David Subar is the Founder and Managing Partner of Interna, an IT Services and Consulting company focused on making product development successful and effective. We discuss the foundation of the lean startup approach, failure as a catalyst for learning, and ways to deliver better products faster.
Listen to the podcast here
Deliver Better Product, Faster with David Subar
Our guest is David Subar, the founder of Interna, which enables technology companies to ship better products, plan and execute successful product roadmaps, and achieve product markets fit more rapidly, and deploy capital more efficiently. David, welcome to the show.
Thanks for having me, Steve.
Great to have you here. And I’m very curious to get into your story and your blueprint and all that stuff. So let’s start at the beginning where we always do your entrepreneurial journey. So how did the idea come about for Interna and what was the road that led you to it?
I had been chief technology officer and chief product officer of a number of different technology startups. And I’d done that for quite a while and I’d learned a lot. I learned a lot of things that you shouldn’t do, I learned a lot of things that you should do. Some of the lessons were hard learned. And I realized after a while, what I started out doing was I wanted to take technology and build something that people recognize. That’s where I started my career. I started my career doing research and development in AI and machine learning. But over the years, as I started managing bigger and bigger teams, I realized what I was really good at. It was really more, more important for me anyway, for companies was building teams that could build great product. Writing software is very different than building teams and they’re both important skills, but they’re different. And so what I realized was many companies particularly, I work in technology companies, don’t know how to build teams that can build great products, which means that they spend a lot of money. And it takes a long time and they release something, it’s the wrong thing. And they use capital inefficiently, which means they don’t get to have the impact they want to have, and they don’t get to be successful company. And so that was what started driving me is like, “How do we build teams that build products that really have an impact, have companies have an impact, have companies be successful?”
Now, David, your approach, is it tailored specifically for tech teams or it’s for any team?
It’s for any team, but tech teams have some particular advantages and particularly software teams, but it starts with, you’ve got to understand who your customer is. Who do you serve? What do they want? What creates value for them? Creating profit and revenue is a side effect of creating value for your customer. If you don’t understand who your customer is. They’re not going to likely want to buy your product or service, and therefore you won’t make revenue and you won’t be profitable. That’s where it all starts. That’s true for any kind of company.
Yeah. So, so this is basically the first step of your blueprint. So this, on this podcast, you always try to isolate framework, business blueprint that people can relate to, can capture and then try to apply in their businesses. So I recall this blueprint “The Deliver Better Product, Faster Blueprint.” Like step one, deliver customer value. If you don’t deliver the value, there’s no revenue, there’s no profit. The business is not sustainable, that’s clear. So what is step two in your framework?
Step two is, so you know who you’re going to serve, you believe you know what they want, you’ve got to put stuff on the roadmap. And everything on the roadmap is a bet. You’re not going to get it exactly right. You get a certain number of bets because you have a certain amount, you don’t have unlimited capital. You have a certain number of bets. So you’re going to put things on the roadmap. Everything on the roadmap is going to have an epic statement. An epic statement looks like, we believe by doing this feature – for this user – they will receive this kind of value. And we’ll know that when we see this metric.You get a certain number of bets because you have a certain amount, you don't have unlimited capital. Click To Tweet
So let me stop you right there. So you see that every step is a bet, essentially. Is this feature going to work? Is it not going to work? Is this going to be useful for customers or not? And then what would be the vision for that? If it really worked, what is the impact of it for the customer? And then, so before you move on, how do you decide which bets to go after?
That’s a great question. There are a million bets you can make, a million ways you can achieve value for the customer. You filter those bets from which of those are aligned to your company strategy. You can do a bunch of things for customers, but you want to do things that also benefit your company. “Hey, we want to increase market penetration.” We’re going to pick those bets, increase market penetration. We want to decrease churn. We’re going to pick those bets that decrease churn. Look at those bets, compare it to your company strategy and filter by those and say, “Okay, here’s the ones that create benefit for the customer. And they also create benefit for us.” Now of those bets, so now we’ve filtered some, some of those bets are going to be more or less expensive, going to take more time, less time. Pick the ones you can learn quickly, pick the ones that take smaller amount of time to build and release. So you can see if you’re actually having forward progress with the market or for your business. So those are the two steps of filter, which of them align to your strategy and which of them cost less capital, take less time to get on the market so you can learn quicker. This is all about learning quickly. We’re making bets. Some are gonna be right. Some are gonna be wrong. How do we learn as efficiently as possible about whether we’re progressing for the customer and for us?There are a million bets you can make, a million ways you can achieve value for the customer. You filter those bets from which of those are aligned to your company strategy. You can do a bunch of things for customers, but you want to do things… Click To Tweet
What about the third dimension, which could be the impact on the customer? Because you can have multiple bets that are in line with your strategy. And then you can learn maybe some of them you can learn quickly. What about the dimension of the impact on the customer? Cause I I’d rather you learn something a little bit slowly and feel more impact. Then learn as quickly, have minimal impact, take it into account as well.
