WEBVTT 00:00:00.000 --> 00:00:07.000 So our CEO and his incident was no one to update a bunch of dependencies before we launch. 00:00:07.000 --> 00:00:16.000 And I can't get the audio to work at all. I first put the app, I restart my phone, and I'd get this audio spinner of gas. 00:00:16.000 --> 00:00:25.000 And that's when I kind of like raised a flag and was like, I don't think that this is a product we should push to market right now. 00:00:25.000 --> 00:00:36.000 Hello and welcome to episode 53 of Metacast Behind the scenes podcast. I'm your host, Arnaf Deca. 00:00:36.000 --> 00:00:39.000 I'm Uliya Bizuos. And I'm Jenny. 00:00:39.000 --> 00:00:53.000 On this podcast we talk about Metacast, our startups journey. On our website, metacast.app, you can find links to download the app. It's currently in open beta, as well as our weekly newsletter and links to our other podcasts that we do. 00:00:53.000 --> 00:01:04.000 Metacast is a new podcasting app. It's in open beta right now. You can do some pretty cool things that you can't do in any other podcasting app right now, such as it's not just audio. 00:01:04.000 --> 00:01:16.000 It lets you read and see the text and use information from inside the text like bookmarks, specific things, or share it with your friends, search through audio, pretty cool things like that. 00:01:16.000 --> 00:01:27.000 With the help of a transcript. Yes. And there's still a lot of things to do. That's why it's in open beta. We'd love for you to check it out. So without further ado, further ado. 00:01:30.000 --> 00:01:34.000 Usually it's a baby's that I do not others. 00:01:34.000 --> 00:01:48.000 Okay, without further ado, so we launched our app on open beta last Friday. So we are recording this on Wednesday, 14th of February, happy Valentine's Day from the past. 00:01:48.000 --> 00:01:57.000 But you're probably listening to it on 21st of February or some day after that in the future. So hello, happy Valentine's Day from the past to you. 00:01:57.000 --> 00:02:14.000 But yeah, we launched the open beta on Friday, 9th of February, and we've been slowly doing some small scale marketing, and we're going to talk about how the launch went so far, what we're finding out and all of the stuff from there. 00:02:14.000 --> 00:02:21.000 Before we go to the content of this episode, can I try something new? So I was listening to Justin Jackson's new podcast this morning. 00:02:21.000 --> 00:02:41.000 They just launched a new podcast called podcast marketing trends. It's Justin and his friend Jeremy ends with also podcasting. And it is an interesting thing. They said, if you go to Apple Podcasts, Spotify, writers review or rate us, take a screenshot, send us an email, and we will give you a free copy of my book, the pragmatic podcast. 00:02:41.000 --> 00:02:49.000 They do something like that as well, but with some PDFs. So yeah, why not? If you want to support us, go ahead and give us rating and also get the free book for doing that. 00:02:49.000 --> 00:03:00.000 And we're asking for an honest review and a rating. It doesn't have to be fake five stars. If you don't like it, tell us. Yeah, yeah, I'm asking for honest five star review. 00:03:00.000 --> 00:03:07.000 And where can they email us, Ilia? Hello, at metacastapotcast.com. Right. Send that over. 00:03:07.000 --> 00:03:22.000 So let's get started with today's topic. So in the last episode, which was 52, we talked about the craziness leading up to how we decided to do the open beta after all, what led to it. 00:03:22.000 --> 00:03:34.000 So if you want a bit of backstory and context, go listen to that episode. Today we're going to talk about how we carried out the launch and what we're seeing since then in the last four or five days. 00:03:34.000 --> 00:03:43.000 So we basically decided we go into open beta sometime end of January about like eight days before we did, I think eight or 10 days before. Yeah. 00:03:43.000 --> 00:03:53.000 And then we set the launch date for February 6th because the episode 52 was coming out on February 7th. We were like, okay, maybe we should tie this together. Let's do it on the Tuesday, February 6th. 00:03:53.000 --> 00:04:00.000 And then we broke the app on February 6th. Yeah, I just want to say we were pretty much really accepted the app stopped playing audio on Android. 00:04:00.000 --> 00:04:08.000 I think on iOS, it was playing very often, but on Android just was working at all, if I remember correctly, because everything is kind of a bit of a blur. 00:04:08.000 --> 00:04:21.000 We had some intermittent audio issues that we were aware of, but we were like, let's just go ahead and launch with this as is get it out there and we'll fast follow fixed these issues. 00:04:21.000 --> 00:04:32.000 So our CTO in his infinite wisdom wanted to update a bunch of dependencies before we launched and I'm the only Android user of the three of us. 00:04:32.000 --> 00:04:45.000 And I was writing in Slack. Y'all it is not working for me. I don't know about you, but I can't get the audio to work at all. I force quit the app. I restart my phone and I get this audio spinner of death. 00:04:45.000 --> 00:04:54.000 And that's when I kind of like raised a flag and was like, I don't think that this is a product we should push to market right now. 00:04:54.000 --> 00:04:59.000 So you pulled an end on court. Do you want to quickly describe what an end on court is? 00:04:59.000 --> 00:05:06.000 Yeah, in the Japanese manufacturing process, which was met by Toyota, DPS, the other production system. 00:05:06.000 --> 00:05:21.000 On the factory line, if a worker sees a defect in any of the parts on assembly, they have this thing. I don't know how exactly it actually looks like in the factory, but it's some kind of like a court that you pull that court and the entire assembly line just stops, just grinds to a halt. 00:05:21.000 --> 00:05:27.000 Then they go fix the problem to prevent that from reoccurring the root cause it. And then let's how they produce defect free cars. 00:05:27.000 --> 00:05:37.000 And the idea there is by giving the workers control over quality and kind of pride over quality and giving them the tool and also blame free culture, I believe. 00:05:37.000 --> 00:05:42.000 So it's like you don't get blamed. Maybe if you should have pulled it that helps them make high quality cars. 00:05:42.000 --> 00:05:47.000 For example, like you don't get blamed if you decided to upgrade everything that day before lunch. 00:05:47.000 --> 00:05:55.000 No, that is different. If somebody pulls the end on court on somebody else's upgrades, that's the end on court thing. 