Product Iteration with Hacker Noon Interim CTO Dane Lyons

cover
16 Oct 2018

Listen on iTunes and Watch on Youtube.

“To build a good product, you’ve got to embrace change. When you have a product and you change something about the product, there’s really only three outcomes. It’s going to get better, it’s going to get worse, or it’s going to stay the same.” — Dane Lyons

On this special edition episode, Hacker Noon Founder & CEO David Smooke guest hosts to discuss the process of making great products with newly hired interim CTO Dane Lyons (who has since become our permanent CPO). Dane is the founder of v1labs & Knowtify, and he is one of the best product minds around.

Full Transcript:

David Smooke: Hello and welcome to the Hacker Noon and Podcast. I am your guest host today, David Smooke, founder and CEO of Hacker Noon, how hackers start their afternoon. I’m the guest host today because we’ve hired a new person and a new firm, Dane Lyons of v1Labs. I’ve known Dane for a while and he’s going to serve as interim CTO, as Hacker Noon builds it’s own infrastructure. And Dane is a great person for this because he’s built a lot of products, he’s really one of the better product minds that I’ve had the chance to work with in the past. We worked together in his first company. I worked for him at Knowtify and they sold to Kissmetrics, where Dane was an engineering leader there for a while. And before that, a lot of cool stuff with how tags and stories work at, I forget the companies name now. but it got bought by Walmart, Walmart labs.

David Smooke: And anyway, today we’re going to talk a little bit about the future of Hacker Noon. What does digital publishing look like for the tech industry. Then more specifically Scrum versus Agile, moving from wireframes to interactive prototypes, what does it mean build an MVP, how do you evaluate an MVP, an other stuff like that. Excited to have you on, Dane.

Dane Lyons: That’s great. I’m very excited to be here. I’m love the direction of Hacker Noon, to be honest. I think that Hacker Noon, in my mind, is brutal at it’s best. It’s just very unapologetically raw and I like that. I’m really excited to be here.

David Smooke: Yeah, I think the blindingly green does a lot of that rawness, and then, when you open it up to contributors of all walks of life and really, your only qualification is to doing cool work about tech or having your own perspective about tech. You just really open up to a lot of different tones and voices. So Dane, what do you think makes a good product?

Dane Lyons: Well, I think that it’s a complicated question but, for me, I think that to build a good product, you’ve got to embrace change. When you have a product and you change something about the product, there’s really only three outcomes. It’s going to get better, it’s going to get worse, or it’s going to stay the same. And if it gets worse, you can usually change it back so there’s no real harm in change and the up side of changing is very high. So I believe that teams need to find a system that embraces change, whatever that is.

David Smooke: And do you ever fear that too many changes, too many iterations, can lead to lack of identity or sanity?

Dane Lyons: Changes can occasionally be negative. I think that you can always have misguided changes and I think that a lot of founders and a lot of people who are in charge of products, I think that creates a lot of fear. So there is a risk in a change going awry but the upside is so much greater that it’s worth it.

David Smooke: Was it scared money don’t make money?

Dane Lyons: Yeah, I guess.

David Smooke: Could you tell me a little bit about Scrum and whether you think people should be using it?

Dane Lyons: In my mind, Scrum is not inherently bad, but it is not a great process for an early stage company. And the reason behind that, it kind od comes back to what I was just talking about, change. And Scrum is not optimized for change or for bringing about change. Scrum actually has a relatively bureaucratic process in terms of bringing about change. You have you stakeholders and you product owners who basically decide things an through this long process, they convert ideas into tasks and then your individual contributors are responsible for delivering those tasks. And that gets into a lot of … you get into trouble when you need to adapt and it’s very slow to redefine your task.

David Smooke: You think the biggest problem is the most top-down engineering management style?

Dane Lyons: Scrum is definitely not the most top-down engineering management style but I think that it’s deceptive in that people think it’s a very agile and nimble process but it is pretty top-down, in my opinion. At least most implementations that I’ve seen.

David Smooke: And then, how do you think Agile is more bottom-up?

