How I Helped Grow Diary of a CEO to 14M Subscribers

When I started, we were on 100,000 subscribers on YouTube. Now we're at over 14 million.

My guest today is Grace Miller. She's the head of failure and experimentation at Flight Story, the company behind the Diary of a CEO podcast and a bunch of other big shows. If you're growing a YouTube channel, what do you set as your core metrics?

I definitely would say— I just said to Steve, I was like, we really need to do this. And the growth has just been insane.

Grace walks through through the YouTube experience they've run, some that worked and some that failed in a huge way.

YouTube Community, now it's called Posts. Lots of people don't share on it. I'm intrigued to why people don't share because I think it can be a really good growth hack. There's so much on YouTube that can be experimented with. Chapters is a massive one.

What's another experiment that was maybe at a different scale?

A big one right now is—

This episode was a fascinating look behind the scenes into one of the world's largest podcasts. And so if you're as obsessed with creating great content as I am, you're going to love it. What are 3 examples that you're like, okay, every creator should be doing this?

The first one is something to do with whatever content you're posting. Second one is something like a poll or an engagement post, so getting people actually interacting with it and commenting. And then I say the third one—

oh, Grace, welcome to the show.

Thank you for having me.

Okay, so the Head of Failure and Experimentation, that is a title that I have never heard before. Is it a title that you've heard before with other creators?

Before I came into the title, no. Now since being in the title, there's a whole world of experimentation I didn't even know that existed. And I think in the creator space, creators experiment themselves, but they just don't call themselves head of experimentation, right? So it's more of a behavior than I would say it is a title alone.

Yes.

And everyone should be doing it. Um, but it's really cool to have the title and now meet other people in the industry with similar titles as well now.

So I want to dive into a couple of the experiments, maybe some that failed and some that succeeded.

Yes.

Um, and we can talk about, you know, both aspects of it.

Yeah.

But before we do that, for anyone who doesn't know Dire of a CEO, Flight Story, all of that, like, give the high-level context and maybe a few numbers because the scale is pretty ridiculous.

Yes, it's actually my, um, 4 years today at the Dire of a CEO.

Oh, congratulations!

I know, it's flown by.

Thank you for celebrating anniversary here on the show.

I had to come talk about it.

Yes.

Um, but The Diary of a CEO has been around for about 5 and a bit years now.

Okay, that's— I would have thought it was way longer than that.

Yeah, more— when I say 5 and a half years to camera on YouTube, full-scale production. And when I started 4 years ago, we were on 100,000 subscribers on YouTube, and then now up till today we're at over 14 million So it has been insane growth over— well, that's 4 years. Over 4 years, it's gone from $100K to $14 million, which is wild. And it's also gone— almost developed as a podcast from when I first started listening. It was very entrepreneurial, business-focused. Now it's developed with Steven as he's grown up and become even an even more diverse entrepreneur. And it's now covering topics all around the world from people that are the best in their industries. And a year and a half ago, we started Flight Story, which is the media company that The Diary of a CEO is now under. And that's a podcast agency, but also has creators as well. So the thesis behind it is we create media companies or podcasts and help creators develop in that. And so we've gone from just The Diary of a CEO to also podcasts with Paul Brunson. He has a podcast called Winnie's Talk. Um, we have one called Begin Again with Davina McCall, who's a UK broadcast icon. Um, and then The Line with Dr. Kristin Holmes from Whoop.

Okay.

Which is, um, really, really interesting. It's all about her interviewing, um, coaches or sports people and how they have become the best in their industry. And then we recently, um, didn't start this one from the beginning, but Hot Smart Rich with Maggie Sellers. So she started her podcast a year ago. And then a few months ago, um, Flight Story partnered with her to help her build it even further. So they're the current podcast on our site.

Yeah. Oh, that's amazing. So you get to learn not only from Diary of a CEO, but also from all these other shows that, that you're running.

Yeah, we learn from all the other shows and then we also have other divisions. So we have a Flight Socials division, which is creators, um, specifically on short form. We also have a product division, so get to learn a lot from e-com. Um, and then also Flightbooks, which is physical books and helping publish creators as authors.

I love that. Okay. So diving into some experiments.

Yes.

What's an experiment that you ran that you think would be interesting to share with the, like the professional creators who, you know, would be fascinated by that?

That failed or succeeded?

Ooh, whichever one you want to start with.

I think one that was really interesting is we have a DOAC Clips channel, which is still, um, horizontal clips on YouTube, but it's shortened versions.

So, so there's not Shorts, it's—

yeah, it's not Shorts, it's Clips. Um, we call them Clips, and then it gets confusing with long short-form social media, but it's, um, 10 to 30 minute cutdowns of the podcast.

Okay.

And we experimented, we actually put one out every day, and we experimented with doubling output. So we thought, okay, if we put 2 out every day, we'll double views, right? Basic hypothesis. And I think with experimenting, it can be as simple as that to see does it or does it not work. Um, we put it out and it was a massive failure. It didn't double views, it actually decreased reach than what we thought would happen. So we thought, okay, say we get 10,000 views on each clip, we're like 20,000 views. No, it was so much lower than what we actually expected to happen from it. And so it's— we've then pivoted back and we're like, okay, we'll go back to one episode, uh, one clip a day, um, to then retest and see what other areas we can experiment in.

So the measure of success for this experiment, because that's a very— a lot of people are thinking like, oh, let's try this. And they're like, great, what's your hypothesis? What, you know, how are you measuring success? And a lot of people are like, oh, well, I was just trying an experiment. It's like, it's not an experiment unless you define the criteria.

Yeah.

But your measure for success was total views on the channel?

Yes.

So not views per clip necessarily.

No. Yeah, we thought, okay, in a 30-day period, how many more views can we get? And does that equate to the time or effort that's been put into creating those clips? And so making sure it balances out, that the views are justified by how much more work is going into it. And it wasn't— it didn't work at all.

So total views in the 30-day period Stayed flat or went down?

Went— stayed flat, but they didn't double the number. Yeah, our hypothesis was double, and they did not get to anywhere near double.

And then, so for that to be the case, views per clip went down pretty substantially.

Yeah.

Okay, do you have a theory on why?

I actually think it's something to do with the algorithm and maybe posting multiple clips a day. I know other channels can do that. Um, but seeing the growth of our current clips channel, then going to 2 clips a day, it's finding its place in the algorithm, I think. And when you look at the YouTube algorithm, how many times— if you actually really delve into it, very nerdy— but how many times do you see a video in your feed from the same creator over and over? So you're almost cannibalizing your own content if you're posting multiple a day because you're probably not going to show that. Yeah, maybe within the subscriptions feed You might come up multiple times, but I think in the like homepage feed that you're rarely seeing unless it's on different channels. So for example, The Sidemen, I follow multiple of their accounts, so I might get a short, then a long form, then another long form, but from a different channel. When it's on the same channel, I actually don't think you can get as many placements on a homepage feed.

So is that why you see so many creators running 2, 3, sometimes 4 YouTube channels?

Yeah, I think so. And I think, yes, last night I was speaking to some of the creators and, um, one couple said they had 6 channels, and I thought, um, that's interesting. And combined, I think they had about 60 million subscribers.

Okay.

But the proof is there. I think if they're— not only for volume, but I actually think for topics as well, especially on YouTube, knowing why people are coming And so they had a gaming channel or a lifestyle channel, and so they were fairly distinct.

