speaker-0: Hello my dear Bajans! This is a bonus episode. I was the guest on another podcast called One Thing That Worked. And Sergei, the host, was kind enough to invite me there to talk about the show and how I started actually learning Bajan statistics. And I know that a lot of you ask me a lot about my background and how I started the show. So I do talk a bit about my background in this episode. Not too much, I talk more about the show. So maybe if you know another podcast where I should go and talk about my background, let me know. But for that one, we'll focus mainly on the show and what it took to build a podcast that now reached the top 1.5% globally among millions of podcasts. So thank you so much folks for your support and for listening all the time to this show. It's literally thanks to you that we are here. And Doing that without chasing a broad audience or turning it into a business. So in this episode, I share with Sergei how consistency, curiosity, and creating the show I wished existed at the time to help me learn patience tests. Well, in the end, how that show helped me find you guys, this audience that I am so grateful for. And we also talk about choosing a niche, learning from failure, building a podcast around the audience you genuinely want to help. So I hope you will like this bonus episode and keep in touch and of course tune in for the following classic episode coming very soon in your feed. See you soon, my dear patients. And I love them, I have a great audience who is usually very technical and loves to get into the details and being told that they are wrong, which is the best audience you can ask for. And so yeah, now the show is actually among the one point five percent most popular shows out of almost four million podcasts globally. So yeah, I could not have guessed that at all. And interestingly, I didn't create the show actually for it to be Listen to but I create the need for it.

speaker-1: Hey everyone, welcome to One Thing That Work. Today my guest is Alex Andora, who is a host of Learning Bayesian Statistics Podcast. It's the only podcast about the topic in existence. So Alex, welcome to the show.

speaker-0: Yes, thank you, Sergei. well well done on pronouncing my name and the name of the podcast at the same like in the same sentence. I know it's hard. yeah, so thanks a lot. As you were saying, one of my main things is have these podcasts since twenty nineteen now called Learning Patient Statistics. it's not a very original name, but when I created it actually I didn't know anything about patient statistics, so I was reorienting my career at the time I studied management and political science. So nothing to do with what I do today. Now I'm a I'm a senior scientist, a research scientist. And so I do very mathy statistical stuff. and I do a lot of open source development. But at the time I didn't at all. And so that's how I started the show. I was I learned a lot through podcasts. I listened to a lot of those. So congrats by the way on doing so many episodes. Like I've I've looked on you know, I've stalked you, of course, before the show. That's the great thing of being a podcaster. You get to stalk people professionally. and and so yeah, I was super impressed by the number of episodes you have. Like it's I know it's a h it's hard work, so really, really well done. So yeah, I'm very happy to to be no one of your one of your guests and yeah basically the show I started I started it because I learned a lot through podcasts and at the time there were no podcasts about Bayesian stats. Mm-hmm. And and I was thinking, man, that'd be cool to have a podcast like that, you know, regularly so that basically I can get to hear from the best in the field. And and then I was like, huh. Actually why don't I try to do that? at the time I was a really clueless professional out of out of university. I didn't like my job, and so I was like, I have nothing to offer to people I reach out to. But I was like, But if I reach out to them and I'm like, Hey, you wanna come talk for one hour and we'll record it and then afterwards people will listen to it? maybe they will accept. And that's what happened basically. so that was very, very selfish at the beginning because it was like I'm gonna get to basically interview one of the best, statistics professors in the world, somebody who is at the top of his field or her field, and and then I get to suck in the knowledge for one hour and and then share that with the world. So yeah, at the beginning it was a very selfish enterprise, let's say. but now I've been but Yeah, it turns out there is a lot of nerds out there like me and so they listen to the show, which is awesome. and I I I love them. I have a great audience who is usually very technical and and and loves to get into the details and being told that they are wrong, which is the best audience you can ask for. and so yeah, now the the the show is actually among the The one point five percent most popular shows out of almost four million podcasts globally. so yeah, like I I could not have guessed that at all. And interestingly, I I didn't create the show actually for it to be listened to, but I created it for me, basically. I created the show I needed. And yeah, like very happy I did because it it took me on a wild adventure.

speaker-1: Well listen, congrats on on first of all, one hundred and sixty-one episodes. So I o I'm only at eighty, so you got twice as many. So this is really, really cool. Let me proceed to asking you my go to question for this show. What would you say is then one thing, one decision or habit? or strategy that boosted this show, this podcast the most. So what the worked the what that worked the most.