00:05:55.000 --> 00:06:01.000 So by the way, did you say TPS? Is that where TPS reports come from? What is a TPS report? 00:06:01.000 --> 00:06:10.000 Wait, you haven't seen Office space? I haven't. You have to watch that like before this weekend. You will get all the pop culture references and TPS reports. 00:06:10.000 --> 00:06:18.000 Okay, I'll tell my kids that yeah, instead of having a family quality time, I'm going to be understanding the quality system through the Office space. 00:06:18.000 --> 00:06:21.000 Yeah, yeah, it's important for team culture, Ilya. 00:06:21.000 --> 00:06:26.000 Yeah, okay. Yeah, I can make this sacrifice. Okay, but going back to this, I was thinking about it. 00:06:26.000 --> 00:06:35.000 This kind of a thing, upgrade every possible dependency. We would never do this in a place like AWS because the risk is too high. 00:06:35.000 --> 00:06:44.000 There's also a very rigorous release process. So all your tests, especially security related tests and all that penetration testing. 00:06:44.000 --> 00:07:13.000 All of that have been carried out with a set of configuration or piece of software and you're not going to change that last minute before the release versus a place like now, especially for open beta with the cost and benefit ratio being that these upgrades will actually possibly enhance some of the security aspects, which it did like we use Firebase and there were a lot of upgrades in Firebase dependencies that we did, which was good. 00:07:13.000 --> 00:07:34.000 But occasionally we have seen that upgrading things because we rely on open source software does break stuff to I think it's a fair trade off for especially our scale and size right now and the kind of guardrails, which we already have, which is Jenny listening on audio on Android to make sure that it works before we release it to other users. 00:07:34.000 --> 00:07:38.000 It's terrifying that that is the release process. 00:07:38.000 --> 00:07:48.000 If we did not have Jenny who is an Android user, we might not have noticed this. Even though you have an Android phone or not, you don't use it regularly. It's just a testing device. 00:07:48.000 --> 00:08:01.000 That's after Jenny joined us before that. Remember I bought the phone soon because you would test on iOS and I would test on Android. And then now that Jenny is a full time Android user, I like, OK, I don't have to use this. 00:08:01.000 --> 00:08:11.000 Just to note the change that you made was not being picked up in the emulator. Oh, yeah, I was just saying, thankfully I don't have to use Android that piece of junk. 00:08:11.000 --> 00:08:30.000 So getting back to it, it's not just us who have technical problems, right? We just had a pretty major hiccup in our recording software. So if the episode does not transition from the last 20 seconds into what we are about to say very naturally like it does with us most of the time. That's why. 00:08:30.000 --> 00:08:36.000 Yeah, and similar to our app, this podcast might lose a bit of context. 00:08:36.000 --> 00:08:51.000 That's a very inside joke. OK, so long day, Feb 6, we figured out that audio is fairly broken on iOS and completely broken on Android. So what did we do? 00:08:51.000 --> 00:09:05.000 I have to say this. So we were only called the three of us trying to figure something out. And then I think I mean, I was just talking and Jenny was intensely looking at her screen, just doing something like ignoring us all together. 00:09:05.000 --> 00:09:08.000 Multi-tasking, multi-tasking. 00:09:08.000 --> 00:09:15.000 I think Ilya, even I were actually trying to do something in Spanish. We were asking her and she was not responding. 00:09:15.000 --> 00:09:28.000 Yeah, we were just so frustrated and tired, we just distracted ourselves with something else. And Jenny is like, I fixed it. We were like, OK, but obviously, as a man do we attributed this success to our category in the background. 00:09:28.000 --> 00:09:42.000 Actually, that happened twice that day, where we had some really insane problems. We could not figure out what's going on. And you and I start blabbering about something and then Jenny comes back after like 15 minutes of silence and says, I fixed it. 00:09:42.000 --> 00:09:53.000 Well, let's give some context here. I took on a task that refactored and added some things to the audio. And it has been slightly broken ever since. 00:09:53.000 --> 00:10:09.000 So I'm very much taking ownership of trying to improve the audio because I'm not going to say that it's my fault. It just turns out that there are a number of ways these different things can go wrong. And we have to tweak a lot of things. 00:10:09.000 --> 00:10:26.000 So we're slowly inching our way closer to more stable audio experience and to give a little bit more. So before this major change in how audio works in our app, we basically use the third party plugin to do it all. 00:10:26.000 --> 00:10:40.000 But that plugin doesn't allow you to like configure a lot of background audio stuff, which is important for a podcasting app because you want your like car hardware buttons and maybe your Bluetooth headphones, all those like tapping and all that to work. 00:10:40.000 --> 00:10:49.000 And so Jenny re engineered refactored the audio system. And yeah, major change minor problems. 00:10:49.000 --> 00:11:07.000 But the actual root cause of the thing being broken the day we wanted to launch, it turned out was upgrading one of our audio plugins. And realizing that there was actually some hidden changes we needed to make otherwise it was a breaking change. 00:11:07.000 --> 00:11:10.000 So just adding a few lines of code did the trick. 00:11:10.000 --> 00:11:24.000 I mean, it's an open source piece of software. It's a great piece of software. It does so many complex things, but it was possibly a breaking change that requires more configuration when you change that new version that was not in the change list. 00:11:24.000 --> 00:11:30.000 So we didn't think about it. Reading the code we figured out, okay, or Jenny figured out that we have to do this. 00:11:30.000 --> 00:11:47.000 So I'm curious if at Amazon, let's say an engineer broke something the day before launch, the day of launch, how much extra paperwork would have to be done, explain for why it happened, root cause analysis, basically all of those artifacts that you have to create to document what went wrong. 00:11:47.000 --> 00:11:49.000 As opposed to actually spending time fixing it. 00:11:49.000 --> 00:12:01.000 First of all, to be clear, you wouldn't do this kind of a thing at AWS because the stakes are too high. You have customers, companies relying on AWS. So you wouldn't do this kind of change. 