It wasn't just the clips versus—

yeah, it wasn't. I don't think there was volume. It was definitely keeping different topics apart, which is another thing, because I think people go into YouTube and they go, especially if they're more lifestyle or vlog, and think, I'm gonna service everyone. It's like, actually, you really have to know what people are coming for, and whether it's you as a person or a podcast and understanding why they'd come back and keeping the topics similar but still diverse enough to bring in new people.

So how many channels do you have for Diary of a CEO?

We have— so we have our main channel, and then we have the Clips channel, which is the short form— well, horizontal short form clips. Then we have, um, a separate Shorts channel, which we tested on, but now we've gone back to Shorts on our main channel. So that was kind of—

I want to dive into that in a sec.

Left aside. And then we also have Behind the Diary, which is Steve's vlog.

Okay. So 4 channels? 3?

4 for Diary of a CEO, yeah. But we're, yeah, the Shorts one's been made redundant.

Some of these other channels that you're running, um, or sorry, let's take another podcast that you're running. You know, now you're sort of spinning it up and, and, you know, growing a new show. Are you running 1, 2, 3 channels for one of those shows? What's the process?

1 right now. So the goal behind it is grow that 1 channel and then we can spin off.

And what, what's the, the mark where you would think about, oh, maybe it's time to spin off? Is that 50,000 subscribers? 500,000?

I think it actually depends on where— like, a multitude of factors. So we have so many different frameworks depending on where we want something, whether it's a podcast being acquired or how we want an episode to go out. And I actually think it's a multitude of things, from views, subscribers, but also retention. So how engaged are they? Would those people convert to another format and also the length. If your podcast is 30 minutes, you're probably not going to go and create another channel of 30-minute clips, right? And whereas if it's, for example, like a Huberman where he's 4 hours, of course he has the chance to go and cut it down.

Yeah.

And I think it brings a different audience as well. So short-form, short-form clips where they're horizontal— we needed a better name.

Yeah.

Um, it's a distinct, um, but I would say for Huberman, he could definitely do that because they would bring a different audience to who would come and consume the full length. But you then also get the conversion of people jumping over to the full length.

Right.

Okay.

And then something that is widely debated, and it sounds like you've experimented both ways, is whether or not your shorts, the vertical YouTube Shorts, should be on the main channel or should be separate. Like, uh, Pat Flynn would be an example with his Pokémon channel, uh, Deep Pocket Monster. He has his long-form horizontal channel, which I think is 1.5 million, 2 million subscribers. And then he has a Shorts channel that's also 1.5 million, 2 million subscribers. And he says that they're entirely different viewers. Like there's overlap, but like it's not the exact same people. And so he's very much in the camp as last time I talked to him of like, keep them separate. What's your take on that?

We experimented with it for a while and on YouTube Shorts specifically, we tested off-channel and on-channel. We tested off-channel first because we obviously thought, as every YouTuber thinks, it's going to cannibalize your long-form content. Let's bring you a different audience. Um, but there's a lot of missed opportunities. So you can't link the video on your main channel if you're posting a Short and it's on a separate channel. Whereas if the Short is on your main channel, you can link it. So people are going directly from the Short to the long form. Super easy conversion funnel. Um, and so, but again, then it begs to the question of what I said before of taking up places on a feed. Do you still get the same placements if it's all on one channel? We now fully believe on the same channel is the best way.

Okay.

And we tested that across all of our podcasts. So we tested it on some of the newer podcasts first, which is a really interesting dynamic, and they're actually growing faster than The Diary of a CEO at if they were where the Diary of a CEO was now. And so tested it, it actually works, it converts. Obviously they have shorter average view duration because they're coming from a short-form video, but the actual conversion funnel is so much smoother and you're getting all of the views on your main channel. So YouTube's seeing that as more people coming there rather than separating out people going to other channels and not coming to that main channel. So we've seen a lot faster growth since then.

So when you say you tested it, a lot of people would— you look— I look at the tests that they ran and I'm like, you did it and you went off of vibes.

Yeah.

What I want to know is, how would— how did you actually measure those tests? And what, you know, what was the time period? How would you know, like, okay, we fully believe based on this data that this is the right decision for us?

Yeah, we, we have a bit of a cheat code because we have other podcasts, right? So it's slightly easier because we can test things on a podcast and then apply it and test it to other podcasts to see, does that hypothesis still hold? So we tested it on other shows with— that were actually starting out, so you can kind of see the growth more. And it was understanding how many people convert from those short form. Does it hurt the long form? Because I know people say putting short form then cannibalizes your average view duration of your long-form episode.

I've heard that a lot. I saw someone on X saying that, like, you know, they were very strongly saying, like, you should never do this, never combine them.

Yeah.

And then I talked to, uh, the team at 7X Content who edits this show, right? And they, they work with a lot of, you know, a lot of YouTubers, a lot of podcasts. And they were like, yeah, no, I mean, they're in the same camp as you of like, no, I'd keep them on the same channel.

I know it's been almost like a myth for a while, and I do understand maybe it's also different industries and niches, right, that would have different behaviors. Because I think as much as things can be applicable to all of YouTube, there's some things that could just be a specific niche as well, like the way MrBeast edits would be very different to the way a podcast should edit because they're different types of retention and different lengths and understanding. Maybe you're not going to cut a podcast every 3 seconds to a different angle because it would be jarring for someone. But then someone Beast-esque would probably be able to do that and get away with it to keep retention.

Right.

So I understand why people might still be worried about the Shorts, but we've seen really great growth from it. And it's not about— the views haven't dropped, put it that way.

Yeah, the—

yeah, long-form views haven't changed at all when we added Shorts.

Okay, that's helpful to know. I think a lot of people love that example. And I love how you're saying, you know, we have multiple shows and so you can test that, like, right? You actually could have a control group and, you know, it's like, we're not quite statistically significant in like testing this across 100 shows, but you know, you can do it across a handful. What are some of the other myths or mistakes that you see people making where they're like, oh, I, I believe this, and you're like, I don't think that's true?

Ooh, I think, um, YouTube Community, or now it's called Posts. Mhm. Lots of people don't share on it.

Okay.

But I, and I think more because of time, but I am intrigued to why people don't share because I think it can be a really good growth hack. Okay, because you're getting a placement in the feed when you share. I would— not to give away all my tips and tricks, but I would put it on a day that you're not sharing any other content.

Okay, so you might— if you were doing 3 clips a week, or like your long form and then 3 clips a week, it's like, okay, well, there's 3 more days that you didn't post anything.

Yeah, I— and if you see it, you— the more you think about posts, you'll probably see them way more, but they are always in your feed. They're in the subscriber feed, they come up in the homepage feed, and I think they're really underutilized because I've done a lot of deep diving into them. And when I go on pages, hardly anyone uses them. And then the creators that do use them are obviously popping up because YouTube's algorithm says, oh, it's a post, we're going to share that. And so I think posts, people obviously are not even thinking about, or maybe thought wasn't very useful, but that is a big one that is kind of another way to get in that algorithm.

I think from working on the software side of things where I know that you end up with a product manager who owns a certain feature, you know, within a company and they have metrics for usage and adoption and all of that. And so you know that their job is to drive adoption in that. And so I think about something like posts or community where there is a product manager at YouTube. Who their job is to drive that. And so they're like, okay, what levers can I pull? And so like, oh, if someone uses this, then we're going to get it featured here.

Yeah.

And so it's like, give them what they want because they're here to reward you with it.

