Product Agility

9 Million Users from a Full Stack Product Legend - An Interview With Gojko Adzic

September 11, 2024 • Ben Maynard & Gojko Adzic • Season 2 • Episode 33

Share episode

Send us a text

What does it take to build products that scale to millions of users? Gojko Adzic, renowned author and creator of multiple industry-shaping tools, shares his journey from coding enthusiast to full-stack product legend. Gojko has written some of the most influential books in product management and agile practices, including Specification by Example, Impact Mapping, and his latest, Lizard Optimization. His tools like MindMup and Narrakit have garnered millions of users, and in this episode, Gojko opens up about the strategies and lessons that fueled this growth.

Gojko discusses how he built Narrakit from scratch, overcoming product management challenges and optimizing user experience at every stage. He introduces the concept of "Lizard Optimization," revealing how unconventional user behaviour can lead to breakthrough innovations in product development. Whether you're a product manager, developer, or agile enthusiast, this episode is packed with actionable insights.

Key Takeaways:

  • Scaling to Millions: Discover the strategies Gojko used to grow Narrakit to over 9 million active users.
  • Product Optimization for All: Learn how Gojko designed his tools to be accessible to both technical and non-technical users alike.
  • Lizard Optimization: Gojko introduces his unique framework for identifying opportunities from unexpected user behaviours.
  • Turning Feedback into Growth: Hear how Gojko translates customer feedback into valuable product enhancements.

Perfect for product managers, agile coaches, and anyone looking to take their products to the next level, this episode will leave you inspired to think differently about user feedback and product growth.

Links and Resources:

Tune in now to gain deep insights from one of the most influential voices in product management and agile practices!

Host Bio

Ben is a seasoned expert in product agility coaching, unleashing the potential of people and products. With over a decade of experience, his focus now is product-led growth & agility in organisations of all sizes.

Stay up-to-date with us on our social media📱!

Ben Maynard

🔗 http://bitly.ws/GFwi

🐦 http://bitly.ws/GFwq

💻 http://bitly.ws/GFwz

Product Agility Podcast

🔗 http://bitly.ws/FdVJ

🐦 http://bitly.ws/FdVT

🤳 http://bitly.ws/FdW9

🎶 http://bitly.ws/FdWj

🎥 http://bitly.ws/FdWy

💻 http://bitly.ws/GFuS

👤 http://bitly.ws/GFvy


Listen & Share On Spotify & iTunes


Want to come on the podcast?

Want to be a guest or have a guest request? Let us know here https://bit.ly/49osN80