00:12:01.000 --> 00:12:06.000 However, there could be operational incidents like you did something else and that broke. 00:12:06.000 --> 00:12:12.000 And typically for doing any change in production, there is a very rigorous change management process. 00:12:12.000 --> 00:12:20.000 So you do have to write up what you're going to change, what you're going to do. If it breaks, how are you going to roll back and everything must have a rollback procedure. 00:12:20.000 --> 00:12:29.000 But even then, sometimes things do happen. And in that way, I really actually like Amazon's correction of error, COE process. 00:12:29.000 --> 00:12:32.000 And Google has probably something similar every company does. 00:12:32.000 --> 00:12:36.000 They call them also another three letter acronym. They call OMG's. 00:12:36.000 --> 00:12:37.000 Oh my God. 00:12:37.000 --> 00:12:42.000 Yeah, nobody ever says, oh my God. Actually, maybe they're actually might stand for something else. I don't know. 00:12:42.000 --> 00:12:44.000 But they called OMG's an assistant. It's called OMG. 00:12:44.000 --> 00:12:48.000 What I like about that process is again, it's blameless. 00:12:48.000 --> 00:12:58.000 It's not about who did what, but more about what allowed like either process or software or something allowed this thing to happen in the first place. 00:12:58.000 --> 00:13:02.000 Rather than somebody doing something that broke something. 00:13:02.000 --> 00:13:07.000 I don't know if the difference between that two kind of makes sense. But let's put it this way. 00:13:07.000 --> 00:13:17.000 A few years back, there was a humongous Amazon S3 outage turns out it was because a single engineer ran one command line thing that did something we want to get into the details. 00:13:17.000 --> 00:13:22.000 If you want, you can look up S3 outage of 2019 or so. Basically, the whole internet was done. 00:13:22.000 --> 00:13:29.000 I think it's public. Yeah, it was a tens of thousands of instances were removed by that single type of in the command. 00:13:29.000 --> 00:13:34.000 And then suddenly the internet was down for like a day or so because of this. 00:13:34.000 --> 00:13:40.000 But that engineer is still at Amazon or it was when I was there and was doing well. 00:13:40.000 --> 00:13:47.000 And it's because the process is not about blames or mistakes. It's about what do we do? 00:13:47.000 --> 00:13:52.000 What do we change in the process so that a single person cannot run this kind of a command ever again? 00:13:52.000 --> 00:13:55.000 Incorporate is important to have visibility. 00:13:55.000 --> 00:14:02.000 And time to get in promoted comes. People should know you. So this person probably has a lot of visibility. 00:14:02.000 --> 00:14:04.000 War stories, yes. 00:14:04.000 --> 00:14:08.000 What are some other ways we broke that because that was not all. 00:14:08.000 --> 00:14:12.000 Yes. Eliya, do you want to talk about this one quickly? Sure. 00:14:12.000 --> 00:14:20.000 So again, the infinite wisdom of our CDO wanted to make sure that we protect our back end from unauthorized access, which is a very commandable goal. 00:14:20.000 --> 00:14:22.000 To enable AppCheck. 00:14:22.000 --> 00:14:26.000 Well, if this is only my infinite wisdom, I would worry. 00:14:26.000 --> 00:14:30.000 No, I was just challenging my inner journey here using her words. 00:14:30.000 --> 00:14:36.000 Yeah. So AppCheck is this actually, I don't know if it's the package or it's more like a system configuration. 00:14:36.000 --> 00:14:47.000 You can almost consider it like a proxy kind of system that protects fire based back end as well as you can use it to protect your third party back ends to from your app or web usage. 00:14:47.000 --> 00:14:53.000 Yeah. So we thought we should enable this before we launch. It is reversible, but it is a brick and change. 00:14:53.000 --> 00:14:59.000 And it might take 15 minutes to see that it broke another 15 minutes to undo it. So it is a downtime. 00:14:59.000 --> 00:15:11.000 And it was reversible for three of the four things we use, but not reversible for one thing, which was probably the most crucial thing for apps is authentication. 00:15:11.000 --> 00:15:15.000 So yeah, we turn it on and guess what? It broke. 00:15:15.000 --> 00:15:23.000 We were considering actually turning it on later on in production, but because we had a bit of extra time, we decided to turn it on before we launch. 00:15:23.000 --> 00:15:25.000 And well, thank God we did. 00:15:25.000 --> 00:15:29.000 I think it took us a couple of hours of fiddling and testing to figure it out. 00:15:29.000 --> 00:15:33.000 Again, our city of figured out very, very quickly within five or ten minutes, I think. 00:15:33.000 --> 00:15:39.000 But it took probably another few hours to get the change into the app, rebuild the app, test it, and all that. 00:15:39.000 --> 00:15:47.000 But yeah, we are sure glad we did it before we launched because it would have broken everybody exactly the same way, but we would have more users. 00:15:47.000 --> 00:15:51.000 Or it would have been a very expensive upgrade process to do this. 00:15:51.000 --> 00:15:53.000 Why would be an expensive process? 00:15:53.000 --> 00:15:55.000 Expensive by time, right? 00:15:55.000 --> 00:16:01.000 Yeah, expensive as in once you release the app, it's public, it's available. 00:16:01.000 --> 00:16:06.000 You wouldn't want to do this kind of a change directly on that app and on that backend. 00:16:06.000 --> 00:16:17.000 So the process would most likely be setting up a different backend, turn it on there, make sure the app can talk to it, and then turn it on in your real backend, the production backend. 00:16:17.000 --> 00:16:22.000 So it would have taken a lot more time and a lot more effort to do this kind of a change later on. 00:16:22.000 --> 00:16:27.000 But setting up this whole development environment and all that that would have taken probably a few days. 00:16:27.000 --> 00:16:31.000 Yeah. Okay, after that, I think it was pretty smooth, right? 00:16:31.000 --> 00:16:42.000 Yeah, after that was just grinding through some of the things we have did it the site, we also enabled the open beta stage in tram line, which is our CIS Deep Pipeline Software, tramline.app. 00:16:42.000 --> 00:16:44.000 I think it's their address. 00:16:44.000 --> 00:16:46.000 So anyway, we will link tram lines from the show notes. 00:16:46.000 --> 00:16:52.000 Tramline is awesome, also done by our friends, well, our nubs friends, now our friends because we love them. 00:16:52.000 --> 00:16:55.000 And we also had an episode with them to talk about their other product. 