And they put— they create the features for a reason, right? And then, for example, like when you jump on that feature, they are going to push you. Like Instagram with their current translations feature, they're going to push people that say, yeah, I'll add that onto my video. Right, because they want people to be using the feature. So why not actually go and be a first mover in that area and try the feature out? And so I think definitely the post feature. The other ones on YouTube— translations is a big one that's, um, coming out because they've obviously allowed audio tracks now to other creators. And the thumbnail testing and title testing tool that I think is now available to everyone, okay, and is maybe one of the most interesting tools that we've finally been able to use because it's based on statistical significance, whereas all of the other tools before were amazing for thumbnail testing, but they don't have the backend data to be able to prove what does and doesn't work on YouTube. So it's finally, it's like, yes, we finally have a tool that is actually in-app, um, and we can use that has real data.

So what kind of content should someone be putting in posts? Like, what are 3 examples that you're like, okay, every creator should be doing this?

Yeah. I'd say the first one is something to do with whatever content you're posting. So whether it's the next day after, it's like, hey, this just came out, because we've tested that and we've seen when you put a post out, people then go through to that episode.

Okay.

And it's really interesting to experiment with it because you can see a spike when people go to that, from that post to the episode. So it's not just, oh, I'll put out the post and hope it does something. You can really see the data. Second one is something like a poll or an engagement post. So getting people actually interacting with it and commenting so that the algorithm's seeing, oh my God, this person is really engaged with that content, and then they're going to be serving you the other content. Yeah, definitely. And then I'd say the third one, maybe, is something to get feedback from the viewers. So not only in the comments section, but being able to actually have almost like a conversation thread with them. To get real feedback to help you produce better content, whether it's a podcast and what guests they want next, whether it's challenge videos and they suggest new challenges. But actually getting them in that feedback loop can be really valuable first-party, like, information as well.

Well, what's fascinating about that is you could have a basic flywheel that gets you, you know, all of those things. So, so you're saying, you know, I'm going to be posting 4 video content 4 times a week, the long form, and then 3 shorts or clips. And then, you know, fill in the other days with posts. And they're pretty, you know, we're talking it could be a 15, 20-minute lift total for the week to add in posts.

Yeah, it's, and it shouldn't be hard to do something like that. It could be really quick. It could just be a quick sentence and call to action, then a link to your episode.

Have you tried like LinkedIn-style posts where it's actually storytelling or something else that's Yeah, we experimented.

If you scroll quite far back, it was maybe a year and a half ago, um, in our posts, and we experimented with different quotes. Actually, maybe even— yeah, I'd say a year and a half ago. And we experimented with different quotes, photos, behind the scenes, and really tested like what works, what doesn't, um, to see like— even they have like almost like carousel or multi-image options. See like, does that get better engagement than just text. Yeah. Um, so there's so much just in one tiny feature that people aren't even using to then build out.

And you found that it was worth going into that level of detail, or you said you scroll way back, and so that makes me think that it wasn't as worth the effort.

I think it depends on what you're trying to do. Like, that was great for if you just want engagement on that specific post. We do more now linking through to the episodes.

Yeah.

Or alternatively You can also see that we'll, um, do posts that link through to our Clips channel. So we'll post on the main channel saying, hey, this is—

let's do cross-promote.

Yeah. So instead of like what I said before about Shorts not being able to cross-promote if they're on a different channel, it's kind of a hack where you can promote on posts and link another video. So I could— if I had a YouTube channel, I could link your video on my post, basically.

Yeah. Well, and what's interesting is you might pick up someone where it's like, okay, I saw that this 2 and a half hour episode with Jefferson Fisher dropped, and I would love to watch that, but like, I don't have to, you know. But then you link to the clip of it. Yes, that's 15 minutes. It's like, all right, well, I'll— yeah, like, that will be my gateway into the episode.

Oh, 100%.

And go from there.

Yeah, but there's so much on YouTube that can be experimented with. Like, it's really— post is so tiny compared to— yeah, other things like chapters is a massive one, specifically in podcasting. I think people sometimes skip over Because it seems like putting chapters in. Yeah.

Okay.

There's some podcasts still that I'm like, why don't you add chapters? But I don't know if they're maybe experimenting with it for retention and average view duration.

So are you— okay, so let's talk about chapters. It sounds like that's another thing that you believe that other, you know, podcasts on YouTube should adopt.

Yes. Yeah. Or creators as well. I think if you're making a vlog that's an hour or an hour and a half, even 45 minutes, to be able to allow people to skip because YouTube created the jump ahead feature where if they see someone skipping, they'll jump ahead to that. Yeah. And it basically, it pops up, it says jump ahead, and you can click that and it takes you to the section in the video where most people have skipped to, which is almost a growth hack to see. So I could do it on someone else's podcast and look at it and see where people are skipping to. And usually you'll see a lot of people skip the start if it's a trailer or a summary, because they're like, I already know I want to watch this.

I watched the trailer on Instagram. That's what got me here on the shorts.

Yeah. And so you can really see throughout how it progresses in the skip ahead or jump ahead. Sorry. But then chapter-wise, it actually gives people, if you're adding chapters, a section to know what to skip to in a good way because you don't want people coming and then bouncing because they can't find the section that they wanted to come for. But also then you don't want to hurt retention. So it's an interesting balance of making the chapters worthwhile clicking for, but then not too obvious that they then skip to them.

That's an interesting balance. Okay, are there any other, um, either myths or mistakes that you see people making?

Thumbnails. So many different— I think people try and copy other thumbnails, and I actually saw— not gonna name the creator, but I saw they copied the exact thumbnail and title for an episode.

Okay.

Um, and it didn't get anywhere near— I think it didn't even get over 100 views. But the episode on DOAC— actually, this is a great myth. I'm going to restart the story because I think it could be really interesting. Um, a creator copied the exact thumbnail and title of a Diary of a CEO episode, and the DOAC episode got a few million views. That creator only got 100 views on the episode. And I think what's really interesting is people can think, oh, title, thumbnail, copy, algorithm thinks I'm the same creator or channel. And it's breaking that myth of being actually no, right? If you create original content that people are going to like, they'll probably stick around for longer than actually just trying to copy someone else. And you'll probably see a million channels on YouTube now looking the exact same thumbnail style with the white text, red highlight. And it's finding your own spin on it. Even if you have still the two faces of the creator, the host and guest, and still have text in the middle, I think there's a way to take a spin on it. And even the whole format, like I saw, or I know a creator in the US who has started a podcast, interesting name called The Rubdown.

Okay.

But, but the concept, when he showed me it, It gave me goosebumps because it reminded me of early days of Chicken Shop Date with Amelia.

Okay.

And his concept is, um, they're instead of sat in chairs, they're on massage tables getting a deep tissue massage while having a conversation. I don't think you get the same eye contact as our conversation, but I think it's a really interesting spin on podcasting. I'm like, that is nice to see people changing it up, revolutionizing what podcasting is, um, and trying new things instead of just sticking to traditional thumbnails that are out there or the titles.

Right. Yeah. There's so many things in that where, you know, you think about the hook that you're going for and is it memorable? Could you talk about it? And it's like, oh, we have another 90-minute long-form podcast interview with the same guests as everyone else. And you can't really, like, if we think of word of mouth as a growth angle, like I can't articulate that show in an interesting way or that sort of thing. But you're talking about this show, The Rubdown, right? It has a title where you're like, what's going on there? And then you And you're saying, like, this is a little wild, but they're getting massages on massage tables. Well, you know, it gives you something that you can talk about and explain. And someone's like, I don't know how I feel about that, but I guess I'll click the thumbnail and watch it.