Dane Lyons: Well, in my mind, agile should really create a system where individual contributors can make the decisions that bring about change and if you can do that then the process of defining their tasks and … it shouldn’t take two weeks to go and create a bunch od stories so people can bring about change. If an individual contributor has an idea that is going to make a product better, they should be empowered to go execute that idea.

David Smooke: Where do you see ideas going awry though, with an Agile structure?

Dane Lyons: In order for Agile to work well, you’ve got to really spend a lot of tie thinking about our KPIs. You got to have … You’ve got to come up with a theory about which numbers you want to focus on to prove that you’re building a better product and if you can do that well, then you can make a determination whether any change that you bring to the product is actually positive or negative.

David Smooke: Yeah, I really being a results driven operation. It’s just, there are opinions everywhere but agreeing what facts and what numbers we’re trying to move definitely aligns peoples skulls in a way that gives them a freedom to approach the problem in more creative ways.

Dane Lyons: But there is a challenge in your KPIs, where a lot of times people are not very good at determining what KPIs they should actually pursue and a lot of times, either your KPIs are too high level, so individual contributors can’t actually do anything to influence those numbers or they’re … you might be tempted to go after a lower level KPI that is actually not correlated with, say, revenue. Say if revenue is you highest level KPI, you might get seduced into having some other metric that everybody believes in but everybody’s focused on moving that metric, but it doesn’t actually make a better product in the end.

David Smooke: So Hacker Noon is moving from two employees, me and my wife and partner, Linh. And now we’re moving to having an interim CTO, you, and then two to three developers, and so we’re looking at kind of a small team, like many start ups and being lean and small. How many KPIs do you realistically think a company of this size should have?

Dane Lyons: In my mind, I think it’s good to have at least two KPIs. A lot of people will try to focus on a single KPI but I think you need to have your high level compass KPI and maybe that is revenue, in a lot of cases for companies, and then you need to have a lower level KPI, which is a lot more relevant for the product. So maybe, on the product side of things, you might have a KPI like how many new stories are being created or written, how many drafts are being created. You’ve got a lot of options. You can probably assume that the more stories that are created, the more revenue is going to be generated. And if you get into a thing where you’re optimizing for stories and you’re cranking out a bunch of stories and things aren’t going so well, then maybe you’ve got a quality problem and you need to adjust your KPIs. But it’s fine to start out with a KPI that optimizes for the number of stories.

David Smooke: Yeah, I’m kind of gone through that exact problem and I like to think of getting to the 80% and to having that be a qualitative thing. And that’s yes or no and then once that’s yes or no, that yes multiplied by stories published is your real rate of publishing, because in trying to publish a lot, you make mistakes and you figure out ways where you should have quality control higher and I haven’t. But once it’s of a certain level, it’s more up to the community and I think a lot about serving contributors, serving the larger community, the reader, and then, making money. And that’s kind of where the three KPIs will tier down from.

Dane Lyons: That makes sense.

David Smooke: So we worked together, also, on a number of projects that have had different degrees of success, with things like starting 42 Hire and Paid Story. And what have you learned over the last … because you’ve kind of gone back and forth of creating a lot of early products, then being in a, not a large engineering team, but being in more structured set ups and larger teams. How are you approaching building MVPs now, than you were earlier in your career?

Dane Lyons: Well, I’m definitely moving much more towards building functional prototypes, so, as we were talking about the importance of change earlier, I think that it makes a lot of sense to consider the cost of change. And if you’re building a … if you’re introducing change into a fully formed product, like a complex product that’s already in the marketplace, then the cost of change is pretty high. As an engineer, you’ve got to worry about testing, you’ve got to worry about a lot of considerations. There’s so many dependencies to deal with and so any kind of change that you bring about is usually pretty expensive. But you could also go the other route of trying to experiment outside of the product and to build functional prototypes that explore ideas and try to get validation where the cot of change is very low. And I think that there’s a lot of value in doing that.

David Smooke: Let’s just jump around a little bit.

Dane Lyons: Okay, sounds good.

David Smooke: So, Dane, this is the Hacker Noon podcast, could you tell us about some of your work and life hacks?