speaker-0: Mm-hmm. Yeah. Yeah, I think actually there are two things, but I'll start I'll start with one and then you can ask me for a second thing if you want. But so first, before answering, I should probably define a bit what Bayesian statistics is, you know, because otherwise like people will be like, what the hell is that? so there are like there are mainly two ways of doing statistics, if you want, in the world. There is the classic way you're taught at university, which is called frequentist, and there is another Way of doing it, both have a different way of coming at probabilities basically differently. So they define probability differently, and that means the way you then do statistics is different. But interestingly, in the long run, that means as you get more data, both kind of statistics converge towards the same result. So mainly it it it matters when you're in a low data regime. Which is actually most of the case the case most of the time when you work on that. And so why is it called Bayesian? It's after a Reverend, Reverend Thomas Bayes, who in I think it's the seventeenth century if I remember correctly, basically discovered that formula, which now has his name, Bayes formula, which is basically the way to optimally update your belief. Mathematically based on data you observe. Okay, so if for instance your prior belief about me, Sergei, is that I'm actually an alien, you have a prior belief on that, which is a prior probability, okay, which I hope is very low. But then you observe something. You know, I don't know, you secretly come and draw my blood at night, and then you see that my blood is actually blue. So you observe data, you just observe data. Now, how should you update your prior belief about me being an alien based on my blood being blue. You should update it aggressively. Like before your probability should have been very low, but now you should update it very aggressively. But by how much? That's exactly what base formula is gonna will will tell you. Basically it's a very simple formula actually you can Google it and based on what you believed before observing anything and what you just observed. It gives you the posterior probability that time an alien. That's called a posterior probability because it's after observing data. So basically that's that. the that's the way of doing statistics in the Bayesian framework. It's a great framework to use when you don't have a lot of data, when you have some prior knowledge about the phenomenon you're observing, and when you need uncertainty estimation. This is extremely useful. So that's why it's used a lot in medical medical settings it's used a lot by insurances, sports betting, sports analytics, electronic forecasting, lots of lots of things. In marketing it's used a lot also. So yeah, like you'll you'll see that a lot. Basically that's patient sense. Okay. So that's what that's what I do. That's what I talk about on the show with experts. Now, what was one of the main things that worked, I think, for the show being where it is right now? A very boring thing, honestly, I think, is just regularity and frequency. just since September 2019, whatever happened in my life, and a lot of things happened, good and bad, I had an episode released every two weeks. clockwork. And yeah, like basically built a habit not only on my end, but I think also on the listeners, where it's like, you know, the show becomes a part of their routine. I know lots of listeners listen to it when they work out, when they run or when they go for a walk or when they walk their dog or where they do the dishes, stuff like that. And basically, yeah, having that regularly come into their inbox basically made the show part of their routine. So in a way a very boring thing, but yeah, like being persistent often in life is is gonna help you a lot, and here that's the case.

speaker-1: Do you know on on average how many people listen to you monthly?

speaker-0: Yeah, it's and it varies, of course, it depends on the episodes and how you have on there and it's also seasonal. But most of the time it's a few do you s tens of thousands.

speaker-1: wow. wow.

speaker-0: Yeah, I told you there is a lot of nerds in the in the in the world.

speaker-1: This is what I this was my initial thought that you actually didn't go with a broad topic. You went with a really, really niche thing. And you know for a lot of people it's so difficult to choose about among all the options like what you're gonna focus on. So I think your other secret power is like a focus. Focus on one thing. wow. Interesting then.

speaker-0: Yeah, the fact that it's a niche for sure.

speaker-1: Would you say that your network is also something that you wanna thank or network appeared as a result of the podcast more?

speaker-0: Yeah, it's much more like that. When I started the podcast I didn't know anybody, nobody knew me. and actually the podcast gave me back so much because now I know so many people thanks to the show because the show is well known in the field. And and so yeah, like now so it's funny because at the beginning people were asking me, But how are you gonna maintain that? You know, like how are you gonna find topics and guests and so on? And it's literally never been a problem at all. first because

speaker-1: You find guests, these scientists it's not like they're they grow on trees, right?

speaker-0: Well, they kind of do. Like there is a lot of amazing people doing amazing scientific things in the world. It's just like we don't really hear a lot about them. so at the beginning it was a lot just me being curious and I'm just a naturally curious person who loves to learn. So anything I want to learn about, I Google people who work on that and are very renowned for what they do in that field. And then I contact them and see if they want to come to the show. And that's how I mainly did things at the beginning, when nobody knew but neither me or the show. And now it's much easier because I still reach out to to people to come on the show when I see people doing interesting things that I'm curious about. But also I get a lot of people recommending me to come on the show. So that I do a lot. Like when I get a recommendation from someone former guest or people who listen to the show that I know, they are here often, I know about them and they recommend me someone, then I'll check it out and if I like it I'll I invite him. And of course I get a lot of people trying to get on the show by themselves, but that I almost never almost never do it.

speaker-1: Never do. Wow.

speaker-0: No, unless I know them already. But otherwise th it's it's just like it's not a principle, but it's just like it almost never is a good fit when somebody reach reaches out to me. It's it's very funny. Like sometimes it's just it's incredible. Like the show is about patient statistics and I know and I I don't know, like you've got someone who wanna come on the show and talk about business strategies or stuff like that. Yeah.

speaker-1: Okay. They're probably driven by your metrics. They see that you have a coverage and they want to get on, but they're not scientists. and speaking of metrics and speaking of sort of making money from your show or at least covering some expenses, is am I correct that you do have some sponsors?

