea, because like the cost isn't just how much you pay right now. the cost is the the maintenance and Hunter Harris: Mm-hmm. Joel Tosi: it's also the ripple effect on everything else because by choosing to do this thing you're also if not saying no to something else you're at least delaying something else and like Hunter Harris: Mm-hmm. Joel Tosi: that there's the opportunity cost. And so like may maybe we're just bad at economics and software. I don't know. Hunter Harris: I think we are honestly really bad at economics. I would love like I feel like I've increasingly like there's behavioral economics, but there's also opportunity cost. And I find Joel Tosi: Yeah. Hunter Harris: opportunity cost is like the boogeyman that keeps on rearing its ugly head here and there. Where God, there's so many organizations that are like, Yeah, okay, we could do the we could go down this particular path and make a whole big bucket of cash. But we'd rather not do that. We'd rather Joel Tosi: Yeah. Hunter Harris: do something completely different for various reasons. And Joel Tosi: Yeah. Hunter Harris: now now opportunity cost is not like real money. It's kind of monopoly money. So I can understand why a lot of people don't want to focus too heavily on it. But completely discounting it, I think, is incredibly dangerous. Because opportunity cost, in many cases, it in the short term, like opportunity There's like a longer term opportunity cost that I think is monopoly money. But short term opportunity cost of like we invested money in maintaining feature flags versus reinventing our business to compete with its new competitor is real survivability questions for businesses, right? Joel Tosi: Yeah. I mean an opportunity costs don't always have to be big bets, man. They can be small bets to see if you should keep investing. And I think that's that's the other challenge is a lot of people think like, Well, I have so much money and so I can do a big thing. And and it's hard to kind of say like, Well, if I have twenty million dollars, I gotta do a twenty million dollar thing. It's not like how do I think about it in small iterative bets and kind of looking for signal, looking where to go. It turns like You know, we wrote that that the that book and it's it's a lot of the stuff we talk about is really just thinking and like like thinking about the system, thinking about what's happening. And it whether it's a product and you're like you have to be in tune with with what's happening. You have to be thinking about I'm doing these small bets. I have a couple feature flags, where am I getting signaled? Now if I I I can't keep running the same bet f you know, add to infinitum, I have to turn some things off so I could focus, right? And so it's about the constant focus. It's hard, man. It's hard. Hunter Harris: It is super hard. And and I mean I love the call to focus too. I feel like there's so many different flavors of dysfunction, but I feel like so many of them there's obviously the cultural ones, but the other large bucket that I see is always priority driven where there's like one like priority one is big and tangly and so it's scary. So people wanna like delay working on that. So they'll just be down the list on like Priority thirty-six or forty instead of priority one. And being able to just focus crisply on one and doing the legwork to actually understand, okay, cool. Priority one, is this the thing that we need to work on? What's the juice? What's the squeeze? Like ship it or ax it off the list and move on to the next thing, is Joel Tosi: Yeah. Hunter Harris: truly underrated in this phase. Joel Tosi: Having seen this with a lot of organizations and a lot of teams, like like you said, dude, like number one might be super hard and super gnarly and s and just scary. And it might be because like there's 80% uncertainty in it. But a lot of times you there a lot of times you can say, But there's a little piece of it that we know has to be done. Can we just use that to kind of learn and see what the next piece is? And what you end up finding is Exactly Hunter Harris: And chip away at the uncertainty and chip away at the uncertainty. Exactly. Joel Tosi: You start you start learning by doing and then you start to figure out the things of that eighty percent that you didn't need to do. And that's that that's that's the stuff that I want to see people do more of is like small kind of bets and kind of like, we don't know everything, but we know Hunter Harris: Absolutely. Joel Tosi: enough. Let's do let's let's use that to feedback more learning back into our product to see where we should go next. Hunter Harris: Absolutely. I mean, I truly believe that for any project, I mean, there's always going to be unknown unknowns for almost any project. Joel Tosi: Yeah. Hunter Harris: But even for a project that has a large like area of here be dragons, putting the lead legwork to your point doing small shippable chunks, like it has a way of pushing back the boundary of the uncertain and at least Joel Tosi: Yeah. Hunter Harris: taking the