Dane Lyons: My life hack is all about productivity. I find that, as an engineer, there are just so many distractions in life and for me to kind of avoid some of those, I kind of use a variation of the Pomodoro Technique. I like to keep what a call a Captain’s Log, so I will write down a goal for the next 20 minutes and then I will set a timer for 20 minutes, it could be 20 minutes to an hour. So you write down a goal, you set a timer, and you work like crazy. And at the end of that timer, you have to write what you accomplished and by writing what I … by having that accountability to myself to write down what I accomplished, I find that I get so much more work done during that kind of mini sprint of 20 minutes.

David Smooke: Have you been able to keep going strong and for weeks at a time like this?

Dane Lyons: I think the longest I’ve gone is probably about a week and I’m actually quite happy when I’m doing that, The trick that I’ve run … well, one problem that I’ve run into quite often is, being able to manage all of it. So what I would really like is an application where I have an easily … a visual indicator of all of the 20 minutes sprints that I’ve accomplished over a given week and an easy way to dive into those sprints and to, I don’t know I they’re called sprints, I don’t know what they’re called, but an easy way to dive into those notes. If I could have a better tool for organizing those, I think I’d be a lot more disciplined in using it.

David Smooke: Yeah, that would be a good app to create but we diverge. So could you tell me a little bit about v1Labs and where you think you’re going?

Dane Lyons: Yeah, so v1Labs is all about helping companies introduce a discovery track. A lot of Agile teams have what I would consider a delivery track where you have a bunch of tasks that … you have a backlog of tasks that get delivered. And like we we were discussing earlier, you need to … it’s a lot more … it’s a lot … how should I say this? It’s difficult to just have a pure delivery track and to innovate. You really need to be investing a lot of your energy into figuring out what makes sense for your product going forward and to validate those assumptions. So v1Labs is all about that. It’s helping you to get started on building a new feature or a new product, just-

David Smooke: Or a new content management system?

Dane Lyons: Yep, or a new content management system, whatever it may be.

David Smooke: Yeah, content management, man, it’s wild. WordPress is 30% of the internet. No one really likes it. Innovative new things are being built and a lot of them … it’s a very … if you go after the whole content management system market, it’s a gigantic market but it’s a real pain to fit your use case to everyone and I don’t want to fit my use case to everyone. I’m of the philosophy of, cot 90% of WordPress, give me the right 10% and I will have a site that’s 10 times better than if I built my site on WordPress.

David Smooke: Now there’s this thing of where I’ve been thinking more of products of … I just don’t want any extra stuff because it’s not essential. I’ve been kind of forcing the Hacker Noon workflow into the media and publication workflow and it helped us grow but it also limited us with things like, just how contributors communicate with their community and … I’m of the opinion, if you don’t have their email address, they’re not your follower. And it’s just, trying to think about how we’re going to reinvent this and take what’s working and fix how contributors want to work … I guess I’m rambling here.

Dane Lyons: No, that all makes a lot of sense. I’ve actually had quite a bit of experience working with building WordPress websites and blogs. And I think that they were a fantastic innovation compared to what came before WordPress, which was, God, it must be 15 years ago or maybe not quite that long, maybe it was 12, 13 years ago. So they came out with this blogging platform, which was great and they had a plug-in system where you could extend the functionality and it wasn’t a great system. I think, even today, they do some really … they do a poor job of making sure that the plug-ins that you’ve added to your WordPress implementation are not going to break your blog or … and are compatible with the other plug-ins. So, they don’t do a very good job of that. I don’t think they … I don’t think the database is structured very well, to be honest. And I think that they could do a lot of improvements.

Dane Lyons: But I definitely agree that, especially for Hacker Noon, that it doesn’t make a lot of sense to try to take WordPress and to kind of add all the functionality that you need and to kind of pare down all of the stuff that you don’t need. It’s a complete mess for that sort of thing. You’re much better off just building something from scratch.

David Smooke: Yeah, it’s definitely the conclusion that I’ve reached. Yeah, it’s like the deeper you get, the more you just don’t want to be reliant on other people and trusting these other firms to continue supporting the technology that they support. It’s a trust thing and then it’s, how trustworthy is the brand? And if the brand … WordPress is a great 30% of the internet and they’re still not making that much money for being 30% of the internet, and powering 30% of websites.