Yes. So that’s part of the epic statement, we believe by doing this feature by this customer, they will receive this benefit. So you have to understand the benefit that they’re achieving, and then you have to be able to measure it to see if they’ve really achieved it. If you don’t have that fundamental thing of we’re achieving this value for the customer, don’t start. You’re not going to have a sustainable business. It’s got to start, it’s rooted in, we’re creating value for someone else. And then when we do that, those that value creation also create value for us, but it’s got to be, it’s got to start by creating value for customers.
Okay. So we deliver customer value. We basically make some bet. We rate them. Are they easy to learn? Do they have an impact? We describe the impact as an epic statement of the vision of the impact. Then what’s next? What is the next step? I guess that’s going to be step three in your sequence.
Right. So, well, I kind of mixed step two with step three, which is which of these bets align with where the business strategy for the business is going. If you’re running product management engineering, you’re probably not the CEO, a lot of times product management engineering groups become feature factories. Some sales person would say, “We need this feature because I have to close this deal.” Or marketing says, “We need this feature because I want to market it.” Those are things that are aligned to individual departments and a feature factory is one that just takes orders and does what’s told. What you want is having an empowered product management engineering team. If you’re not the CEO, that alignment to where are we trying to drive the customer, which of these bets do we pick, really are about alignment with the CEO because the CEO is trying to increase the market value of the company over a long period of time. And doing that driving to that North Star, gets alignment with the CEO. If you are a CEO, then congratulations, you’re aligned with yourself by definition, but you need to have a strategy.
So aligning with the CEO is the same as aligning with the company’s mission or vision. It’s the same thing or it’s separate?
I would say it’s aligning with the company’s strategy. That’s, at least in theory, the CEO has built a strategy to fulfill the mission and vision of the company. And so, all of the things on the roadmap need to fulfill that strategy. If you’re doing something that doesn’t fulfill that strategy, it’s a hobby. That’s lovely. I have hobbies. I really like them. No one pays you for them. So it’s the alignment with the CEO is alignment to the CEO strategy, which drives the company forward, which gets you to the mission vision.
So step one was to make sure we deliver customer value. Step two is filter our bets and we said that they want to pick the bets that are aligned with our strategy and that are easy to learn and generate impact. Now when we are aligning the best with the strategy, isn’t that the same as aligning with the CEO or there are two levels of alignment?
No, that’s just it. Actually, I would say first step is know who your customer is. Second step is pick bets that you’re going to make. Third step is align those bets to the CEO, which is the same as a strategy.I would say first step is know who your customer is. Second step is pick bets that you're going to make. Third step is align those bets to the CEO, which is the same as a strategy. Click To Tweet
Okay. All right. We aligned it to the CEO and then what, what’s next? So you have your customers, you came up with the bets, the directions that you want to develop that software product, perhaps, it’s fully aligned with the set of the company that the CEO represents. What’s next?
Now you’re going to build it. So you made a budget, you said, “Here’s the best we’re going to make. Now you going to build it and you going to be efficient, as quickly as you can at building it because every company has limited capital. So the, the question is, how quickly can we get this thing in the market and see if we made a bet, if our bets are right or wrong, you should expect some of your bets are going to be wrong. That’s okay, but you want to learn quickly. If all your bets are right, you’re not stretching far enough. You’re not trying far enough, hard enough. I worked with American Express years ago. I was at AI tools firm and we were building something called Authorizer’s Assistant. Authorizer’s assistant was an intelligent system to make credit decisions every time you used American Express card. American express advertises, we have no preset credit limit. It doesn’t mean they don’t have a credit limit. It just changes every time you make a charge on the credit card. If you’re doing well, you have a higher credit limit. If you’re not doing well, you’re lower one. So I said to our sponsor, American express, “Oh, so you want us to make no bad authorizations, every authorization should be good?” He said, “No, that’s a terrible idea.” I’m like, why is that a terrible idea? He said, “Because I can get no bad authorizations very easily by having zero authorization and people are gonna use MasterCard and Visa. I want to make the right number of bad authorization so I can learn.” It’s the same when you’re doing product. Some of them should be wrong, but you should learn quick when you go. “ Oh, we thought the customer was gonna like this. We thought this was gonna be profitable. It’s not.” What did we learn from that? And how are we going to change our roadmap? How are we going to change our bets based on that? It’s all about a very active feedback loop. So the quicker you can release in the market, the quicker you can learn, the better you can tune the product roadmap, the company to creating value for customers, the better you can tune the company for fulfilling its strategy.