I can't get rid of this Bond villain accent. You know that. So what I try to do, it sounds like I'm, I'm, I'm blackmailing Liam Neeson on the other end of the line and, and he's going to come after me very soon. So it is not a good voice for the videos, is it? Welcome to the productivity podcast, the missing link between agile and product. The purpose of this podcast is to share practical tips, strategies and stories from world class thought leaders and practitioners. Why I hear you ask? Well, I want to increase your knowledge and your motivation to experiment so that together we can create ever more successful products. My name is Ben Maynard and I'm your host. What has driven me for the last decade to bridge the gap between agility and product is a deep rooted belief that people and products evolving together can achieve mutual. This is the Product GT Podcast, as you've probably ascertained from the introduction, unless you skip the introduction, in which case I'll say it again. This is the Product GT podcast and I'm your host, Bill Maynard. Today we're joined by somebody that I first met trying to figure this out a long time ago, over 10 years, Kindergarten, Kindergarten that long ago, that long ago when I was working at the Royal Bank of Scotland, we got this gentleman in because we were just blown away by two of his books bridging the communications gap and specification by example. We were trying to release a a big old chunky product and we had began to pay him of automated testing. So we've got this person named and then there's a hand. But since then, this person has just been prolific in what he's done and what is given to the agile and product communities. His name is, I'd like to pronounce kind of correctly. Yes. Excellent. Thank you for trying it. I tried, I tried. I always like to try. And he is an author of specification by example, Bridge, new communication Gap, Impact Mapping, a new book, a lizard optimization and five other books I think as well. Oh, you've got your user story mapping book as well, which I think, or user stories. Yeah, I'm up to 9 now, nine books, creator of a really useful tool I was using many years ago called mind map and now creator of a tool, which is a tool where you can try and convert PowerPoints into of movies and all day using their text to speech. So today we're going to be exploring narrow key, his journey in creating it, the product management bumps and hurdles and really looking to understand more about just what is lizard optimization. So thank you for coming on. Thank you for inviting me to come on. It's always lovely to talk to you. Thank you. You say that and yeah, it's been 10 years, so not that lovely that you haven't. We haven't made any effort to do this earlier. OK, That you got me there. So not that lovely, no, but they got an experience for me, so it's tolerable. It's tolerable. No, I am yeah. I've always been a huge fan. I'm just in awe and I'm I'm genuinely, genuinely over the moon. You've made time to I mean, just have a conversation with me, right. The fact it's recorded and it's going to go out to all of our hundreds. Just a bonus for them really. I am yeah. I'm I'm really excited. I wish that we could have done it face to face, but maybe one day if you when you get out from behind your keyboard for more of the coding you're doing, maybe somewhere anyway. Anyway, for those of you for those listeners maybe don't know so much about you and want to know a little bit more, could you give us a quick potted history into how you got to where you are today? Yeah, I, I kind of, I, I've always enjoyed developing code. I, I started building code when I was a kid, copying and pasting from a German magazine about Commodore 64 and, and kind of making the computer do crazy stuff that I couldn't even understand what I was copying and pasting. Yeah, I, I, I kind of loved it so much that I really never considered doing anything else as a career apart from kind of programming. But when I, I started hitting all these bumps where people were giving me stuff to build that made no sense and they were kind of working in a way that created lots of bottlenecks and things that. So I started learning all I can about all the other stuff like testing and product management and things like that. And I I ended up writing a few books, mostly as a way to free up my short term memory. I stole that idea from Henrik Nieberg. I saw him talk at a conference where he said he writes books because he doesn't have enough shorter memory to, to keep everything he said and he's interested in too many things. And yes, some, you know, some of these bottlenecks that I ended up hitting other people hit as well. And, and they liked kind of my ideas on how to solve them. So I became a consultant, ended up travelling like mad, then realized that I like coding more than travelling and ended up building my own product. So, and I launched my map in 2014. I think we started building it in 2013 and I kind of started building narrow kit in in 2018. So it's been a while. And yeah, I'm these days mostly kind of building my own products and doing everything from product management, development, sales support, presales, after sales, troubleshooting OPS. Yeah, I'm, I'm, you know, making coffee. So I'm, I'm the as well So yeah, I, I learned how to make good coffee as well during the experience. So that that's kind of my my journey to here. I think you're such an incredibly unique individual, like you're so I don't know. I know it's amazed. It's like you do stuff and you do it really well, whether it's writing the book, whether it's creating a product. I just, I just I yeah, I fed. Speaking of brilliant. Well, you know, I'm really passionate about these things and I, I think. You know, when you're passionate about something, then the details matter. And that's like, you know, for the book that we did, I, I worked with a graphics designer called Nicola and we, we like, spent a lot of time on, on tiny my new details. You might not even notice when you're reading the book, but the end result is something that's very pleasurable to to read. Can't wait to read it. I must I haven't read it yet. I am I'm behind the curve. I have a huge well, it's not come out yet. It comes out this. So you're not OK go, but the podcast will come out after that. So I've got an excellent job here. So but but by the time the podcast comes out, you might have already read it. So you're aligned to your audience. Oh, it's hurt in my head. Is time linear? Does that even come into the conversation right? It's always time. Time is always a problem, especially in software systems. I had a it was a bank holiday in the UK on Monday and I woke up and the first week of four was God, I'm so pleased I don't work in a bank anymore having to do with systems and bank holidays and figuring out kind of trading calendars. I'm so, so happy birthday longer part of my life. Anyway, enough. The question is, you know, why are you allowed to take a bank holiday if you don't work in a bank? That's I've got all my money in banks. Maybe I've it's it's of interest. I get a day off. Funny. The one needs some holiday and and you're you're buying application holiday yourself. Yeah. Bank holidays in the UK used to be great because everything shut, you know, it was like and it was like locked down, you know, it was just like it was great. Nowadays everything's open like they like like, no, no, not everyone gets a tail for there were bank holidays would never work in survey because people don't keep their money in banks. They nobody trust the banks. So maybe on