00:16:55.000 --> 00:17:01.000 We wrote a few updates for Reddit, yeah, it was basically just like grinding through a bunch of these updates. 00:17:01.000 --> 00:17:07.000 And then on Friday, I think we had the links finally up, but that's what I found interesting. 00:17:07.000 --> 00:17:15.000 So we had version 56, which was functional, but it had a viewer bug with a playlist, but we shipped it to open beta anyway. 00:17:15.000 --> 00:17:17.000 And Apple approved it within the same day. 00:17:17.000 --> 00:17:20.000 We did a few hours, I think, like a couple of hours. 00:17:20.000 --> 00:17:25.000 But then version 57, which had the fix for that little bug, it took probably three days. 00:17:25.000 --> 00:17:35.000 I think we submitted it on Friday and then on the weekends nobody was working, I don't know, because sometimes we get reviews on weekends as well, but not this time. 00:17:35.000 --> 00:17:37.000 And then it was only approved on Monday, I think. 00:17:37.000 --> 00:17:42.000 Yeah, there's like some manual review always happening on Apple's idea. 00:17:42.000 --> 00:17:45.000 Yeah, did we even have to wait for Google to review anything? 00:17:45.000 --> 00:17:50.000 No, Google only does configuration level reviews. 00:17:50.000 --> 00:18:05.000 The actual app changes looks like it's automated testing because every time we submitted, we do see Google users that we've now figured out that they're like test users, not humans, bots, signing into the app, doing things. 00:18:05.000 --> 00:18:07.000 And just to make sure that the app is functional. 00:18:07.000 --> 00:18:18.000 Yeah, so the other thing that we found puzzling is that being in beta on Google Play Store basically means that you just show up in the Play Store and anybody can download the app. 00:18:18.000 --> 00:18:21.000 There's a quarter, we said it to 1000 users right now. 00:18:21.000 --> 00:18:26.000 So if more than 1000 users use the app, I don't know if it will not show up or if it will not be able to download it. 00:18:26.000 --> 00:18:31.000 I don't know how it's going to work, but it was surprising that anybody can just go to the app store and just find it. 00:18:31.000 --> 00:18:33.000 But that's how Google Play just looks. 00:18:33.000 --> 00:18:38.000 Like you can search for Metacast or you can search for like podcast player and in that it'll show up. 00:18:38.000 --> 00:18:43.000 And there is a very small disclaimer saying this is in a beta or something like that, right? 00:18:43.000 --> 00:18:48.000 Which is actually we wrote that it was not Google doing anything automatic to show that. 00:18:48.000 --> 00:18:51.000 Yeah, so you put the word beta right in the title. 00:18:51.000 --> 00:18:53.000 Whereas with Apple actually, it's almost the opposite. 00:18:53.000 --> 00:19:01.000 They give you the link that has very clear instructions, but you have to install test flight, which is an app that Apple acquired already about 10 years ago. 00:19:01.000 --> 00:19:08.000 And it used to be a third party app that lets people like us distribute private apps in test mode right before they released the app stores. 00:19:08.000 --> 00:19:13.000 So yeah, people have to first install test flight, then they click the link and then they are able to install. 00:19:13.000 --> 00:19:15.000 And it kind of expires after 90 days. 00:19:15.000 --> 00:19:16.000 So it's very tightly controlled. 00:19:16.000 --> 00:19:19.000 Whereas with Google, it just feels like free for all. 00:19:19.000 --> 00:19:21.000 I don't know which approach I like more. 00:19:21.000 --> 00:19:23.000 Google is definitely easier for the user. 00:19:23.000 --> 00:19:27.000 Apple feels like it just gives a bit more control to us developers. 00:19:27.000 --> 00:19:44.000 And maybe a bit of safety angle too, because Apple's beta software on iOS, you basically cannot install unless you go through the hoops of installing test flight and acknowledging that this is beta software and you're taking on the risk of installing it. 00:19:44.000 --> 00:19:52.000 Whereas Google says almost accidentally you can search for something, find something that's in a beta, not even in production and install it without knowing it. 00:19:52.000 --> 00:19:53.000 Right. 00:19:53.000 --> 00:19:59.000 So Apple also you have to have the link. If you don't have the link or the code for test flight, you can find it, you can discover it. 00:19:59.000 --> 00:20:05.000 Yeah, I think different philosophies to running different sorts of app stores, that's all fair game. 00:20:05.000 --> 00:20:07.000 Alright, let's talk about the marketing site. 00:20:07.000 --> 00:20:10.000 Like what did we do after the soft launch? 00:20:10.000 --> 00:20:12.000 Yeah, I think it was Friday afternoon. 00:20:12.000 --> 00:20:20.000 I wrote a short post about how crazy this week was and congratulating our thousand logic metacast to open beta with all the links. 00:20:20.000 --> 00:20:26.000 It got 6.7000 impressions, which is pretty good. You can measure the clicks in LinkedIn. 00:20:26.000 --> 00:20:29.000 Yeah, it sent out a newsletter on Monday. 00:20:29.000 --> 00:20:33.000 So yeah, that wasn't Friday, right? The LinkedIn post on Monday sent out a newsletter. 00:20:33.000 --> 00:20:40.000 We have just under 400 subscribers, generally about 50% of those open our newsletter. 00:20:40.000 --> 00:20:43.000 So give or take 200 people would have seen the email. 00:20:43.000 --> 00:20:48.000 Yeah, but then unexpectedly we also got mentioned on a newsletter called PadNews. 00:20:48.000 --> 00:20:59.000 And we learned about it because somebody wrote a huge comment under our read post saying that they've downloaded the app and they had a couple of questions and reports to make about the app. 00:20:59.000 --> 00:21:02.000 The person mentioned that they learned about us on the newsletter. 00:21:02.000 --> 00:21:04.000 So we asked them for a link and then we found the app. 00:21:04.000 --> 00:21:10.000 There's a small mention, but apparently you know people read that and they found us this way, which was pretty cool. 00:21:10.000 --> 00:21:17.000 Yeah, so far basically I think every day I've been posting something and mentioning metacast in my LinkedIn post. 00:21:17.000 --> 00:21:26.000 So that gets a bit of traffic and we still have 300 people who enter their emails in our landing page that we have not reached out to. 00:21:26.000 --> 00:21:30.000 I do want to get the audio in a better shape before we start reaching out to those people. 