Exactly. And now, and now I bet you'll see it in your feed somewhere and you'll be like, oh my God, that's that podcast. And it really stands out. And it was like Chicken Shop Date. She's in the chicken shop. I know it technically wasn't a podcast, but it is a YouTube format. And I think it's finding that Even the format of MrBeast now, I would say people call the MrBeast format, whereas really it's almost like challenge videos or— yeah, um, but yeah, almost becoming the new genre of your YouTube or podcasting format is quite an interesting new experiment and strategy to go down.

Okay, so now I want to go into an experiment that worked.

Yes.

What's one that comes to mind that you're like, okay, this is something that you learned a lot from and and it drove results for the channel.

A small one or a big one?

Let's do both. Let's start small and then go big.

A small experiment we've been running is the collab feature on YouTube where the video appears on the creator's page. Well, both creators' pages, right? So if I posted a video, it would appear on mine, but also if I tagged you, it would appear on yours. And it's a really interesting dynamic because you're getting the views and subscribers from both pages engaging. and so you're able to see how it reaches new algorithms. And I think it's worth playing around on every channel because you'll have different experiences to what we would have experienced, but you'll find it interesting to see like where does it get served, and also understanding how does the other person who's tagged have an audience that comes and maybe has a different retention, or they might come from different places rather than the homepage. They're all coming from external, so seeing how that works, um, it's been a really interesting— and not only the collab but also the tag feature in titles. So seeing whether that does or doesn't work, I actually— I think collab works really well. I'm not sold on the title tagging feature yet.

How did you design the experiment around collabs?

Yeah, it's more— less statistically significant and more— we— I kind of— not to go too deep into experimentation, but I kind of see it in a few ways of Some are great statsig— statsig, what did I say? Some are great statsig experiments, some are more knowledge-based experiments. So I see it as if we can test this and see, can we learn more about the algorithm that will then help us with other experiments, that's going to be a win. So it's more for the tagging, what can we learn that will help us? So can we learn more about other creators' algorithms and how they work? Do they work the same as ours or not? But also setting it— like, we set parameters of how many views or new followers do we hope to get from this episode? And did it hit that, or did it just exceed it by tagging that person as well? And what sort of growth has that had from it?

Yeah, that makes sense. Okay, what's another experiment that was maybe at a different scale?

Yeah, a big one right now, which I touched on before, is localization and dubbing. And I was chatting last night to some of the creators about it as well, and I didn't realize how many other people are doing it. And the industry is really picking up on it right now. So not only YouTube, but also Instagram now has the feature of the translations. And we've been experimenting with it for— we started 3 years ago testing it. Um, it wasn't the time, but we started testing, seeing what worked, what didn't, and tested different channels. So there's kind of different ways you can do it. You can do human translations or AI translations, and then YouTube's obviously added the feature now where you can add the translations to your video. So instead of what MrBeast used to have was his English channel and a Spanish channel, he now has all of his audios in one channel. So you're not only having English subscribers, you're able to have subscribers all over the world on one channel.

Okay.

And I would bet money on it that that's why his growth has been so big, because he's had so many languages at his disposal.

It's a global audience.

Yeah. It's such an easy growth hack. Obviously, it takes time and effort to make sure the quality is good. And I know a lot of theirs were still human dubbing, but with AI dubbing now—

So is that where— are you cloning a voice in ElevenLabs and then—

There's a few ways. You can do so many. So YouTube has an in-app feature.

OK.

I sound like I'm selling for YouTube.

Um, I mean, you know, YouTube has done a lot for, for you and Steven, so, you know, you can—

um, we, we actually can't use it because— so YouTube's in-app feature, but you can only use it for under 1 hour. So your clip has to— or episode has to be under 1 hour.

And you're like, that's not how we work.

Yeah, I'm like, why, why not slightly longer? Um, but so they've built an amazing tool that you can use in-app. I think there's a lot of work to go though. Like, it's good, it's not controversial, not amazing, right? Um, even though I just said amazing tool, but I think it's, um, a really good build on what other platforms have offered because they currently— a lot of platforms weren't offering localization. It's the fact YouTube are now offering localization across so many different audios, and then being able to get creators to post that, they're not only reaching like a wider audience themselves as YouTube as a platform, but creators are also benefiting from it as well, which is nice. So to use that, it has to be under 1 hour. I think there's some improvements to go right now. From conversations I've had, a lot of people are still using outside platforms and importing the audio. And then the other part is it's audio only right now on YouTube, so you can't do lip syncing, right? Whereas Instagram's feature is incredible. I got a video the other day in Hindi and I could— it was English. I didn't even realize she was speaking Hindi until I saw the tiny translated button. So I think seeing what Instagram is doing is really, really cool, how innovative it is. But it's— I love YouTube because they experimented it and they were a first mover of testing what worked and what didn't, and they're building it themselves to try and get better.

So, and they've got translated thumbnails as well now, which has been a really interesting So we do this thing on the podcast where we always talk about how to grow creator businesses, and I've realized I don't actually talk about my business. Sometimes people will ask in the comments, so what do you actually do? Are you just a podcaster full-time? So I'm interrupting this show to tell you that I actually run a software company that is built entirely to help creators grow their online business. It's called Kit. You can check us out at kit.com, and it's just the place where all the top newsletter creators, podcasters, authors grow their email newsletters, connect with each other to grow faster. And so much more. So if you're looking to increase revenue, automate more in your business, and just generally make this year even more successful, go check out Kit. You can sign up at kit.com. We have a full migration service where we'll bring you over to Kit totally for free. It's been a labor of love to build this business over the last 13 years. We've got a dream list of clients on there. Everyone from James Clear and Tim Ferriss to Dua Lipa and Matthew McConaughey. So it's a pretty special group of customers. And a really great team just who's really here to help you be successful in your business. So then is the, the DOAC team is translating or dubbing offsite, off platform, and then uploading these files?

Yeah, so it started as an experiment. Okay, so there's no— we don't even have a production team doing it yet. Our production team is still doing English, even though, um, Berta, who's one of the amazing producers, um, she is actually Spanish, so it could be perfect. Yeah, because, because she'd be able to fully translate and then quality check the episode. But the translation started as an experiment. We were the best way possible having conversations with people and heard localization, translations coming up more. And I just said to Steve, I was like, we really need to do this. We've been testing it for a few years, but this truly is the year. And the growth has just been insane in 2025. And then now obviously seeing it grow in 2026, we've got some big goals. But we just started the Spanish Spotify channel as well. On the 1st of January it launched. So very new things to come. Spotify doesn't have the feature of adding the audio yet, but we'll see what happens in localization.

And does YouTube split out the views based on, like, do you get analytics of how many are viewing in each language?

Yeah, you get everything from views, retention, Well, ABD, being able to see even the countries based on them. So if it was Spanish, are they Spain or LATAM? And really understanding, but it's just the whole world out there of how do you dub? I know some people have created, or some creators have built their own dubbing tools because they didn't find any on the market that were good enough. So I think there's lots of tools out there that are good enough though, and it's about what level you're at. But I actually don't think any are 100% perfect yet because the AI has still got a long way to go. But I think the best part is trying it and starting to get the views and experimenting with it. And then when the AI is 100% ready, you're already so far down the line that you're not starting from zero.

Yeah. Okay. So I want to know, going back to experimentation as a whole, how you think about designing an experiment.

Yes.

Like, do you have specific criteria that you're going through a process and then how you measure it? It's the time that you're going to run the experiment with the original hypothesis, all of that.