something you have to have just holiday, but then that has like a a different kind of meaning to it, doesn't it? Hey, let's let's get back on track a little bit. So, so Narakit, why is it called Narakit? And what? And, and I gave a really crappy overview, like to tell us a bit more about Naraki. Because you started this development, what, six years ago? Yeah. So I was like trying to do lots of demo and promo videos for my map. That's the other tool we built. And I, I hated the experience because first of all, I'm not a video editing professional. I, I can double in it, but it takes me a long time to do anything useful. The second thing is. I mean, I, I can't get rid of this Bond villain accent, you know that. So what I tried to do, it sounds like I'm, I'm blackmailing Liam Neeson on the other end of the line and he's going to come after me very soon. So it's not a good voice for 15 videos, is it? And so he used to take me about maybe two or three hours to do a 5 minute demo video of something new release. And I as a developer, I thought, well, you know, most of this has nothing to do with me. It's all repetitive. You have like an intro video, an intro sequence, an ultra sequence, a bunch of screenshots and some explanation there. Can I automate this? And and then I built a bunch of shell scripts that does the video editing thing and text to speech engine started becoming good enough around them. And I mean, they know they increased in quality significantly since then, but they were just becoming good enough. And I realized I can actually plug in a text to speech engine and just do the whole thing from a markdown script. And it used my time to make promo videos from, you know, two or three hours for a 5 minute video to about 10 minutes, which is wonderful time saving. And I thought other people might like this as well. So I kind of repackaged the shell scripts and built a tool around it and it was wonderful because the there was so much learning and so much discovery Every time I I did anything with that, because obviously the audience is not the same as I am. People were getting stuck and struggling in lots of different places. So for example, instead of just kind of building the script and things like that, I did a bunch of custom potential user interviews and it turns out lots of them were actually using PowerPoints to plan the video. They would use PowerPoint as a very quick way to storyboard things because people know how to use PowerPoint. And then from that storyboard, they would build a, a screencast or they would build a, a training video or something like that. And I thought, well, why am I asking people to build the PowerPoint to storyboard and then export the PowerPoint into images and then add voice overs and, and, and a bunch of other things. Why not just let them upload the power? And it was a huge revelation because then, you know, people started building a lot more videos than they were with scripts. It was a lot easier. And there was a wonderful example of, I think Kathy Sierra has this wonderful book for product people. I think everybody should read it. It's called Badass. And he talks about this kind of sax zone and, and the sax zone is the initial period of engagement where people don't feel confident enough that they're productive and most users get stuck in the sax zone. And a good product helps users escape the sax zone very quickly. And instead of this whole, you know, process where they would plan and then export and upload by letting them, I was allowing users to escape the SAX zone very quickly. And that significantly increased usage. It created a whole new category of users where marketing people started using it and marketing people got in touch with educators. So educators started using it. And this was kind of around the time where COVID struck and it was kind of, I guess everything was going digital. Lots of university professors had their lectures in PowerPoint. They needed to put it online and, and it was a godsend for them because it was very, very easy to kind of convert PowerPoint into online videos there. And he basically grew from there. It's, it's had about has about 9 million active users last year. People are building hundreds of, of thousands of things every day with it. I'm, I'm quite happy with how it turned out. And, you know, especially given the fact that I'm still the only person working on it. Yeah, I mean that. Yeah. I mean, what a story. And so it started. You were solving a problem for you. Yeah. You scratch your own itch. I mean that I think that's how lots of products get started by by scratching your own itch because you kind of see a problem and you you want to solve it. And then you did some user interviews. You learnt more about how people were kind of preparing for the, the process of creating a video and just thought, well, why don't you just shortcut that and just to kind of use that raw. Yeah, the raw deck they were using to plant the video to create something themselves. And so the videos that get created, are they are they just it just transitions through the slides or does it do anything more? Well, I mean, mostly it transitions through the slides, but you can include like videos into PowerPoint as well. So a lot of people are using it to, you know, just marketing agencies are using it to prototype their videos with customers. Educators are using it to create online lectures, people are using it to do product demo videos, kind of explainer videos and, and screencasts and a bunch of other things. And so then from all of this, you've come up with this idea of lizard optimization. Which is the title of your new book Yeah, that's, that's the new book. I, I, it was because I'm the only person working on this, it was relatively easy for me to kind of short circuit the distance between product management and support. In lots of organizations you have this kind of full customer engagement, customer service in customer support part where it's totally disconnected from product management. And yeah, the, the people are facing problems every day or they're getting stuck. And it, it, it's when it's propagates to product management, it's often, you know, triaged put either as a bug word or as a kind of potential improvement. And then very, very rarely kind of it gets into the actual product unless it's a critical thing. And I was able to kind of short circle that and complement the research that I've done before doing something with feedback after it was done. And, and it was amazing because unlocked exponential growth. And, and that was a big lesson for me how optimizing the product for people who are like really getting stuck with it because of weird edge cases. And you might call them non-technical, you might call them stupid, you might call them, you know, whatever you want offensive and, and, and developers often do that. But the fact is that users are getting stuck somewhere. And sometimes these people are just, you know, can't be helped. Sometimes they're getting stuck because they're trying to do something malicious. But sometimes they're getting stuck because they have a genuine use case and, and they are not being able to solve it. And then they end up doing crazy workarounds. They end up doing crazy things you, you, you never thought about. And by doing, looking at that, I've realized I can extract very useful signals for improving the product. For example, there were a bunch of users who were paying me to create blank videos. And sounds like it sounds like good money to me. It sounds like good money. But it was something I really couldn't