00:21:30.000 --> 00:21:37.000 To be fair, I think of those 300, there's probably some that already are on the beta, but there's still a good amount of people. 00:21:37.000 --> 00:21:40.000 And then our personal networks and all that will reach... 00:21:40.000 --> 00:21:43.000 Yes, yeah. And then I have a rant about this. 00:21:43.000 --> 00:21:52.000 I recently came across a screenshot or a short video where people are talking about this tool called Tapleo that helps you automate your social media engagement. 00:21:52.000 --> 00:21:56.000 And you can like DM multiple people with a predefined message. 00:21:56.000 --> 00:22:02.000 All that kind of stuff. And I was looking at this and I'm like, okay, so this is the emails spam but for social. 00:22:02.000 --> 00:22:17.000 I think you mentioned it on LinkedIn itself, but there are LinkedIn browser extensions that will basically increase your engagement by replying to whatever post you're seeing and using AI to generate a comment. 00:22:17.000 --> 00:22:21.000 And those comments are usually very easy to tell apart that these are AI generated comments. 00:22:21.000 --> 00:22:27.000 Yeah, they would be something like, yeah, generated comments are very easy to tell in the Raspersial tools to detect them. 00:22:27.000 --> 00:22:31.000 What other means do you think we could use to detect AI generated comments? 00:22:31.000 --> 00:22:34.000 That's exactly the structure that they use. They're purpose what you said and ask a question. 00:22:34.000 --> 00:22:39.000 Yeah, it's a hook. It's just so lame because now everybody knows they're AI generated. 00:22:39.000 --> 00:22:47.000 I was thinking maybe we should use a tool that would automatically send those people an email, something like, no, that was Tapleo, but for email. 00:22:47.000 --> 00:22:52.000 But like, there's something wrong about it. I think the energy of using an automated tool to send 300 emails. 00:22:52.000 --> 00:22:59.000 I don't like it. So I'll probably start with maybe like top 50. I'll just say take 50 email addresses and actually manually send in the emails. 00:22:59.000 --> 00:23:03.000 And see what the response rate of that will be. 00:23:03.000 --> 00:23:15.000 I think it's this scale I can afford to spend an hour and a half to just send everyone an email. I mean, it will be the same copy-pasted text, but I could probably deduce the names from the email addresses and just send it from my account. 00:23:15.000 --> 00:23:20.000 We see how that works. There's another thing that is not a rant. We were looking at this software called Fondo for taxes. 00:23:20.000 --> 00:23:27.000 I think I entered my email there to get a discount or something. And I got an email supposedly from their CEO. So it's named by the CEO's name. 00:23:27.000 --> 00:23:33.000 And that email looks like it's personal email and it ends with something like, let me know if blah, blah, blah, right? 00:23:33.000 --> 00:23:38.000 And I just ask him a question in response to that email. Well, it's been about a week. I haven't heard it back yet. 00:23:38.000 --> 00:23:44.000 And that's what I don't like about this kind of personalized emails that make you believe they are personal but they're not personal. 00:23:44.000 --> 00:23:49.000 That's why I have this hesitation of using an automation tool for just a couple of hundred of emails. 00:23:49.000 --> 00:23:53.000 For our scale, yeah. Personal purchase is probably for our scale. Yeah. 00:23:53.000 --> 00:23:58.000 Obviously, if you have like 5,000 or more, like you can't possibly do this with a couple of hundred, you can. 00:23:58.000 --> 00:24:04.000 But at that point, like, would you want to make that email feel like a personal email? That's another choice to make. 00:24:04.000 --> 00:24:10.000 That's another thing. Right. You can make it impersonal because it is impersonal. But if you try to fake it being personal, that's what I don't like. 00:24:10.000 --> 00:24:17.000 Yeah. Okay. So it's been like a handful of days since we launched it. Let's talk about a few numbers. 00:24:17.000 --> 00:24:25.000 I'm going to check out prep more numbers for this recording. So just yesterday, often when I started working on a dashboard, I mean, we had a dashboard. 00:24:25.000 --> 00:24:31.000 I guess maybe let's talk about the stack there because we use Firebase and we use Google Analytics for events. 00:24:31.000 --> 00:24:38.000 Actually, both of these have very nice integration with BigQuery. We have a feed of all analytics events in the BigQuery. 00:24:38.000 --> 00:24:42.000 They all anonymize and all. And then in BigQuery, we can make reports on that data. 00:24:42.000 --> 00:24:47.000 And then we try to use local studio, the free version of it for visualization of that data. 00:24:47.000 --> 00:24:53.000 When I do something simple, it works. Something simple means like a simple chart or a simple table. 00:24:53.000 --> 00:25:04.000 But I was trying to do something more sophisticated. And I didn't like that basically made me conform. The way I like to do reports, I like to have a view that has a bunch of data in there. 00:25:04.000 --> 00:25:09.000 And then I can just like slice and dice it in different ways. It's almost like I get a lot of data into the same table. 00:25:09.000 --> 00:25:12.000 I use something like PivotTable functionality to get different insights from it. 00:25:12.000 --> 00:25:18.000 Local does not work that way, unfortunately. So I would have to write multiple almost like purpose-built queries to get the data that I needed. 00:25:18.000 --> 00:25:26.000 And I discovered that in Google Sheets, you can actually connect a BigQuery source. And then in Google Sheets, you can use PivotTables. 00:25:26.000 --> 00:25:33.000 And then with PivotTables, you can do charts from them. And I started doing what I really enjoyed doing in both Amazon and Google. 00:25:33.000 --> 00:25:44.000 I set up a bunch of Sheets that have different representations of that data in PivotTables. And they have one sort of huge dashboard sheet that just pulls the data from all of those other sheets. 00:25:44.000 --> 00:25:53.000 So I can have all of the numbers in the same view in a very nice spreadsheet format that can be easily formatted and moved around. And you can build charts from it. 00:25:53.000 --> 00:25:56.000 So yeah, I started working on that, but I haven't finished. 