Yeah, we, so we look at the company overall goals and KPIs first. So whether it's yearly, 5-yearly, whatever your team sets us, maybe even monthly. And we'll look at overall company or podcast goals, then the team goals and say, what do we want to work on? And there's a really interesting photo our manager showed us of arrows pointing in all different directions and then arrows all pointing in one direction.

The same direction, the same number of arrows and all of that, but like Are you making a tiny bit of progress in many areas, or are you—

and I think that's the key to experimentation, is people can be, oh, I'm gonna— or people, I think people will say, I'm gonna try something new and experiment, but it might be so off-piste towards what your goal is that actually, what is the point in doing it? So say your goal is to hit 10,000 subscribers on YouTube. What do you need to get there? Every single experiment should be aligning to either learn new knowledge of how you can get there or gain more subscribers. And if you're doing something that's like, um, so far removed from that experiment, I think it's about refocusing. And sometimes we've run really random experiments that actually— like, I'm so far down the rabbit hole that we're like, bring it back, what do we need to learn from this? And going back and saying, okay, how do we get to that answer? And That's what helps the most. I think the ones— they don't have to be, as I said before, they can be really simple experiments and don't have to be massive life-changing things, but the little ones, for 1%, as lots of people say, can add up to those goals that you really want to achieve.

If you're growing a YouTube channel, what do you set as your core metrics? Like, maybe what are the things that people get caught up in that you're like, that doesn't actually matter, versus what, you know, on a day-to-day basis or, you know, monthly, you're like, no, this is what I'm actually measuring and judging success based on?

I definitely would say, well, it really depends on what the goal is. If it's having a small community of really engaged people, then you'd look at subscribers and retention. If you are wanting more mass, it could still be subscribers, but maybe it's views, and you're willing to sacrifice retention. I think you can do all three though, I think, but it takes a bit longer.

Okay.

And what I think the other people— the other people— I think people forget is, um, it is a long game. Lots of YouTubers and creators I've met have been doing it for 10+ years, right? And obviously they've got lots of subscribers now, but at the start, like, um, Sean Does Magic, who did the talk yesterday, he had 800 subscribers for 2 or 3 years until he started getting subscribers. And I think a lot of people would quit and go, oh, but I'm just going to stop now. I'm not even getting views. And so it's the mentality to keep going. And but in terms of what you'd actually look for, I think for the long game, if it was a podcast or a creator that wants to do similar content, building up your subscribers, but also keeping views and retention high. So how do you keep it engaging throughout? How do you— whether that's chapters so people know where to go and can continue to be engaged. Whether it's engaging with comments so people feel like they actually know you and they keep coming back, whether it's the post and asking what people want so you can make actual content around what people are looking for. And then the other side of it is being so interested and actually caring about the analytics to find out what is valuable to you and what's not. So one of my favorite analytics that YouTube added recently is like new versus returning versus casual viewers.

Okay.

And seeing how many people are new to your channel. You obviously always want a specific percentage of new people. Returning, good. Casual— how many casual viewers do you want? You want a lot more returning. That means they're coming back for more. They liked your videos, they're hooked. Yeah. And I think casual, fine, it's good that they come back. You're never going to say no to a view. But I think you really want to try and build out that returning or more engaged area and say like, hey, keep coming back? How can we keep you— obviously they'll move to a different section later on, um, but how do we keep you returning and interested in our content?

Is there anything that you've done across either DOAC or the other shows that has made the biggest difference to getting those returning viewers?

We've actually— it's one of our experiment focuses right now.

Okay.

So, um, on Begin Again, you'll see, um, we've done almost like a themed month around health. So all the guests are around health, it's around wellness, obviously start of the year, perfect timing. So, um, for January they all have a similar topic. So I was thinking if someone's interested in one health guest, they'll probably be interested in multiple. So being able to hopefully bring them back. And so that experiment is still running. Other ones are things like a, um, playlist.

Mm-hmm.

I should have said this before when you're like, what YouTube features? But obviously we can talk about it now. Um, playlists are maybe a slept-on area as well.

Okay.

If someone likes your content and they're in that playlist, they're gonna be recommended more of the same videos. So whether it's finance, whether it's content creators and learning how to develop their, um, platforms, the more you can almost tailor that content to them, So they finding it better. And also then the algorithm realizes and they're like, oh, they've maybe watched 4 videos out of this playlist, even if they haven't clicked in the playlist. I'm going to serve them more in the recommended. And so we've seen that. And also playing around with the names of the playlists as well.

What's an example of that?

Instead of just finance, it's like, um, the best ways to make money or the best ways to save money, how to build wealth.

Yeah, something along those lines, rather than searchable or engaging, rather than just finance. Let's be a little bit more interesting.

And, and we've seen that, like, if you even played around with testing different names, so you chose one playlist and over the space of 3 months— I know that sounds long, but because you want to give the algorithm enough time to find it— um, changed it every few weeks, you'd see how the, yeah, different views change.

I mean, it's fascinating the amount of information that YouTube gives you.

Yeah.

Like both YouTube and Instagram are amazing as platforms of like, look, we'll tell you exactly what works and why and all of that. And a lot of creators are not actually using that. And I always think back to the, you know, the product manager.

Yes.

Who has their own KPIs and they're running their own experiments of like, how do I get, how do I drive user creator behavior on the platform?

Yeah.

And so it's like they're giving you the tools so that you can do the thing that they want you to do. And so like work with them.

It's spot on. And they wanna know as well.

Right.

We get so many of the specific product managers. There's been some like Spotify ones, for example, that come and say, can we get your thoughts on that tool?

Right.

Because they really wanna know, did it work? What could have been better? Um, so there's so many that we've been asked about, and I'm sure other creators or even could share feedback as well to then build that relationship. of providing what does and doesn't work, because that's the only way they're going to be able to develop better tools as well, is by getting real feedback.

The other thing that I would say is you can find out who all these individual product managers are because they're all on LinkedIn. They're all, they're all out there. And you might say like, okay, of course Spotify is reaching out to Dyer, the CEO, because it's one of the biggest podcasts. But, you know, me, I have this little show. But, you know, knowing from behind the scenes in software, like the product managers want to hear from the whole range because they know, oh, this works for big shows, but we actually have to get the long tail going.

Yeah. And that's how they found me for that one of the meetings. It was on LinkedIn. I posted about a feature and said, this is a really cool feature Spotify has launched. And the product manager literally messaged me and said, can we go on a call so I can learn about what you think of it?

And I was like, it's so, so good because they all have goals around user research. And understanding this. And so you can talk to them and then you can just ask, hey, would you mind giving me feedback on this? And they're like, oh, we're actually behind the scenes, we're trying to push this metric. And so this is how it works.

Yeah, it's— I think the connection side as well of learning from other creators or the platforms and understanding actually it's not so much a competition. Yeah, like I listen to podcasts, I'm not only going to listen to one podcast, right? I can listen to other podcasts or YouTubers. Like, there's no— there doesn't have to be the secrecy. I think the sharing of knowledge can be super valuable to help everyone get better as well, because everyone's trying to learn how does the algorithm work, right? What works, what doesn't.

Um, so I want to go to even on a show the size of Diary of a CEO, you know, you're 14 million subscribers. There are episodes that get millions and millions of views, and there are episodes that get thousands of views. And, you know, you all put out a lot of episodes. What do you notice in those trends where, you know, you might expect this one, okay, this one's going to do really well, and it actually doesn't come anywhere near to reaching the same number of people versus the ones that, that do quite well?