Okay, that’s awesome. So build it, make some mistakes, there is optimal number of mistakes to make so you can learn and you’re basically swinging for the fences. You’re not playing it too conservative and then you measure against your objective. So what about the objectives? Are they tangible up front? So up front we talked about aligning with the CEO’s vision, but is this more of a broad stroke alignment, or is this a very granular where you actually have measurable objectives alignments? How does these two alignments come together?
It’s about being granular and measurable. Everything, that’s where the epic statement starts. We believe by doing this feature for this user, we will see this metric move, but we will create this value and we’ll know if we see this metric move. Everything you’re doing has a measurable value you’re creating, and then also for the strategy, here’s a strategy. We wanted to, once again, this is an example, we wanted to decrease churn. Did churn decrease? Yes or no? Oh churn decreased more than we thought. Oh, that’s interesting. What did we learn from that? Churn didn’t decrease, it increased. Oh, that’s interesting. What did we learn from that? It’s about everything needs to be instrumented. So the objectives you think you’re going to meet, you need to know whether you actually met them or not. And then after you measure it, you need to ask that question. Oh, here’s what happened that was different than what we expected, either more or less positive, what do we learn? And then we start to process again. Do we actually have the right customer? Do we have the right stuff on the roadmap? Are our measurements right? Are we aligning with strategy? Can we build faster? And so you just start the whole process again, you iterate.It's about being granular and measurable. Click To Tweet
Now I said before, software, it’s easier with software. Why is it easier with software? Software, my time to market is very very small, particularly for consumer software, I can release a new version of the app. B2C software, it’s a little bit harder if I have salespeople that go out. Maybe I’m selling to insurance companies and they have a long time to implement things or, large B2B enterprises, when you’re producing hardware, it’s harder still. So you need to modify these techniques for whether you’re doing consumer package good, I’m building, I’m writing. Whether you’re building a piece of technology, hardware, you’re doing software because your velocity of learning is different.
Yes. So in a consumer and market, you can experiment easier because there are bigger numbers. Statistically the more significant tasks, I guess you can A/B test much easier there. When you’re in the enterprise market, you have to be not upset your enterprise customers, and maybe it’s a few smaller number and you have to stab more carefully. Is this why the difference or there are other dimensions?
It’s that, but it’s also just the consumption of new releases for B2B tends to be slower from B2C. I had 25 new apps on my phone today. I woke up, looked at them quickly, just said, “Okay, great. Accept them all.”. Those were 25 bets that different companies were making with me. My new iPhone, I’m waiting for the iPhone 15 to come out. I’m one who, you may not expect this, but I’m one who keeps technology for a long time so I have an iPhone X. I have the new Apple watch, but I have a really old iPhone. So that the velocity of me consuming new hardware is much slower than software. Now, the good news is my wife has an iPhone 14. Other people have iPhone 14 and iPhone 13. So Apple in aggregate can learn. I’m not the only slow consumer, but there’s a lot of people who are consuming the hardware, but that’s a once a year release. Software can be multiple times a day and so the ability to learn is slower. You’ve got to be more thoughtful about your bets when you’re doing slow release cycle products than when you’re doing quick release cycle products.
Okay, that makes sense. Now you describe this process as it’s kind of a lean startup approach. So tell me about this lean startup. So is this just for software and technology companies, or is it also for professional service businesses, maybe traditional service businesses who deliver maybe retail service, food service, household service businesses as well, or just for software businesses and why?