00:25:56.000 --> 00:26:11.000 But what we are seeing now, after launching the data, we had about 20 users on average or so using the app. And coming back to the thing that you said before our nav, every time we submit to Google, Google Play Store, we have these fake users signing up to the app, creating some activity in app. 00:26:11.000 --> 00:26:20.000 So it's some kind of testing. If you look at Firebase reporting, it would show you we have something like 170 users or something. It's a very big number compared to what we actually have. 00:26:20.000 --> 00:26:24.000 I think more like 250 or so. That's what it shows. 00:26:24.000 --> 00:26:33.000 Whereas what I think we should be measuring is more like active listeners. So people who started playing an episode, but can also play an episode. 00:26:33.000 --> 00:26:44.000 So I guess it's going to be an error built in there. So yeah, I looked at a number of episodes, people play and also people who play at least one episode. So an active user is defined as a person who played at least one episode during the day. 00:26:44.000 --> 00:26:49.000 And also we look at a completion of episodes, which indicates that people actually stick around and use that. 00:26:49.000 --> 00:27:02.000 So I haven't built anything more sophisticated yet. But yeah, what we saw is about 20 users, about 40 episodes that they start, about like 25, 30 that they finish, which is much more than we used to have before we launched the open beta. 00:27:02.000 --> 00:27:11.000 I'm looking forward to actually building out the dashboard in full. And by the time this episode comes out, I think we should include a couple of graphs into the newsletter so that I can share those numbers. 00:27:11.000 --> 00:27:22.000 And maybe the next episode we could actually go over a little bit more because there's a lot of metrics like what are we trying to measure? Why are those things important to us and all that? 00:27:22.000 --> 00:27:24.000 I think that would be an interesting part. 00:27:24.000 --> 00:27:34.000 Yeah, also another meta point here is when we first set up the integration, the BigQuery, I built a few dashboards, which sort of made sense at the time during the close beta time. 00:27:34.000 --> 00:27:44.000 But they were also hypothetical because we barely had any usage. And now that we have actual usage, I'm starting to rethink what kind of metrics you want to look at, what kind of shape the reports should be in. 00:27:44.000 --> 00:27:50.000 Whereas if you do it theoretically, it's almost guaranteed that you will have to redo what you've done. 00:27:50.000 --> 00:28:00.000 I think it's a good insight for the future that it's best to think about what events you want, instrument them, and then actually have the data and then use the data to shape it into the format you want. 00:28:00.000 --> 00:28:04.000 As opposed to trying to predict everything in advance. 00:28:04.000 --> 00:28:11.000 The other, I think the Reddit people joining the subreddit and engaging that has also been pleasantly surprising. 00:28:11.000 --> 00:28:20.000 It's nothing astronomical, which we never expected anyway. But it's been pretty cool, right, to see about 30 people also so far have joined. 00:28:20.000 --> 00:28:23.000 There's some chatter in the Reddit posts and all that. 00:28:23.000 --> 00:28:34.000 Yeah, we have 32 users last time I checked, but they said it was an 11,000 percent growth from the three members we had. 00:28:34.000 --> 00:28:39.000 Maybe that's how we should start our LinkedIn post like we grew our community 11,000 percent. 00:28:39.000 --> 00:28:50.000 Yeah, one more quick step I was just looking at analytics is looks like the Android versus iOS playt is almost even at 50, 50 now. 00:28:50.000 --> 00:28:56.000 Although some of that is probably contributed by the Google automatic testers, which don't seem to happen on iOS. 00:28:56.000 --> 00:29:01.000 That's cool because we had very few one or two only Android users still now. 00:29:01.000 --> 00:29:04.000 So what is some feedback that you learned so far? 00:29:04.000 --> 00:29:09.000 One thing is feedback also has been coming in from many different channels. 00:29:09.000 --> 00:29:16.000 We've got emails, we've got Reddit, you have received telegram, I received WhatsApp feedback over emails. 00:29:16.000 --> 00:29:21.000 So they're like coming from all sorts of angles, which is cool. It's great to see all that. 00:29:21.000 --> 00:29:24.000 Yes, people sending us DMs, people who know us. 00:29:24.000 --> 00:29:28.000 Yeah, that's going to both blessing in the course. It's great to have all those. 00:29:28.000 --> 00:29:32.000 It's just difficult to share them with the team because you can't just send them a link. 00:29:32.000 --> 00:29:39.000 In terms of actual feedback, I think somebody mentioned to you that the phone was heating up and then another person also said, 00:29:39.000 --> 00:29:44.000 and we had never heard or experienced this ourselves. So that was brand new. 00:29:44.000 --> 00:29:51.000 Yeah, today it was heating up as well for me, but I was listening to the podcast while also being in LinkedIn. 00:29:51.000 --> 00:29:58.000 So I actually don't know which app caused the overheating, but when I looked at the screen and put it down, it became cool again. 00:29:58.000 --> 00:30:05.000 Yeah, this will be interesting things for us to also diagnose because none of us have a background in app development. 00:30:05.000 --> 00:30:10.000 This is the first time we're doing something like this. So I'm super excited to tackle these kind of things. 00:30:10.000 --> 00:30:15.000 The other thing I kind of felt is even on Monday, we were a bit hesitant. 00:30:15.000 --> 00:30:22.000 Like, should we send out this to the newsletter or not because there were still intermittent issues with the audio, right? 00:30:22.000 --> 00:30:31.000 But I think looking at some of the analytics data, at least feels like there is a good number of people actually starting or that had used it already by that day, 00:30:31.000 --> 00:30:38.000 had played episodes, completed episodes and all that. So we decided to ultimately just send the newsletter anyway. 00:30:38.000 --> 00:30:43.000 So now that we've launched, I really like how Amazon was putting this. 00:30:43.000 --> 00:30:50.000 Launch is just day one. Launch is just the beginning, even though you might have put months or even years into building the thing, 00:30:50.000 --> 00:30:57.000 it only starts when you actually launch. And that's how I feel now, because yeah, all of the effort that we've put in is all great, but now is the moment. 00:30:57.000 --> 00:31:03.000 Not necessarily moment of truth, now is the moment when it all starts in terms of like, it meets the reality. 00:31:03.000 --> 00:31:10.000 Is what we built what people actually want. One more realization I had after launching. 