Yeah, I think sometimes we're surprised.

Yeah.

For example, Dr. Tara Swart, who's a neuroscientist, she is one of our best performing episodes. and I don't think anyone would have ever expected it. No, which is amazing. I love that when it is an outlier and you're like, wow, what can we learn from that?

Mhm.

Um, I think the other way around is when something doesn't go as well, and I think there's a few reasons. It could be that the topic and the guests just aren't right. And also, to preface, there's a lot that don't go out as well. There's a lot that don't even make it to going out.

What percentage of recorded episodes do you think actually get published? Is it 90%?

Yeah, I, I say 90% or maybe higher. So it's only a small percentage, but it's if this— the quality is not good enough, it needs to be really valuable. Like, there has to be so much value in that podcast. And especially Steve is amazing. He has the highest standards I've ever met in the best way possible. And he cares so much about the quality of the podcast. And I think that's why it has been such a success, is making sure that everything that goes out is somewhat valuable. In whatever topic it is to people instead of putting something out because, oh no, it's a Wednesday. We don't release on a Wednesday, but it's a Wednesday and you have to get something out. Yeah. Which I, I know is what a lot of people do because you're like, oh God, I promised to do this and I have to put something out. Whereas actually having quality episodes is what is going to keep people coming back. And the more I think about it, the more I think, gosh, if you have one chance maybe one chance to get someone to listen to an episode and they happen to click on the episode that's subpar and then they never come back.

Forever they think of the show as like, exactly. And I was going to check out Grace's content, but I actually wasn't that great, you know? And you're like, that was one.

Yeah. And it's, that's scary. So it's like, why would you put out something subpar? Because that could be the one piece of content that people see and go, and that's when they make their judgment. And even though that sounds crazy, it is probably True. Like, people do judge so fast on whether they like or dislike something. And I listened to one podcast, there was a man which I thought was kind of motivating, but he was shouting down the podcast microphone. And I was like, okay, it could be motivating. But I thought, I wonder if he does this every episode. And then he did.

Okay.

But, um, but I thought it's interesting, the dynamic of, for example, the one you did with the whiteboard. I was like, I love that. I want to go back and watch more. And consistency, I think, and quality is super important.

I would see it as a red flag on a podcast if you release 100% of your episodes. And I think that might be controversial because you have to— in order to not release 100% of your episodes, you have to offend people. And because it means like we recorded an episode and it was not good enough to release. So I—

you're going to tell me that after?

This is my nice way of saying that. Him. Delete the files. No. So there was an episode, I guess that I've, I've done a few over 100 episodes on the show and, uh, there's 2 episodes that I haven't released. I'll use one as an example. I won't like throw out the guest in particular, but, um, we were recording in a hotel room, which is always hard to do. Like got the hotel suite, set up all this gear, hired a, um, a videographer to do it and everything. And it just wasn't there was a bunch of things working against us.

Yeah.

Like the fire alarm had gone off at 4 AM that morning in the hotel. And so that was a little rough, you know?

Yeah.

And then I had made the decision I was recording 4 podcasts that day, which that was really pushing it. You know, we had all day for it, but like that was before the fire alarm at 4 AM. We had some AV issues, all of that. And I didn't bring the energy. I like could not get You know, the guests didn't want to share specific numbers, and that was hard, like, because it's something we do on the show is we talk very specific things. They didn't want to. I was like, I think that'd be fine. I couldn't draw it, you know. Yeah, all that didn't work. And so going back to the guests later and saying like, hey, we, you know, we owned it, like, this was our mistake, you know. And we said, we will fly you out to our studio. Like, we're not going to release the episode. Yeah, we would love to re-record it if you're up to it. We will fly you out, we'll pay all the expenses. Well, you know, all of this, but if we're going to do it, it needs to be at this bar for quality. And I'm so sorry. Like, it's on us and all that. And they took it well. Yeah. I don't think they were too terribly offended.

I think it's good if they take it well, though, because it's also the self-awareness of them realizing maybe they didn't perform as well.

Yeah.

And how can they— we've had some that have said, like, how can they improve, which is actually really nice. Like, that's So self-aware to be like, what can I improve on that didn't go right? But you've gotta have the high standards cuz that's how people keep coming back.

Yeah. And so I guess for, for anyone listening, I, I would ask that as a flag of like, if you've done hundreds of episodes, yeah. Is there at least one that you haven't released? Mm-hmm. And if the answer is no, then I would argue that you're not holding the right bar for quality.

Yeah. Yeah, it's true. It's so true.

Another thing that I've heard, mm-hmm, is, uh, well actually I can say specifically who it was. Yeah. So I was talking to Ramit Sethi. He recently hit 1 million subscribers on his YouTube channel for his podcast and all that. And he told me that maybe a year or two earlier, as he was maybe 100,000 subscribers, something like that, he was like, okay, I want to grow my YouTube channel specifically. And he has this amazing podcast about money for couples. And so it's like psychology and money and relationships and all this stuff together. And he doesn't care about personal finance tips. He cares about like why as a couple the husband has the weird spending habits and how they're, you know, like any of these things he's getting into it all. Right? His whole thing is he loves a Reddit post about like, my wife or my husband is doing this. And he's just like, please come on my podcast. You know, he posts all the time.

I need to listen. I've listened to his dark episodes, but I need to listen.

It's so good. But he was asking all these YouTube experts, okay, how do I grow my podcast on YouTube? Because all of his videos are putting out and they were like, make YouTube videos. Not podcasts, you'll make YouTube videos. And so what he did is he started interspersing, you know, he still has the podcast is the main thing, but he started going and making, you know, 15 to 25 minute YouTube videos around a specific topic. It might cut in clips from the podcast, but it was cut for retention and all, and it was like a classic YouTube video. And that is what made the difference in growing the show. I was worried as you were saying that, you know, I was like, wait, isn't this going to like fragment our audience on the channel and people are coming for different things? He said it worked really well. But I'm curious about that of podcasts for retention and average view duration is a very hard thing to do. So one, how do you think about that? And two, is it something that you've considered of like going and making much more YouTube-specific content rather than pure podcast?

Yeah, I, that's really interesting. I wonder if it's also because it's a specific genre. Yeah, finance. And maybe people are happier with shorter clips to quickly learn. And whereas podcasts, or some podcasts maybe, are more educational or storytelling, so you're willing to sit there for the length of a movie.

So that, I mean, that's a good point of being really cautious, not, or not, not even cautious, being really aware of who you're taking advice from and if they're in the same genre or these other things where you're like, oh, what this gaming channel, I'm going to implement that. And you're like, Nathan, you don't have a gaming YouTube channel. Like, you're as far away from that as possible. Like, you should probably, you know, you could be curious about it, but like, please don't copy and paste.

I usually have a slide in my presentations. Well, one says, are you curious? Um, but one slide says, don't trust what others say. And I say that I do trust—

Grace says, don't trust people. I don't trust anyone.

Um, I do trust what others say, but I, I think the whole point of that is take what they say and go and test it for yourself, right? So obviously he did. He went and tested it on his channel, saw if it worked. I'm sure if he tested 5 and went, oh my God, this has way worse performance, then he'd be like, great, let's pivot and stop and go back to the long-form podcast. But I think you can take it, learn, and then take, like, put your own spin on it to see how could it work for your channel, rather than just taking something going, I have to do this. There's some things that could work across everything, whether it's something analytical or a specific feature like a post. But I think for retention-based and specifically content, um, that is a style, whether it's podcast or YouTube traditional or vlogs, I think they all have different styles of retention. Like vlogs, people might play in the background most of the time, and like, I'm happy to listen to it from far away. I'm not really going to skip usually, whereas MrBeast you might watch and really intentionally watch.