speaker-0: so I used to and now I don't. because yeah, like I the the post podcast's not really for me a source of income. I I've never seen it that way. I really do it because it's a way for me to stay on top of my field, be curious, learn from great people, and and you know, just meet people. So now mainly there are some expenses because I need to, pay for the podcast hosts and the website and the web developer and the people who handle the s social media and stuff like that. But that I cover mainly through the Patreon of the show. So thanks to the patrons, basically it just like more or less covers the expenses of the show and and that's all. so what I get from from the show, I get a lot from the show. a lot of value from the show, I just don't get lots of money value out of it, but it's mainly my choice. Like I'm sure I'm sure there would be more ways to make money out of the show, but I don't because I also don't wanna mix work in in hobby. So that's more mostly a a choice of mine.

speaker-1: Yeah, is there a point of growth development of this show where you might consider putting it more on the f on the fun financial rails and considering it as a primary earnings driver?

speaker-0: Yeah, maybe. at some point down the road. We'll see. the thing is that I think for that I'll need to expand the audience. And so for that I'll probably need to either be I I probably will need to be less niche. and I'm not sure I I mean don't wanna do that right now and I'm not sure I will ever want to do that. Because what's like the fun part of the show is precisely that it's a niche show and I don't I don't go into, you know, like broad Topics that I don't necessarily care about. That's one thing. And second thing, as you attract more people, you attract more detrimental people. My audience is great and and yeah, like I enjoy interacting with them. Whereas If you grow you necessarily interact with people you don't wanna interact with and so yeah, it's not something I want in my life right now.

speaker-1: tell you the one lesson that I've extracted from our conversation and and I will let you finish finish up strong by asking the main lesson for you from your podcasting journey. What I've heard is that the key to your success is not thinking about others but but thinking about the right others and thinking about yourself about thinking about your growth thinking about as you said saying on top of your field And can contributing to people who you really care and who you wanna you know bring value to in terms of your scientific community. So I think for all the podcasters who want to get to thousands of listens and views and downloads, I I think the lesson here is to stay patient and not think about so much you know other people, but maybe just grow your own game in in whatever field. a better specific field of your of your choice. This is my takeaway. And what was what will be your main lesson? To to share practical

speaker-0: Sharon of

speaker-1: lesson, something that people can can do. Yeah.

speaker-0: Yeah. Yeah, of course it's it's gonna depend on people's objective and what they wanna do with, let's say if they want to create a podcast, but yeah you know it's mainly in the in like in the book world, you know, in the author world, usually authors say that they they write the book they needed to read. and so for me it's the podcast was exactly that. Like create I created the podcast I wanted to listen to. something that's gonna be very helpful in these endeavors is always to think about okay, how can I help myself basically and help people and your people. So probably if you have a problem, a lot of other people have the same problem. And so if just think about a way to solve your own problem.

speaker-1: You're helping yourself and you're helping others in the same breath. Exactly.

speaker-0: Yeah, my problem was that I don't know how to learn Bayesian statistics and I need to do that right now in a medium that I like and trust, like the audio medium. So let's try a podcast. And that's just that solved my problem, but I'm obviously not a special snowflake, so it solved the the problem the same problem that a lot of similar people were having. So that that was that's one of the main things Think about the avatar basically of your show, right? who do you want to help? Yeah. What what does that person look like? What do they do? What do they need? Why would they even listen to that show? That's the main one of the main things. And then the other one that's extremely important is just being able to try stuff out and, fail. this has been one of the main most important. The other most important thing for me has been related to at the same time I was doing the podcast, I learned how to code. And Now, unfortunately with AI, it's a bit blurred, but at the time you need to learn how to program. And learn how to program is awesome because you're gonna fail all the time. your program and your model are gonna fail all the time until you get to a point where it's like, okay, that's good enough. And so learning how to program was amazing for me because it it taught me that failure is fine and actually it's awesome because it gives you quick feedback. And having these quick feedback loop on how to improve your program and your model. And then get to a point where it's good enough and you don't need it to be excellent and perfect. You d need it to be good enough. That was amazing because it it it drilled into my mindset that, well, you need feedback to interact with the world. And that's how you're gonna grow and make this better. And so yeah, that's also how I made the show better thanks to that, because I tried things and if it didn't work, I'm like I didn't let that being personal, but I wanted to take that as feedback. So yeah, that. And and also the third lesson is honestly I do the show again that I like. It's just like sometimes people are like, why don't you talk about that or this? I'm like, yeah, I don't care. So and it's my show. So like I I I invite people that work on stuff I care about. So that's also important I think because this is Hard work, this is it takes time, it takes persistence, and if you don't like it, you're not gonna stick around.

speaker-1: Wonderful way of finishing the conversation. Thanks so much, Alex. Thank you.

speaker-0: Yeah, thank you, Sergei. That was a a pleasure to be in.