00:31:10.000 --> 00:31:18.000 So we have been working on this from about what March, I think March, that was the first commit for the app March last year. 00:31:18.000 --> 00:31:21.000 End of March or so. Do you remember, Ilya? 00:31:21.000 --> 00:31:25.000 I remember that in March, we had a demo that we used for the white combinator. 00:31:25.000 --> 00:31:36.000 Right, or maybe it was February then February and when we started probably it's almost a year. But throughout this thing, very early on, it was completely bare bones, right? 00:31:36.000 --> 00:31:41.000 Like you couldn't browse different podcasts, you couldn't find anything. It was basically what we had put in there. 00:31:41.000 --> 00:31:49.000 The interface itself was quite janky. There was a lot of gimmicky hard-coded stuff just for the demo, I think. 00:31:49.000 --> 00:31:56.000 But over time, until now, I had never thought about how does it feel in the user's hands. 00:31:56.000 --> 00:32:02.000 And two things change. One is the marketing, the Metacast.app website, the image that you put in there. 00:32:02.000 --> 00:32:14.000 I am H.O. in my humble opinion, it looks great. But also like two or three people have told us that they really like the look and feel and how the information flows and all that. 00:32:14.000 --> 00:32:25.000 And that kind of finally made me realize that this is actually a real app. Now it's not a personal experiment or hobby thing anymore. 00:32:25.000 --> 00:32:30.000 Yeah, I've personally become hypersensitive to little bugs now even more. 00:32:30.000 --> 00:32:34.000 Because previously we'd be like, whatever we haven't launched yet, we can leave with that. 00:32:34.000 --> 00:32:40.000 But now every time something happens and a bunch of things happen now because it's fairly unstable right now. 00:32:40.000 --> 00:32:46.000 I always have this at the back of my mind. Would this be a trigger for somebody to say, okay, I'm done with this. 00:32:46.000 --> 00:32:50.000 And then I'm going to move to something else. Just because people have different tolerances for bugs and all. 00:32:50.000 --> 00:32:55.000 Of course it would be, right? We are kind of personally invested in it, but most people are not. 00:32:55.000 --> 00:33:04.000 So yeah, basically for us right now, these stability issues are the number one priority because audio fails every now and then. 00:33:04.000 --> 00:33:10.000 Then the app loses state every now and then which can be really annoying actually today was searching something. 00:33:10.000 --> 00:33:15.000 And it lost state basically it went from me typing something in search back to home screen. 00:33:15.000 --> 00:33:18.000 So I lost all of that progress of typing. It was very annoying. 00:33:18.000 --> 00:33:26.000 These little things they happen every now and then. And this is just core functionality. They have to be fixed before we go on to like build cooler stuff. 00:33:26.000 --> 00:33:35.000 Right. And I think some of these things are very hard to debug because they are intermittent. We can't really see them on our devices or in emulator. 00:33:35.000 --> 00:33:40.000 So that's also the next setup things we're doing is to add more diagnostics features to the app. 00:33:40.000 --> 00:33:44.000 Yes, one of the things that we'll be doing actually, I don't have you already built it or not. 00:33:44.000 --> 00:33:47.000 It's kind of semi built about 60%. 00:33:47.000 --> 00:33:50.000 Yeah, basically we'll have a feature to enable logging. 00:33:50.000 --> 00:33:56.000 The user can choose to enable logging and then send us those logs. Obviously we're not forcing this onto everyone. 00:33:56.000 --> 00:33:59.000 So it's only if you want to help us troubleshoot something you can opt into this. 00:33:59.000 --> 00:34:05.000 That's something that we are learning with the mobile device. It's a computer in somebody else's hand that the app runs on. 00:34:05.000 --> 00:34:10.000 Especially in Android, there are so many different devices and it's just almost impossible to get hold of the logs. 00:34:10.000 --> 00:34:12.000 So we need to have some tools in place. 00:34:12.000 --> 00:34:21.000 And Jenny has fixed a lot of like the current problems already in a PR so that should go out very soon. 00:34:21.000 --> 00:34:25.000 So yeah, let's hear from Jenny because Jenny, you've been quite quiet today. 00:34:25.000 --> 00:34:28.000 We are going to our favorite section of this podcast. 00:34:28.000 --> 00:34:31.000 Maybe we should just have a podcast, book recommendations. 00:34:31.000 --> 00:34:34.000 That's not what it's going to listen to that. 00:34:34.000 --> 00:34:38.000 So yeah, Jenny, what are you up to now in terms of content consumption? 00:34:38.000 --> 00:34:44.000 I am listening to the Elon Musk book on audio book from the library. 00:34:44.000 --> 00:34:46.000 It's about 20 hours or so. 00:34:46.000 --> 00:34:49.000 I'm listening to it at 1.7 speed. 00:34:49.000 --> 00:34:52.000 And I have three weeks to finish it. 00:34:52.000 --> 00:35:02.000 So I've told my company the two guys I work with that I'm actually not using the medicast app very much right now because I got a crank through this book. 00:35:02.000 --> 00:35:10.000 So I'm listening to that and then my daughter has a little bit of an entrepreneurial streak and she wanted to start a banana bread business. 00:35:10.000 --> 00:35:13.000 So last night we watched an episode of Shark Tank. 00:35:13.000 --> 00:35:17.000 I've never seen it before, but it gives her a little bit and me. 00:35:17.000 --> 00:35:25.000 All of us, my family, a little bit of insight into company building, funding, different things that investors are looking for. 00:35:25.000 --> 00:35:27.000 I don't know, it's kind of interesting. 00:35:27.000 --> 00:35:34.000 Yeah, it's a very different perspective than I think even what you'll normally find in investment circles. 00:35:34.000 --> 00:35:38.000 I mean, literally it's called Shark Tank because of a reason, right? 00:35:38.000 --> 00:35:44.000 In terms of podcasts, so Vision Pro launched I think about a week ago now as of this recording. 00:35:44.000 --> 00:35:47.000 So there's been a bunch of podcasts. 00:35:47.000 --> 00:35:49.000 The HardForg podcast has had it. 00:35:49.000 --> 00:35:51.000 Brand McAle's tech meme has covered it. 00:35:51.000 --> 00:35:55.000 So I've been going through all that trying to figure out so far it sounds intriguing. 00:35:55.000 --> 00:36:05.000 If I had like $3,500 plus tax to like spare on a version one, probably half not working device, I would. 00:36:05.000 --> 00:36:08.000 But yeah, it sounds so far pretty interesting. 