understand. And I thought it was a bugging decision because the videos were all coming out white with sound, but white. And I was thinking like, this is insane. There's a bug in the, you know, video production pipeline somewhere people are going to ask for chargebacks. I need to fix this. And I spent couple of days trying to figure out what went wrong and I couldn't. So I started talking to a few of these people to say, you know, maybe they're using some other tool to create PowerPoint and my PowerPoint reader is not compatible with that. So it's coming out as as a blank screen, but no that we're actually using PowerPoint. And then asked like one guy, well, you know, if you open the PowerPoint, do you see anything there? Do you see anything in the slides? And said, no, no, it's blank. And I'm like, well, OK, that's so PowerPoint is broken. So no, it's not broken. I, I, I don't have anything in the slides. And I was like, well, why would you be paying me to create the white video? I mean, and it turns out that why, you know, doing this kind of optimization for, for people who are getting stuck, I ended up creating a very easy way for people to configure text to speech conversion that was much easier than with alternative tools. And some people just needed the audio. They didn't need the video and, and they were going through the hassle of creating a blank PowerPoint just so they can configure the text to speech conversion. And, and then they were converting it into a video to extract an audio track to use for voicemail messages, for audio books, for podcasts and, and things like that. And again, this is one of those examples of, you know, people getting stuck in the suck zone. Why would you force people to do that? Why not remove obstacles? And I kind of kept removing obstacles and, and, and this really unlocked lots of new use cases and unlocked exponential growth for, for the product. So the idea of lizard optimization comes from an article I read by Scott Alexander, who wrote about this thing called the Lizardman's constant. He invented it in research and and he he was doing some research that combined demographics with kind of psychological research. And they were trying to figure out would there be differences if you account for demographic data and how people here behave. And they were constantly finding weird answers in the demographic data. So, so American as agenda, agenda. Yeah, there were people who were doing Martian as nationality. And, you know, some of these are just malicious. Some of these are people clicking the wrong button by mistake, Some of these people being distracted or not understanding the question and things like that. And he said you need to kind of account for all these crazy answers because there will be some percentage of people who just, you know, do things you can't explain. And the percentage is not actually that small. He, he says that the lizardman's constant is 4% and out of just fun whenever I'm faced with some kind of total brain melting puzzle, like why would people be doing this? I I started thinking about, well, maybe you know, it's not a human, it's a lizard. And, and I kind of, you know, it's one of these, listen, men's constant thing and they have their own lizard logic and things like that. So I thought, well, you know, I'm, I'm actually optimizing the product a bit for lizards here. And that's where the idea of lizard optimization came in. But basically what's interesting about this is if you can reduce churn by 1%, it, it compounds growth massively. If, if you can help 1% of your users not get stuck and, and not drop off and, and, and be more successful, it can unlock a lot of growth. But it can also point to some really kind of interesting feedback. Because it doesn't matter how much research you do, the research has to stop at some point. You know, we, we, we, I think the, the product management community is much more aware of doing cheap and dirty research and easy research and iterative research and, and, and build, measure, learn and, and all of this stuff. But research has to stop at some point and you have to kind of get, get something out. And but when you get something out that the fact that you got something out changes the behavior of your users. That's a very interesting concept to think about. Before they had something and after they have something, their behavior changes. Bruce Tolazini, who's like a legendary UX designer, talks about this as a, as a complexity paradox. And he says basically the commuting time is fixed. The, the, the time you have for the task is fixed. And then if you give people a better tool, they will take on a more complicated task because they still have 10 minutes to do something or two hours to do something. So they take on more complicated tasks. So the very fact that you gave something to users changes their behavior, which means that the research you've done before releasing is not necessarily relevant for their behavior after releasing. And, and it kind of impacts itself. And this is, I think I really got me to think about all these cases when I was working as a developer and, and business analysts were giving me requirements and you do something and then the user wants something else and you think, well, the users don't know what they want, Not necessarily the the very fact that you're giving them something changes what they can do and, and they can do more. So basically the, the, the idea is to, to look at what the crazy people are doing, look at the things you don't understand and, and those are your blind spots in research. Those are things you couldn't spot because. What you delivered changed the context and now people are doing other things and then kind of to figure out good signals from that to improve the product. And I'll give you another example. So we create, I created this kind of screen where people can just upload the script instead of creating a video, just get a get a voice over. And that took off really nice. That was kind of optimization for the PowerPoint thing, but then people kept uploading files that I wasn't supporting. So I started supporting Word documents and PDFs. People started selecting APK files and Android packages. I, I, I I don't know what the right way to read APK file is. I still don't know that I have a couple of 100 people every day trying to do that. Still, my best guess is that it's one of these, you know, where you leave a USB key with a virus in front of a building and then somebody kind of loads it just to see what happens. So maybe that maybe maybe they think if I upload an APK file, somebody's going to load it on an Android app. And yeah, I thought explained, there's no way to explain it, but there were a few people every day uploading subtitle files. I never thought about that. And I well, you know, it's kind of reasonable to expect to produce a voice over from a subtitle file. And my first thinking was, OK, you know, I'll kind of ignore the time stamp, so I'll just read the text. And then people kept complaining how the audio is faster or slower than the original subtitles. I was like, well, yes, because this voice is reading faster or slower That's and then, you know, I started talking to them why they're doing it. And lots of people like they had a video with subtitles and they wanted to have an alternative audio tracking in a different language. So they were trying like dubbing. Dubbing. Yeah. Or, or just like not, you know, you have a corporate corporate training video in English. You want it in French and Spanish and things like that, but you don't want to waste time re