It’s for everyone, just easier for software businesses. But let me talk about how you do it if you’re not a software thing. Let’s say you’ve been online service, let’s say you have a housecleaning service, and let’s say you have a large house cleaning service and you have a thousand people that go to houses across the nation and clean their houses. It takes a while to train a thousand people for them to learn whatever new technique it is. An easier way to build this kind of model is go on the website and put up a service that you don’t yet offer and say, let’s say, I’m just completely making this up. Pool cleaning. We do pool cleaning now too. We do house cleaning and pool cleaning. Pool cleaning people put a button is learn about our new pool cleaning service. Test how many people actually click on it. That is a cheap way to understand demand for the service and then say, coming soon, sign up here for our new pool cleaning service is going to come out in three months, making that up, obviously. So there’s ways of testing consumer demand or customer demand without launching the whole service in a much less expensive way. Now you need to do it in a way that, that is not deceitful to your potential customers, right? But there’s ways of measuring demand and then you can go like, “Oh, people actually want the pool cleaning service. I’m going to do it in Brooklyn, New York only.” I don’t know if there’s pools in Brooklyn, New York. Probably there are. I’ll make it easier. I’m going to do in Costa Mesa, California, which pulls the Costa Mesa, California. I’m not going to start it nationwide, I’m going to start it in a small and I’m going to test the service and then I’m going to expand it. The key is once again, small, quick releases that don’t cost much where you can learn quickly. And then you could iterate in the service and you can expand the geography.The key is once again, small, quick releases that don't cost much where you can learn quickly. And then you could iterate in the service and you can expand the geography. Click To Tweet
Okay, so it’s basically a running sprints. So let’s say I have a pool cleaning service and I create a process. Maybe a 15 step process, how I prepare and then how I clean, empty the water from the pool and then the detergents I’m using and the equipment I’m using and the process how I’m going to go through and then refill the pool. Let’s say that’s the process. And then I test how successful the process is and then iterate and maybe tweak the process. Is this how it’s would apply to a traditional service?
Exactly, but it’s not just iterating, tweaking the process to make the process better, but making sure the customer received value and the company was able to achieve its objectives. It’s just not the technical components of the process. Like I can tell you a lot about how software is built. I came from the software side of the world. I can talk to you about algorithms and efficiency. I can talk about some large language. And you know what? Most people don’t care that’s technical stuff that nerds like me are really interested in. What people care is when I get on Google maps, do I get efficient directions to where I want to go? That’s what people care about. So those processes, I want to test them. So I know that I’m efficient as a company doing what I want to do, but the ultimate measure is whether I’m creating customer value and whether it’s pushing the company forward. And that’s what you fundamentally have to measure.
So how do I create a practical test environment where I’m only delivering a limited number of the service in a limited number of customers? So I say, if I end up a pool cleaner, maybe I’m starting a company. I’m trying to figure out what would be the optimal process or maybe innovate on the existing processes that are out there. Let’s say I figure things out. It’s a 15 step process. I have one crew, they can deliver that maybe to three or four houses maximum a day. So how can I get data that is reliable? Is this applicable to a small business service or you have to have a certain number of, a certain population of users to use it in order to have a statistically significant data so that you can actually use it to run your next sprint in your business?
Right. So system significance is a really good question it’s having larger numbers to test against, you’re going to get a better result. If you’re a small pool cleaning company with a few customers a day, it’s better than not having data, but it’s, it could be suspect. This is another reason, like I came from the software industry, it’s easier in the software industry because you can do A/B tests on here’s the old method, here’s a new method we’re doing it at scale, we’re seeing things trend. But even startup software business by definition, have few client, and so you do your best you can with the data you have. And that’s why iterating quickly becomes very important. “Hey, I’m a software business. I only have a hundred clients in the beginning.” They seem to say this, “Okay, let’s double down and try more of that. Is that working?” “Oh, now we have a thousand clients.” And you know what? That was 30% right, 70% wrong. Great. It’s about being quick at doing the releases and learning quickly. It’s all about learning quickly. The lean startup model came very much from the software industry and we work internal works, in the technology industry. The techniques I’m talking about are applicable other places. We work with lynda.com, we worked with Disney, work with lots of companies, but it’s easier when you have large numbers. And, but in any case, it’s required to release quickly, learn quickly. Think about what happened and do it again.
I’m just thinking, trying to relate it to my coaching business. I started I don’t know, six, seven years ago. And I didn’t have large numbers to test things on, but I did perceive sometimes a better reaction or a less good reaction. And of course it depend on the client as well, but even with a few repetitions, I kind of get a feel for what is a better way. And then I did more of that and did less of the other and kind of it, it helped me. And along the way, even without a scientifically proven testing methodology and statistical number. So that’s a great blueprint. So make sure you deliver customer value, you formulate an epic statement of your bets that you have chosen and make sure it’s aligned with the CEO’s vision, build it and then measure it back to the objectives.
So, David, if our listeners would like to learn more about your process or would like to contact you, talk to you or explore about Interna, where should they go?
So our website is www interna.com. So it’s I N T E R N A .com. You can find me on LinkedIn. So in turn, by the way, is like internal without the L. My name is David Subar. It’s like Subaru without the U, S U B A R. You can find me on LinkedIn. You can hit me up there. You can hit us up on our website. You can find us there. Feel free to reach out. Someone just wants to spend a few minutes talking about these things. I’m glad to chat with anybody anytime.
David, thanks for coming to the show. Really enjoyed it and have a great day.
Thank you, Steve. Glad to be on.