David Smooke: There’s always a balance between building to help other people’s business versus building to help you own. When your own business is struggling, you see it in how you start to treat other people around the internet and so I really want to have sustainability and sufficiency from the ground up. And a lot of people are in spots like me, so could you talk a little bit about taking wireframes and moving them to interactive prototypes?

Dane Lyons: Actually, I wanted to come back to the WordPress thing, just a second, and then we can move on to the wireframes. So one thing that I think is pretty interesting about using WordPress to run your product, is their plug-in system. As we were talking bout earlier, when it comes to introducing change, you actually don’t have a lot of change when it comes to WordPress. You have a lot of plug-ins that you can just pop into your application and add functionality but to iterate on that functionality, it’s really difficult because you’ve got to go and learn the code base of each individual plug-in, and sometimes you’ve got to just take it upon yourself to go and change the functionality if it doesn’t do what you want.

Dane Lyons: So, a lot of times people will go and they’ll rotate through plug-ins to try to find the perfect commenting plug-in or the perfect caption plug-in, and nothing really kind of solves the problem so then you have the burden of going and trying modify one of the existing plug-ins. It’s not conducive to change, I guess is what I’m trying to say.

David Smooke: Yeah, I think software has a longer reaching problem of, a lot of it’s with the incentives of how software is sold, where it’s just, you want to be able check this box, that it always has more, like this RFP structure. And then, whenever you’re saying you can do any type of software on top of this software, it’s like, “Oh, yeah, you make it so easy to add software but you make it so hard to remove.” It’s like, you can’t change if you can’t remove. You should be removing as much as you’re gaining. At the end of the day, you change stuff but you come back to an equilibrium and if your whole point is just adding more and more and more and more features, their going to get in the way of each other.

Dane Lyons: Oh, absolutely. I think that you really need to pare down your feature site to the minimum number of features required to satisfy the use case.

David Smooke: Yeah, and then it’s, how many use cases do you support? How big of a company are you? Which of these use cases is really your business and which one is not your business? It’s cool to think about where software and strategy overlap and when they work together well, you understand how it serves the contributing write and then, what contributing writer do you want to serve. If we’ve become a destination where just brands publish press releases, we’re not going to be a very good site. And by the same token, if we become a destination where it’s just only product managers talking about why Agile is better than Scrub, then we’re also not going to be a very good site. But how we support the contributors will drive the editorial line, then the editorial line will drive how we support the contributors.

Dane Lyons: Right. Yeah, it’s a delicate balance because so many start ups are pretty much forced to go up market way too fast, and to not focus on smaller use cases, just because they’ve got to satisfy an investor and I think if you kind of go up market too fast for the wrong reasons, it can kill you.

David Smooke: Yeah. Yeah, some investors are on the other side of things, where they only satisfies the users that really love you and don’t worry about losing your 80% middle of users that use you but may not tomorrow.

Dane Lyons: Right. Yeah, from my point of view, I think it makes sense to really go and try to find the best way to satisfy an individual use case and to be relatively focused and once you’ve kind of cracked that nut, then you can try to understand if your product makes sense for other adjacent use cases or is you need to build a complimentary product. But you can kind of expand that way, just kill it and satisfy one use case at a time, I think. That’s kind of the way that I think about it.

David Smooke: Day-to-day, you’re living in San Francisco, where I used to live, once upon a time, and it’s an expensive city, it’s a unique city. It’s got people from all walks of life. How are you hacking your life to live a better San Francisco?

Dane Lyons: Oh, man, I’m probably not, to be honest. I probably should not be in San Francisco, to be perfectly honest, because I’m not really taking advantage of the community here. I would love to spend a few weeks or a month trying to think about this problem and trying to understand how to maximize my time in San Francisco and then to try to build around that because I-

David Smooke: No, that’s time away from spending it in San Francisco, with other humans and what-have-you.

Dane Lyons: Yep. But I think that there’s plenty of technology that can help solve this problem because I mean, there’s so many events that happen and it’s kind of an informational problem. You need to be able to be exposed to the information to know when the events are available and which events you should be attending and who you should be networking with and all those sort of things. Almost like a personal CRM or something like that.