editing the video for the differences in language and. I said, well, you know, this is like a couple of days of work to, to, to speed up or slow down the voice to match the timestamps. And it turned out to be like one of the most profitable things I've ever done. Because there's a big enterprise software company that kind of has 200,000 videos in, in, in I don't know how many languages. And for compliance reasons, they need to have it in every language that that they kind of operate in. And they basically just like found this thing. And instead of spending hours and hours in every video, they could use this to produce an alternate audio tracking minutes. And yeah, they were more than happy to pay for that. I have churches who have sermons and, and, and, and things like that, you know, then they want to operate in, in lots of different markets and they, they have a priest talking, they record them. Then they use some other software to produce the subtitles that translate and use my software to do it. So it's all these kind of crazy use cases that are really interesting, but that also led to some kind of quality of life improvements for everybody because they needed to check different segments. So I, I allowed them to kind of check segments easily and that allowed to be kind of wonderfully useful for other things. So generally the idea of wizard optimization is to look at the usages where people are misusing or abusing a system or, or using the system in the ways you don't expect and, and use that as signals to improve it. And kind of I nailed it down to four steps and, and the four steps for, for easy to, to make it easy for people to remember they're starting with LZRD, like listen. And the first is kind of learn how people are misusing your system. The second is to like 0 in on one behavior change because there's going to be a lot of signal here, a lot of noise. We need to figure out the right things to improve. That's consistent with our current strategy, with the current goals, with the current talkers and things that you don't want to improve everything all the time and you don't want to follow every lizard there is. We want to use that data to extract the useful signal. And then we need to remove obstacles to user success. That's the art part. And The thing is interesting, we have lizards interacting with your system. All the acquisition and marketing is already done. That's the most difficult part in my opinion, because I'm a developer and kind of a double in product management and testing. I'm really bad at marketing and sales. So if somebody's already interacting with my product, the, the, the most difficult part is done, but they're hitting a roadblock. They're hitting a boundary. So they, they remove obstacles to their success is, and that's how you know, optimize, get them out of the sax zone, remove PowerPoint, get them to kind of do, do a video and optimize as much as possible what people can do in the 1st 5 minutes, 10 minutes with your software and reduce the time it takes to do a complex task, get people to be productive, remove obstacles to their success. And then the last part is really to double check and to make sure that what I thought was a good idea is actually a good idea. Because remember, we're talking about people who don't follow the same logic as the product management at this point. And, and what you think is a good idea might not be and they might do something different. They might not get there all the way, and this is where operational metrics and observability and and user behavior metrics come in quite a lot. So there's like a slightly less formal and easy way to remember this kind of LCDRD is Lizard Alert, zoom in, rescue them and then double check and, and kind of going through this cycle complements research. Going through the cycle complements research both in a way of kind of identifying blind spots that we didn't know about, but also identifying the changed context after we've done something. What now, how do people behave differently now? How do you figure out what not to do something about? Because if you have a product or let's say trying to follow some kind of plan as a paradigm to this, because I'm thinking here about agile and responding to change over following a path. Some people see that as always want to respond to change. Whatever the change is, let's just go and do it Because surely the respond to the change is more important than sticking to the thing we're doing before or achieving our product vision or reaching our goal, whatever it might be. So how do you make the decisions for your product as to what stuff you should then go and do something about versus do you know what Mrs. Lizard doing it and I would never, can never go down that route. We don't want it to be used for that particular purpose. So I think there's, there's a couple of things there. One thing that really helped me as a mental model with, with my products is the five stages of growth from the linear analytics book. And I think it's a wonderful model. It's, it's not that well known. And I think more people should kind of know about that. They, they, they looked at lots of different organizations and, and, and growths, product growth and how the goals changed through, through the growth for successful products and less successful products. And they created a sequence of goals that's logical to look at. And they, I, I, I, I wrote about that in the user stories book and I mentioned it in the lizard optimization book, because I think it's really a good first pass for aligning on, on big ticket items and goals. So the five stages are basically empathy where we need to prove that we're actually solving some problem that that's kind of important. Then they talk about stickiness is the next stage and, and stickiness is the stage where we need to prove that there is a potential market for it. And, and the way they talk about it is that if people are using your product a lot and if if they reach for your product when they have this problem, then there's a potential market for it. You've solved a problem that's important to them in a good way. Then they talk about kind of the next stage. They call it virality. I really hate the name because it's kind of, you know, startupy and you can't talk about virality in, in large organizations, but I, I, I think kind of, if you rename it growth, then then it, it makes more sense. So at this point, so you've proven that you've solved the problem that's important enough. There is a potential market for it. Now you need to prove that you can actually reach that one because maybe the market is out of reach, that the competitors are too strong or maybe the market is not big enough or you. So kind of in that stage we're looking at usage numbers, raw usage numbers, user numbers, retention numbers, and things like that. Then the, the next stage is revenue, which is kind of self-explanatory, but kind of for nonprofit orgs, I like to talk about extracting value from the product and, and figuring out how can we scale the value extraction at this point? Not the value extraction for the users, but the value extraction for you or or the business or, or. And then the step after that they call scale, which is kind of all about consolidating things, reducing operational costs, reducing the cost of delivery, optimizing how we work. And what they talk about is there's a logical sequence to these things and you shouldn't really focus on, on optimizing how cheaply your product runs. If not enough people are paying for it, that that's because you can always optimize how cheap