Yep.

And then for a podcast, you might put it on your TV. And so, and TV is massive on YouTube now. And so we're seeing, I would actually say there's really different behaviors and retention across all the different genres on YouTube. And so taking knowledge from one into the other can be valuable, but making sure it actually applies and testing it on your own one is worthwhile first.

In your title, I'd expect it to be the head of experimentation, but you have this word in here of failure, a word that people are like, well, I don't like that. I don't want to fail in my relationship. I don't want to fail. I especially don't want to fail at work where, you know, my job might be on the line or any of those things. How do you go about first building your own relationship to failure where you're fully embracing it? You know, you and I are on stage yesterday talking about failure and showcasing all of that. Because that's what the, you know, the team here at the Billion Follower Summit, they're like, do talks around failure. And so, okay, we can do that. But how do you shape your own relationship to failure?

Well, my microphone was actually failing yesterday on stage, which was hilarious. And I was going to call out and say, this is a failure.

I was very glad that they sorted out the microphones before I came on right after you. So thank you for being the guinea pig.

That's right. But no, I think, yeah, when we're younger, we are taught to avoid failure, right? So I— as when you get into adulthood, it's really hard to be like, oh, now I just want to fail, because you don't have the experience. Like, why do we study for a job? Like, why do doctors study? It's because they need to learn. It's like, if we're not taught that failure is okay throughout school, throughout university, or in life, how do you then get comfortable with it? You don't. And so the whole mindset of, and why I'm trying to share more about failure to the world, is it's the biggest brands, celebrities, like even entrepreneurs, they have all failed. And I know you talked about that yesterday as well, but without those things and without them being comfortable with failure, they wouldn't have got there or they wouldn't have tried to get there. And my favorite one is obviously J.K. Rowling. And was turned down by 12 book publishers before they published Harry Potter, which is the best-selling book series in history. And I just think, God, how do you have the motivation to go to 12 publishers and still think this is something? And it's really the self-belief to continue going. But in work, it can be really hard because you're working for someone else. You're like, oh, I have KPIs to hit. I have to impress my manager if I want a promotion or a pay rise. And so how do I actually make sure I'm doing my best? And but for failure, it really has to be a company-wide thing to encourage it. So whether it's a culture change, whether it's management saying we want to start doing this, but it's really a mindset of saying failure is going to help us grow because the more things we can learn from what didn't work, the faster we can grow, right? Because if we're not trying anything, we're not going to learn at all. We're just going to keep doing our regular day job. If we're trying things, some things will work, some things won't. And whether that's a subject line on an email, you might test two. One might flop, one might do amazing, or maybe both will flop, who knows? But we're actually testing it to see what works, what doesn't. You're just doing one, you're never going to know. You're going to have put that one subject line out and gone, cool, could it have done better? Maybe. Or maybe it would have done worse, but we would have learned from that. And so the whole mindset that we've— or Steve has created, which has been amazing— is making sure that no matter what, when someone fails, it's— as long as they share a learning from it, it is congratulated. It is like, thank you for sharing that, because you have helped to do something new in the company or try something that's different to help us grow. Like, you are helping the company grow by failing.

Oh, that's so interesting. So the learning is the key thing. Like failure is not celebrated. Failure and learning from it is celebrated.

You can't celebrate failure alone. You have to actually add a learning to it because failing, cool. Like if I smash this glass in the water, fine. On the floor, sorry. Yeah, I could smash the glass. Yeah, I've smashed it. What do I do next time? Don't touch the table. But I think what's the learning side is super important to then say, okay, How do we do this better next time? Do we continue down that experiment and try and find the way that does work, or is that so far removed that it's just gonna keep failing and we know don't do that again, try this other way. Um, but it is, I think the best part that Steve says when someone shares a failure, he's like, thank you for actually going and trying something because that shows you care.

Right.

And like you are willing to go and try something new to help this company grow as fast as possible. And he doesn't care that it's a failure. He's like, you just wanted to try, and that is what matters most, rather than you just sitting there going, I'm not even gonna try.

There's— I've heard someone phrase it as, in life, I want to commit, you know, sins or crimes of commission, not omission. Meaning like, I never want to get in trouble or not succeed in my business or my goals or whatever else because I omitted something, because I didn't take action. Yeah, I want to do it because I took action and it ended up being the wrong action and I learned from it. Right. And so, so many people, so many people are paralyzed to try things because they're like, what if it doesn't work? And, you know, it's one of those— it's a total cliché, but it's so true of great. If you don't do it, it didn't work. You have 100% chance of it not working if you don't take action. And then so anything, any possible chance of succeeding requires taking that action.

Yeah. And I think you always regret not trying, but I don't think you'd ever regret failing and finding out the answer, right? So why would you not at least try and see what does and doesn't work? But it's also hard because you can be in the mindset of the public embarrassment of failure, and I think that's probably what stops people most. And even— I grew up in Australia, and I think definitely there, when you're surrounded by all the people you went to school with or university or your first jobs, and you're like, oh my God, all these people are going to judge me. And then when I moved to London, completely changed. Nothing had— I didn't care about public judgment. I— whether that's because no one in England really knew who I was, but I was just willing to try and fail. And even though they could maybe have seen what I post online, and judge, what am I going to do? They are the ones judging. And I loved what Vin said yesterday in his talk. He said when people say negative comments to him online, he said, "I do not want your negativity. You can keep it, and you can go and have the negativity." And I was like, that is such a great way, like a mindset to be in of if you failed or have a negative experience, you do not have to worry about public perception or take it with you. It is like, move on, learn, and we can grow.

Yeah. Yeah. It's so interesting of thinking about like, what's the worst case thing that could happen, right? You put out content and it's not good and it's like, well, you know what YouTube's going to do? They're going to see in the stats that's not good and they're not going to show it to anyone. Yeah. You know, so you're like, it's barely going to be seen.

No, but that's true. It's like, and it doesn't matter. Like, it's— but also then you can learn from it and be like, why didn't that work, right? Why did it not get the views? Was I boring? Did I have no inflection or tone or intonation? Was it a topic that wasn't popular online and like finding out why. And you know what's interesting is I actually think we learn more from that than all of the viral ones. Because sometimes you have a viral video and you'll go, what did I actually do in that that I can then replicate? You can do it most of the time, and you'll see YouTubers or content creators that will rip off a video 10 times, which is amazing. Get more views, go do that. Um, but I think you can learn a lot more from what didn't work and be like, why did the retention only last 3 seconds on a 30-second clip? Or finding out those answers.

I love Vin's example of like just rejecting someone's negativity because it's like, okay, the thing I'm most afraid of is people commenting in a negative way or that sort of thing. My go-to response is when someone says something kind of nasty is I just go, you must be fun at parties. And that usually gets a bunch of, but you know, his example of like, oh, I, I reject your negativity. Like, you can keep that.

Yeah, it's like psych—

or the other one that I've seen people use is, um, uh, I think that would have been better as an inside thought. Some people like where you just reflect it back.

I just have never— I can't even fathom not to go down the topic of comments, but I have never commented anything negative. So I just don't understand how people can comment something negative, especially because people are putting themselves out there online. And I'm like, that takes a lot of guts because they are willing to maybe fail. So to then go and comment something negative on someone's post, like, how sad. Um, that we should all be— obviously everyone's entitled to their opinion, but we should be celebrating people going and trying and doing something and trying new things. But I love that content creation in general is such a new industry and career. Like, there's so many people doing it as a career now, which is incredible.