00:36:08.000 --> 00:36:12.000 You see the Zuckerberg video where he talks about Quest vs Vision Pro. 00:36:12.000 --> 00:36:18.000 Not yet. I just saw a mention of it this morning. I have not seen it yet. 00:36:18.000 --> 00:36:22.000 Yeah, I saw it enough time in my feed that I decided to watch it today. 00:36:22.000 --> 00:36:24.000 It's about three minutes. It's actually quite interesting. 00:36:24.000 --> 00:36:28.000 I mean, obviously he says that Quest is better than Vision Pro. 00:36:28.000 --> 00:36:32.000 But he said a quote. Sounded like he made it up, but apparently somebody else said this before. 00:36:32.000 --> 00:36:35.000 The best way to predict future is to invent it. 00:36:35.000 --> 00:36:38.000 I love that. That's why we are on a quest to build Metacast. 00:36:38.000 --> 00:36:41.000 Our vision is very pro. 00:36:41.000 --> 00:36:44.000 Oh my god, yeah. That's an Apple Lava code. 00:36:44.000 --> 00:36:50.000 And then the other thing I've been reading is an audiobook called The Midnight Library. 00:36:50.000 --> 00:36:54.000 It's incredible. I don't usually read this sort of stuff. 00:36:54.000 --> 00:36:57.000 It's more of, I think, philosophy and things like that. 00:36:57.000 --> 00:37:05.000 If you have a bit of, like, science fiction plus philosophy nerdism in you, I think you would really enjoy this book. 00:37:05.000 --> 00:37:09.000 It's about the state between life and death. 00:37:09.000 --> 00:37:19.000 And it's portrayed as if there is a place, a midnight library where you go to, and basically every book in there is a decision in your life that you would have taken differently. 00:37:19.000 --> 00:37:22.000 And how your life would have turned out differently. 00:37:22.000 --> 00:37:23.000 It's amazing. 00:37:23.000 --> 00:37:25.000 Nice. Who is over? 00:37:25.000 --> 00:37:27.000 Matt Hague, I think. I'll look it up. 00:37:27.000 --> 00:37:28.000 Is it modern? 00:37:28.000 --> 00:37:30.000 Yeah. It's like contemporary. 00:37:30.000 --> 00:37:33.000 Maybe 15, 20 years back, that's about it. 00:37:33.000 --> 00:37:34.000 Yeah. Matt Hague, yes. 00:37:34.000 --> 00:37:38.000 It's incredible book. And it's got really good reviews too. 00:37:38.000 --> 00:37:42.000 Cool. Yeah. So I've not been reading much recently. 00:37:42.000 --> 00:37:47.000 I'm still trying to finish the Children of Dune book, which is the third book in the series. 00:37:47.000 --> 00:37:53.000 I had to, like, look into the plot because it's so interesting, but I don't have enough time to read it. 00:37:53.000 --> 00:37:58.000 Yeah, apparently one of the people that returned into the sandworm at the end. 00:37:58.000 --> 00:38:04.000 So I'm actually looking forward to reading this because that takes this science fiction slash fantasy to the next level. 00:38:04.000 --> 00:38:08.000 Also, I had one podcast discovery this week. 00:38:08.000 --> 00:38:13.000 There's a podcast called Acquired. It's very popular. I heard about it, but I never listened to it. 00:38:13.000 --> 00:38:19.000 But their host David Rosenthal was on Justin Jackson's Build Your Sass podcast. 00:38:19.000 --> 00:38:21.000 That's how I got to know about it. 00:38:21.000 --> 00:38:28.000 So I listened to their interview. Then I listened to the interview with Jensen Huang, who is the founder of NVIDIA on Acquired Podcast. 00:38:28.000 --> 00:38:35.000 And I really liked the way these guys, they have done three episodes on NVIDIA where they break down NVIDIA, they do all of the research. 00:38:35.000 --> 00:38:39.000 So they said they've done about 500 hours of research on NVIDIA. 00:38:39.000 --> 00:38:44.000 And then they got to interview the founder himself. So their questions were very prepared. 00:38:44.000 --> 00:38:49.000 In a way that they don't need to be around the bush. They know so much about the company that they can go really deep. 00:38:49.000 --> 00:38:53.000 I really enjoyed that. And I really enjoyed hearing Jensen Huang talk. 00:38:53.000 --> 00:38:57.000 I didn't know who he was until recently. He's very wise man. I really liked listening to him. 00:38:57.000 --> 00:39:04.000 Acquired is actually one of my favorite podcasts. Their episodes are incredibly in depth. They're very long usually. 00:39:04.000 --> 00:39:14.000 My favorites are the Costco one was amazing. Like I didn't know about how Costco came to be and how much of an influence. 00:39:14.000 --> 00:39:19.000 All of these like Kmart, Walmart, Costco, how much they've had across the industry. 00:39:19.000 --> 00:39:26.000 And the other one I really liked was the Nintendo story. They go all the way back like 100 years and start from the beginning. 00:39:26.000 --> 00:39:33.000 In fact, that's the episode I used to demo our app at the last Vancouver launch event in like end of January. 00:39:33.000 --> 00:39:37.000 That's nice. Yeah, they also speak very clear. So the descriptions work pretty well. 00:39:37.000 --> 00:39:41.000 Oh, they had an interview with Charlie Munger. That was the very first episode I listened to their podcast. 00:39:41.000 --> 00:39:48.000 But they were sitting in the restaurant. It was just a few months before Charlie passed. So he was 99 years old. 00:39:48.000 --> 00:39:57.000 And his speech has this kind of crackly kind of sound like old man's speech. And I could barely understand anything that he was saying. 00:39:57.000 --> 00:40:03.000 So I actually ended up not finishing the episode. But I was impressed that these guys must be big deal if they talk to Charlie Munger. 00:40:03.000 --> 00:40:08.000 I remember this because they mentioned Costco. I think Warren Buffett and Charlie Munger, they invested in Costco or John. 00:40:08.000 --> 00:40:11.000 And they had a conversation about Costco. 00:40:11.000 --> 00:40:18.000 I think Charlie Munger or Warren Buffett has some famous quote about Costco and investing in it. I'm forgetting the exact thing. 00:40:18.000 --> 00:40:26.000 It wasn't a Costco episode though. Cool, cool. Yeah. It was a great low energy but she'll kind of episode onwards and upwards. 00:40:26.000 --> 00:40:38.000 And upwards to our beta to the top and right. Hecky stick. So yeah, again, I don't have enough energy to tell people to leave us reviews and reach out to you. 00:40:38.000 --> 00:40:46.000 But if you're soliciting to this, please do. Yeah, you can find us on metacast.app. And there you can find all the links to our newsletter. 00:40:46.000 --> 00:40:57.000 And this podcast, the other podcast we have called builders gonna build. And of course you can send us an email at hello at metacastpodcast.com. See you next week.