your product runs. If there's not, you know, if there's no market, it's pointless. And looking at this as as a sequence really helped me figure out what went wrong with a couple of things I did in the past. We had one case where we really nailed down the first stage, you know, retrospective looking at this. We, we, we solved the useful problem, but we didn't solve the problem of stickiness. I, I don't remember the exact numbers, so I'm just inventing some numbers now. But let's say that out of 1000 people that came to the website, one of them stayed on to do something useful. And then we kind of skipped that stage and we tried to do the growth stage and we invested a lot in marketing. We brought lots of people to the website, but very few people stayed on. If we had inverted that, if we optimize the stickiness part and then did kind of marketing and growth, the, the, the outcome would have been completely different. So this product went bankrupt at the end because we spent all, all the money we had foolishly looking at this, you know, for Ericita was like very, very systematically following that. And, and this gives you the goals you want to focus on. In the first stage, you focus on kind of things around solving, you know, this specific problem, figuring out, optimizing that, getting people to escape the SAC zone and things like that. Then in the next stage, you focus on things that help people stay around for longer, come back more frequently, do more tasks through you and, and, and things like that. Then on the next stage, you look at how can we get our users to help us grow? How can we grow easier? How can we optimize marketing and and things that and then kind of I think the product is, is now in kind of the revenue extraction stage. So I'm looking at both. Kind of, you know, how do I get people to do more through me to pay for larger plans and things like that. And while at the same time, I like to use the other stages as constraints, as kind of what makes a box, you know, around what's viable. So for example, in the growth stage, I would look at how do I grow but not damage operational costs too much or how do I grow but not damage revenue too much or not damage stickiness and things that. So I use one of these things as something I want to achieve and the others as like bounding the solution to make it viable. And and that helped me quite a lot. For example, there was an insane episode a few years ago where the usage doubled overnight and I was trying to optimize the frequency of usage. I was like, oh, I must have done something amazing. You know, the usage doubled overnight, but you know, if something's too good to be true, it usually is. So then I started looking at the metrics and, and thinking about is there a bug? Maybe I'm just resubmitting every task twice. Maybe you know, that maybe there's, there's something weird going on and I couldn't figure that out. And I, I started looking at lots of different data sources to figure out is this genuine or not? And I realized that Google search is sending me like twice as many users as it did the day before, which is amazing because I, I did not optimize anything around kind of search engines and things. And so Google occasionally changes their algorithm, you know, through no fault of you, yours or no, no credit of yours, you might go up or down in search results and pay, you know, maybe nice benefit. And then kind of the usage doubled the day after as well. It's like, wow, this is, this is totally amazing. It's so there was a point where people are building so 100,000 audio and video files per day jumped to 200 to 500. And at this point I was like, you know, ready to celebrate, but I, I had a box around this for operational costs and revenue. The, the product uses a freemium model where basically the commercial users are paying for the cost of free users. And there's lots and lots of free users, small percentage of commercial users. When they balance, it works out OK. Unfortunately, we have a massive growth of free users. There's not enough commercial uses. They need to catch up. They need to experience the system well enough to catch up. And because this is all self funded, when your growth accelerates so much and, and, and the usage increases 5 * / a couple of days, the operational cost increase five times, the profit doesn't. And it was kind of, if it continues like that, the product will run out of money. It's my money. So I can't allow that. And I, I had to like act very quickly and figure out what's going on. And the usage pattern of the new people was significantly different from the old people. They were spending the free tier up to the limit. And then very, very few of them were upgraded, which was really weird. And, and then I started looking at what's going on and almost everybody, like 90 something percent people were coming from India not upgrading. Now my, my first thought was, OK, I have a payment processing problem with II has these weird banking rules. People don't have cards and things that, but that wasn't true because if I looked at like I had customers from India before, there was a much, much smaller percentage even of those upgrading. And in order to avoid basically going bankrupt, I realized that the growth is just, you know, growing too fast. So the bounding box around viability is wrong. So I temporarily blocked people from India by GGIP and that kind of solve the problem for, for a day or two. But then the rest of the world just started kind of popping up like mud and people were building Hindi language audio from all over the world, which makes no sense. So then I realized, OK, there's something, there's something really weird going on here. And I, I, I kind of tried to figure out what to do and one of the kind of statistics that came out that was really weird is my bot detection algorithm was, was being triggered all the time. Most of this free usage was coming from data centers, not from residential Internet connections. So realized, OK, I've shut down India. But then what happened is. They just started using VPNs to switch and, and, you know, go go around the world. So I kind of blocked VPNs and and then I started getting hundreds of weird emails every day. The guy who was saying that he's too spiritual to use a regular Internet connection, he needs to use a VPN. Yeah, there was an insurance company and, and they were telling me how they blocked taxes in the system now because they're using VPN. I said, well, you know, that's not good, obviously. Can you give me your account name? And he said, oh, we don't have an account. We're using the free system. It's like, well, you're an insurance company, man. It's, it's like, and and they, they were, they were insisting that they turned their access back on because it's critical for the business and they've been used for, for, for a year. It's, it's like it's critical for the business workflows, but they're never considered actually creating a commercial account. And there were a couple of these kind of crazy emails, of course, you know, but there was one guy who was sending me emails from this throwaway Outlook numeric e-mail addresses where you can't even reply to them. And I know it's the same person because he sending the same e-mail like I'm blocked, I'm going to sue you. And you know, it's like, well, it's insane. It's just insane. And he then a week later the same person contacted me from a Gmail address and said, you know, he paid he paid quite a lot of money actually to create an account and and and and he was paying for access. But now I've blocked VPN. He was blocked