unknown knowns and flipping them to known unknowns. That then Joel Tosi: Yeah. Hunter Harris: you can go research, answer, and in many cases, those in turn flip more of the boundary into known unknowns. And that that chaos and that uncertainty ends up evaporating much faster than most people realize as long as you put in the concerted effort to do it. Joel Tosi: Exactly. And you have to I I think it's the awareness of it too. Like you have to be like and sometimes it's like when we're building a product, we don't want to say, Man, we don't know a lot about this. We wanna say, we gotta do some more research and then we'll be confident. Hunter Harris: We know everything. Joel Tosi: Instead of like like we need to almost embrace uncertainty and kind of say, There's things we don't know, man, but the only way we're gonna find out is by chipping it away at a little piece at a time, kind of leaning into what we do know, Hunter Harris: Mm-hmm. Joel Tosi: trying to learn faster. That's that's the magic. Hunter Harris: And that's part of the beauty of stuff like CI C D, right? Like, Joel Tosi: Yeah. Hunter Harris: if if you can build these small shippable chunks, if you're continuously deploying, that also means if the experiment fails or goes the opposite opposite direction that you want, you have it at your fingertips to just blast that change away and go back to the way the world was previously. Too many teams like they really focus on risk mitigation in a way that actually increases risk. Like let's move slower instead of faster. Joel Tosi: Yeah. Hunter Harris: Let's ship less often instead of more often in the interest of not breaking things, but in actuality, changes pile up, larger changes are shipped instead of smaller changes, and reverting from those changes becomes increasingly non trivial instead of easy. Joel Tosi: Yeah, but like even like there there's a whole inherent risk of just larger batches and more, you know, more change happening concurrently. But there's also like the risk that people don't really talk about is like the risk of just doing the wrong thing. And like you don't know Hunter Harris: Of course. Joel Tosi: longer, you don't know longer if you did the wrong thing. And then you get that some cost fallacy and it's just like, but we already built it, can't we just make it better? It's like, well, we probably shouldn't have done it in the first place, man. And so you like it like the the larger batches and like when we don't do CICD well, it tends to kind of cascade itself and cause more expense and more risk. Like like you said, the more risk aspect, but I see I think it actually cost more. Just because we then invest in the wrong things. You know, best case scenario, Hunter Harris: It certainly does. Yeah, like Joel Tosi: like the pipelines are green and we we have good artifacts we can roll back and the defects are low, but if you're building the wrong thing, it doesn't matter. Hunter Harris: Absolutely. Well, Joel, I wanna be respectful of your time. Would you like to plug some pluggables? Joel Tosi: look, the only stuff we got to plug is we have a beautiful book coming out at September twenty second with IT Revolution, signals and levers. and then Elizabeth and I, who was a few weeks on a few weeks ago, we'll be at various conferences around the states and around the globe. Come hang out. I promise you we we will help you think about things in new ways. And what what I would even say a little bit further, not just things in new ways. I believe that if we could do a little bit more critical thinking in some of our work. Our work actually becomes easier. I think it becomes a little bit less scarier. And I think that's what we want to do is make like software engineering should be fun, man. It should be fun building things. It shouldn't be Hunter Harris: Mm-hmm. Joel Tosi: scared and kind of like bureaucratic. It should be like, I don't know, let's see how this works. Let's let's have fun together. Let's let's bring some joy back to software. And I think that's what we're trying to do. Hunter Harris: A thousand percent agree. I mean, that's the thrill of building and the thrill of delighting the customer is a big part of what Joel Tosi: Yeah. Hunter Harris: got me into this in the first place, right? So preaching to the choir, and I'll signal boost more on signals and levers As I mentioned on the previous episode, I read that book. I enjoyed the hell out of that book. So I really recommend it for everybody here to check it out. and also check out the previous episode with Elizabeth Hendrickson, too. We talk a lot. I mean, we talked. more generally in this conversation, but that particular chat was very signals and levers focused specifically. but Joel, thank you so much for joining us. This has been fantastic. There's a ton of gold here. And for all of my listeners, I hope that you've really enjoyed this. I hope that you've learned a lot. I hope that you're inspired to get out there, build, experiment, iterate, and I hope you all have a fantastic day. Peace. Joel Tosi: Thank you.