David Smooke: And then artificial intelligence will monitor your activity and recommend related activities.

Dane Lyons: Oh yeah. Yep.

David Smooke: I don’t know. I hope … You’re always been a little more optimistic of the machines taking over than I have.

Dane Lyons: Yeah, I’m pretty optimistic. I think that … I don’t know. I think machine learning is just a fantastic thing. Recently, AlphaZero beat Stockfish in their first, machine learning, or Google’s first machine learning foray into creating an AI for playing chess and it’s just fantastic, to see how creative this AI has become. But the AI that Google used to create this chess-playing machine, is not going to hack your vehicle and drive you off a bridge or something like that. It’s just … that’s not really something that I worry about too much. If there is some kind of hacking attempt that I fall victim to, I think it’s probably a much more malicious intent, rather than AI going around. I think that the risk of that is far higher.

David Smooke: Yeah, we publish Google’s Chief Data Science Officer and one of her best posts, she puts, “Machine learning is the simplest thing it is, is a thing labeler.” And that’s where it all starts. How can you label this thing properly and then what do you learn from labeling a million things in a million different places for a million different whatever. It is kind of … thinking about searching photos, is kind of an area where Google’s done a lot and it’s just interesting, especially like there’s a piece of art in my office, it’s abstract and there’s shapes and there’s circles, is this going to be searchable under rectangle and circle or is it searchable under art? And then, within the art, what other things does it recognize? Does it recognize that this is a sunset or is that possible? An it is, yeah.

Dane Lyons: It’s a very purpose built application. It’s not like that AI is learning how to walk or to do things that … or even has the ability to do the things that are malicious in nature. Certainly there is a possibility of that something like that could happen but for the most part, I don’t think that the risk of AI, as we see it today, is very high.

David Smooke: Yeah, but it’ll make mistakes and that’s why, when a Tesla car crashes, it’s in 10 million sites and whenever the crash that happens outside your office happens, it doesn’t make the newspaper. It’s under a microscope right now because of the implication of switching all drivers to self-driving cars. It’s massive on a labor level, a safety level, all that, and a performance level. Self-driving cars, that’s the one where I just can’t get my eye off that industry, you know what I mean?

Dane Lyons: Yeah, definitely.

David Smooke: It hits those things of those bigger tech themes versus practical day-to-day application of running into everyone’s lives from a business perspective and a user, a writer. Dane, what do you think I should prioritize in the new Hacker Noon?

Dane Lyons: Well, I think that you should embrace a system that empowers your contributors. I actually don’t necessarily … Well, I definitely don’t like to call contributors “resources”, that drives me crazy. I think that contributors should be called collaborators. And so the engineers that you hire and the designers that you hire, I think that you need to empower them to make change and to feel like they have the ability to do that. And that, I think, should be the number one priority.

David Smooke: Cool. So we’re now hiring front-end, back-end, you can apply at jobs.hackernoon.com or you can just email Dane, Dane@hanckernoon.com. That’s D-A-N-E, Dane, that’s you.

Dane Lyons: Yep, that’s me.

David Smooke: And where else can people find you, Dane?

Dane Lyons: Also Dane@v1labs.com, or on twitter at duilen, that’s my handle, and that’s pretty much your safest bet.

David Smooke: And duilen, your handle, where’d that come from?

Dane Lyons: I don’t know. I was just a … it’s a name that I invented probably about 15, 20 years ago and it just kind of stuck wit me.

David Smooke: And that’s your internet name, you internet given name, self-given?

Dane Lyons: Yeah, I think I probably made it up in AOL days or something, I don’t know. It’s just kind of stuck with me.

David Smooke: The early days of the internet. And anything the internet should be doing to be a better place?

Dane Lyons: There’s a lot problems with the internet. There’s a whole cyber-bullying thing and people just need to be more, I don’t know, open and tolerant to their neighbors and just good neighbors, I guess. If people were just better neighbors on the internet, just like in real-life, I think that the world would be a better place.

David Smooke: I do too. Hey, thanks for doing the Hacker Noon podcast. You can find more episodes at podcast.hackernoon.com or just visit hackernoon.com and I’m sure you’ll bump into a great tech story. Peace.