and this was like a wonderful case of a lizard where basically this use case I've never even considered. I, I don't use VPNs in anger. I don't need to use VPNs in anger. You know, it's, it's, but he, he lived in a country that was kind of spying on, on the residents. He was a bit paranoid about the content he was creating and, and I think there's really no reason why I would block that, especially he's a commercial user. So what I did later is kind of allow people to use VPN, but only if they're a commercial user. And if they're not a commercial user, a nice kind of message comes up saying, well, you're trying to use a VPN, you know, to prevent abuse that's blocked for free users. If you're a commercial user, just sign in. Brilliant. If you're not a commercial user, here's a button to, you know, upgrade. As a result of that, basically we ended up the month with a 25% increase in revenue, which is quite good. There were lots of people like him using VPS for different purposes for whatever reason that I didn't consider before. And you know, now we can talk about, OK, that's a gap in my research. We can talk about how I should have done that, but effectively. But by figuring this thing out, I was able to, you know, very quickly, first of all, unlock growth and, and the second thing is prevent abuse. And I ended up actually finding at the end what caused the growth. There was a a video on YouTube of somebody demonstrating in Hindi how to abuse my system. Really. Yeah. And, and I, I couldn't understand Hindi at all, but I played the video and and I found that kind of at some point he talks about VPN, VPN, VPN, VPN. And you can see he's standing on VPN and, and kind of the interesting thing about this was that kind of the, the, the, the, the person was suggesting going through Google, because lots of sites that have a paywall kind of allow you to access content through Google. So the Google search increase was a byproduct of this where basically he was teaching people completely irrelevant for my website. I wasn't doing a paywall or something like that. But for newspapers and things that if you go through Google you, you get slightly better content, I guess. So the uptick in Google search was the result of this video and this person recommending for people when they're abusing the system to go through Google search rather than going directly. But as you know, the benefit of how Google works, the more people you click on it, the more relevance they give you. I got like a hammer effect of that for months and months and months. So going back to your original question, how do you figure out what to work on? I, I kind of tend to focus on the goals that are around the five stages of growth that I want to grow in, but then use the other stages as a bounding mechanism to kind of put constraints to put goal posts to put like the, the, the so not goal posts, the railguards or the bumpers on the road to prevent me from going off in, in, you know, the wrong direction. And when doing that, you can spot very easily if something you're doing is, is like not delivering the effect that you want or delivering a wrong effect or maybe delivering a different effect. But it's also key. And, and as long as you're in the right lane, why not? And, and that's how kind of this discovery works really well. What was the name of the book that came from Lean Analytics? Fantastic. All right, so our time is like coming to a close. I have I've learnt so much and there's so and I'm I'm going to be doing this afternoon. He's creating an account on narrow and having a play around with it because I'm I'm I'm intriguing upload an Android package file just so you can bump the analytics. So I will I will. Have you ever been tempted to install one of the APK's or like a No, no, no, I wouldn't even have what to install it on. I don't have like an Android mobile device, but maybe I should buy one. And just, you know, it's, it's so fascinating. Yeah. We just have to be connected to anything, you know, like, who knows? I mean, yeah, that's just me just being a little bit nefarious. Oh, it's so much. I mean, I've, I yeah. Just one of the narrow keep stories is fantastic. And the fact you're doing all of this is just is wonderful. And the way that you've approached it and just the, your ability to experience these things in real life, to get a products into market, to be coding it, to play all these roles. And still, you know, then find something new to to write about and to think about it. Empty your short term memory in a book, which is, I'm sure it's going to be fascinating. I can't wait to read it. It's what you were saying earlier about product management and support and closing that loop. You know, I think it's so critically important and I don't think enough people really appreciate that the, the teams at the end there who are going doing the custom support and didn't have the inquiries. Actually, that's a gold mine of information. And I think I was talking to Jeff got health about this and they're saying, Oh yeah, we were talking about like sensing and responding to what's happening. Actually, that is such a huge wealth of really valuable information. It's your sensor, you know, and, and kind of I, I for my map, because again, it's, it's a, it's a small self funded product and there were only two of us and we hated doing support. We, we kind of tried to optimize the product every time somebody came with a support case. And I've, I've later read about this from in the Amazon way, I think, or something like that. They mentioned how Jeff Bezos had this thing where he said where that people people should talk to their friends, not to their vendors and and. The fact that somebody's calling Amazon is a sign that something's gone really, really badly. They try to kind of use those interactions as a signal to figure out what went wrong. Why did you have to contact Amazon as a person and then to systematically prevent these things. And I, I, I, I really love that as an idea. And it felt like really consistent with what we're doing for Myanmar because in a sense, if you make the product good enough that it doesn't cost support, it is better. It's better for you and it's better for other people as well. Yeah. And kind of this whole listed optimization thing is just kind of taking that to another level, like not just solving the support stuff, but actually using it as an insight on where to start growing the product, what kind of ideas to to build so that you can solve new use cases, not just kind of, you know, fixed, fixed support cases. Was it? I don't think it was even told me this, but I think I read this around the time that we first met about the, IT was a tool that was bought by Cisco and it was originally started off as a support tool, but they found that no one was using like a lot for support functions, but the screen sharing functionality was being used like incredibly consistently. And that was all people were using it for. So they, they, they looked at that and said, OK, well, maybe that should be the product they carved out. And that became the, yeah, you know, this the lovely example in, in founders at Work by Jessica Linston. And, and in late 90s, like you had these people who were incredibly smart and they built a very, very efficient cryptography library that was able to run on low power devices very efficiently. And, and they, they tried to figure out the use for it. So they thought, well, cryptography, low power devices, maybe some kind of money transfer. So they, they build this thing where you can like bump 2 LB pilot devices together and transfer money from from 1 device to another to, to utilize the library they built. And then they built a website. It was late 90s. So, you know, web was coming up and they built a website to promote it. And the website was very rudimentary. It was a demo. It was not, not kind of intended to be used in anger. And then all of a sudden somebody realized that people are using the website for money transfers a lot more than they are using the pole pilot app. The the whole idea of the website was it was going to promote the pole pilot app, not kind of take over but. The company was actively fighting against money transfers on the website because they thought that's kind of distracting from the goals. It's not achieving the key results, it's not achieving the objective. And they were fighting for for more than a year against these people, prohibiting them from using it. And and there was kind of inexplicable to to the product management why, why this is taking off so much. But when they looked at the numbers at the end, they realized, well, this is the product is not the pump pilot transfer the product, you know, the website is the product and, and that company's PayPal, you know, they're now one of the biggest payment processes in the world # pilots no longer exist. And, and at the point where they figured out what to do, the website had 1.5 million users and, and the # pilot up had like 12,000 users. So it was kind of obvious what to do. But the, The thing is, so many organizations, product management fights against this like they did. We waste so much time trying to keep our vision clear and and, you know, steer towards the goals we set instead of trying to figure out, well, you know, this thing has taken off. Let's let's be opportunistic about that. I think that's yeah, for me, This is why I love why I love and also hate and have very little time for product visions. I think a good product visions for me it should be doesn't matter the like the the channel you're using the mechanism by which trying to solve the problem ultimate if you really understand the problem you're trying to solve and it doesn't matter how that manifests. I think when people are so, you know, for good reasons and for bad reasons, you know, just obsessed with the particular channel or the particular outlet if they're trying to solve the problem rather than actually saying they but what is the the problem here? Yeah, I mean, we get caught up just wasting our time and fighting against actually something which which is the thing that's going to make the big difference to us. I mean, there's, there's a lovely example of that, you know, totally unrelated to software in seeing like a state. That's another wonderful book everybody should read just for kind of systems thinking examples and things that talk about how you have like building a university, especially like several buildings and, and people need to go across and, and things like that. Lots of universities when, when they're kind of building up a new campus, they, they try to build roads or paths and, and things like that. And then they get angry when students are going across the lawn and and not going on a path and, and then they start to put up kind of rules and things like that. But the opposite is basically just not build any roads, build alone. So kind of students could go anywhere they want and then look at where the students are going to actually build paths. And that's kind of you, you complement the research you've done. You have some ideas, you have kind of your your plan, but you actually figure out where the people are going and and then you follow them to remove obstacles. And what's that desire lines? Yeah, I guess, yeah, probably, probably, yeah, yeah. And that that the book you mentioned there was that the 50s I I, I I found about that in seeing like a state. Yeah, OK, because I remember reading that I came it was in a book that I was reading when I was becoming a large scale scrum trainer. I'm the reading about that because Craig and bars talk about it in regards to architecture. Yeah, I mean, it is I I can you know, it's it's it's not particularly related to software or a specific type of product. I think figuring out what the reality and and understanding that reality might not match a plan is a wonderful thing. And it's like, what even match your vision? And then the question is, you know, do you follow the reality or do you get stuck in your vision? Yeah, exactly. Love it. Okay, thank you so much for giving me your time. In fact, it could be gracious and I just, yeah, I just, I love listening to your stories. There's so much great information. I think there's so much people there to listen and to make notes on, and so if you are listening to this and there's something here that you've really liked, then please do get a share on LinkedIn, share on Instagram. Even if we quite do some following Instagram, maybe even tick tock, I don't mind, but tell the world about it. I think there's some really awesome stuff. I think more people should hear if people do want to find out more information says about narrow keep, then is it narrow keep.com? Yeah, Narrow keep.com. Excellent. And if they want to find out more about you, you've got a website, haven't you? Yeah. Yeah. I, I kind of, I'm relatively easy to find online. Basically, you know, going with such a real name that it's, it's it's very unusual even for where I come from. And if you Google for my name, you like to see a a footballer who sleeps with supermodels. That's obviously not me. Or you'll see kind of a an East German actor in in spaghetti westerns. He's bare chested with feathers in in his hair. He's playing of American Indians. That's not me as well. I, I can't like don't have enough care for that. I'm, I'm the third one. So the third one I see to find online. Well, we'll make sure there's some links in the show notes so that people can find it easy to save them, save them typing. So yeah, thank you so much for coming on. It's been awesome. I'm going to spend my afternoon mucking around on narrow key. I think somebody I'm quite looking forward to. Do you have a Google Slides integration or do I have to use PowerPoint? I don't yet have Google Slides, but I am planning to do it at some point. Lots of people trying to do that. So that's part of the wizard optimization. OK, yeah, I'll try to do anything too weird to give you false signals and to see get you investigating stuff, which maybe it's just me digging around with it. So yeah, thank you so much for coming when everyone thank you so much for listening. We're going to be back again next week actually before we're up because anything else you want to share with people and if else you want to tell people about? Well, I can create a coupon code for the book and people can download the book at a discount if they're listening broadcast and maybe you can distribute the link. Yeah, absolutely. That makes us. So give that the. We will get that, we will put that in the show notes and we will use it on some of the promotion we're doing for this episode as well. This episode will be coming out around mid-september. So just for a bit off to productized where our number of listeners just always grows exponentially. So it'd be great times and a great market for you there I think as well the interesting people in regards to your book. So we'll get that done. Wonderful. Thank you. Brilliant. OK, so everyone, thank you for coming on. Everyone, thank you very much for listening. We're back again same time next week.

People on this episode

Head Shot

Ben Maynard

Host