Yeah. So I want to go back to the culture inside of Flight Story around failure because that— there's, you know, these unspoken rules or, you know, that shape a culture. And I want to know how Steve and the rest of the leadership, like, really shape that and say, okay, not only do we celebrate failure, but how we learn from it. And like, what are those rituals or those little things that if you were implementing it in another company, you would do?

We, for the public recognition side, we have an Experimenter of the Week trophy. So it's this big trophy and a voucher that, um, someone gets for sharing an experiment. Doesn't matter whether it succeeded or failed, but it's the awareness in the team that this is what we celebrate. We celebrate someone trying, and they have got this for this week because they did something. Um, so that is one of the, like, public recognition ways. The other side is Steve is— he just really cares about our team. And he'll, if he sees someone experimenting, trying, even failing, he'll drop a message and be like, thank you so much for doing that. Or I can see you're working really hard. Thank you so much for putting in all that effort. Or the quality of what they're doing. And it's so nice that he cares. And I think when you— to create a culture where people are willing to fail the people at the top, so all of our amazing C-suite as well, our CEO Georgie, they have to have relationships with the people on the team so they feel comfortable enough to say, hey, this didn't work, right? And this is why. And then they go, okay, well, what do we learn from it next time? It's not always positive. Yeah, like, it's not like we go, woo, we failed!

Yes, like, we didn't hit any of our goals.

Yeah, yeah, like, it's not always like that. Yeah. Obviously we want to hit our goals. We wouldn't be the second biggest podcast in the world if we didn't want to hit goals. And it's what we're striving towards. We're not— we don't go, oh, we are going to fail. We want to succeed, but we know there's going to be failures along the way. And that's— it's the accepting of that. And so yeah, Steve, Georgie, our amazing C-suite team are always willing to go, okay, it didn't work, what can we learn? Right. And they even say— Georgie's famous line which I love, and everyone in the team is encouraged to say, is, "I don't know." Yeah, I don't know the answer to something. And not to look stupid, it's to be genuinely honest and say, "I don't know," instead of making up something, right? Um, and so it's the whole psychological safety of the team feeling comfortable to say, "I don't know," or, "I failed and this is why," and being able to build on from that. And you're going to learn, and the learning is so much more valuable because that is what we know what does and doesn't work on YouTube, or what does and doesn't work in social media marketing.

Well, yeah, especially when you're taking it across not just one show but a bunch of shows, and you're saying like, look, this is part of what we bring to the table if, if a creator wants to launch a show with Flight Story, right? Is that you're getting all these learnings.

Yeah, and we do it with brand partners now as well. So like when we have ad reads in a podcast, we go and then test the ad read. How do you get the best retention in an ad read?

Mm-hmm.

Because most people skip an ad.

Right.

So how do we make people want to stay? And it's not, yeah, it's not just experimenting on the podcast. It's experimenting in any area possible.

Yeah. And carrying that through.

Yeah.

I'm thinking a lot about what gives status in an organization because whether we like it or not, humans love to play status games.

Yeah.

You know, usually subconsciously. And so if you think about like organization design and culture, it's like, okay, well, what gets status in the organization? Is it Is it working super long hours? Is it achieving, you know, key results and like that being really visible? And so what I hear you saying is that experimentation gets status because, you know, you have a specific award around it. Like it's pretty overt. And so thinking about that in each organization, like at Kit, we do a lot of experimentation. But we have— we don't have those rituals around really celebrating it. We more tend to celebrate the wins than the experiment itself.

Most companies do.

Yeah. And so that's something that I'm working on shifting because then it, you know, right now it's around how do I get a win, which is great. We should be trying to do that. But, you know, you're saying actually it's, it's the pure number of shots on goal or that sort of thing.

It's always the more shots you take. So if you're uploading content on socials, if you upload 1 video a week, you have 1 opportunity.

Right.

You upload 10 a week, you have 10 opportunities to go viral. So what is the better option? The 10, obviously. Obviously it takes more work, but you have way more opportunities. And so in the 10 experiments I shared, 7 are probably gonna come back average. You might get 2 failures, maybe 1 success, but that success could be exponential.

Yeah.

And the failure side of it, it's interesting. Obviously it's amazing to share successes or reward success, but I think what can be missed is actually what didn't work. And we do that across all of our shows and bring it together. It's like, is the same thing not working? Right. But also, if something didn't work, could we then make sure that it doesn't get tested everywhere else if it doesn't need to be? Like, if we know solely that that is not going to work. Don't waste your time. Like everyone else, the 10 other— if you had 10 shows, 10 other shows don't need to go and waste their time testing it if we know it doesn't work, right? So it's valuable both ways, of the successes of what does work, but also you don't need to go and test things if it doesn't work.

How do you log all of the, you know, the shared knowledge from those experiments? Yeah, because you have a growing team, uh, people get hired between other teams and companies, you know, you're not going to retain everyone one for all time. And, and as you're like, okay, we, we went from 20 people to 40 to 100 to, you know, you need to transfer that knowledge. Look, what systems do you have in place for that?

Yeah, it's currently a database of all the experiments, but the goal this year, which we're currently building, is some— a platform where people can go and input their own experiments too. So previously it's them sending them, but now they'll be able to almost like own it, and you can go in, see your experiments, see other people's, their face will be on it. So you can see, hey, Nathan's always running heaps of the experiments. What's going on here? Is he the only one experimenting in the company? Um, whereas then we'll be able to see like what teams are experimenting the most.

Okay.

Are there lots of social learnings we can then take on to another area? And also I think what can be maybe siloed in companies is different areas of social or marketing or production, whereas there's actually learnings we can take across multiple of them. And so for example, if a specific type of content worked really well on YouTube, we could cut it down short form, then put it in email, then maybe we go and put it in our, um, do a spin-off or a special segment of it for our paid community. Then maybe we could take it to a live show if it's something that could be created into a live show. So the experiments can apply to so many areas that it's not just one team, which is, yeah, I find quite interesting of it shouldn't be ever siloed, it should be the knowledge sharing.

Yeah, so I mean, the key takeaway there is you have to have a database where you are logging your experiments and that could be as simple as a Notion table where each thing gets submitted there and it's tagged by the team and the category. Like this was a copywriting experiment, this was a retention experiment. And then what you can do is now you can have AI, you know, summarize learnings from that. Or you can say, hey, I'm designing an experiment, uh, like this to solve that. Has, has anyone in the company done an experiment like that? Claude would be, or Notion AI or whatever, would be very happy to say like, well, here's the 3 most similar experiments that have been done in the company. 2 of them were done before you joined the team or whatever else. Um, but no, the exact thing you're testing has not been documented in the same place.

Yes. And, and also retesting things. So if someone's tested and it's 6 months later, you, you might go, oh, actually we need to retest that and see— especially on socials, things move so fast— does it still hold, does it not hold, um, and how long should we retest it within? And keeping all the data in there. Like, I know Booking.com has an amazing one. Obviously it's easier because they own the platform, right? So you can get proper statistical significance, whereas some of ours, as I said, are more knowledge-based experiments, um, but we're still able to put them in, learn, and then apply them to other places. But it's the knowledge sharing 100% that is so valuable.

Yeah, that makes sense. We— I touched on AI for a second. How are you using AI across Flight Story, and especially in the podcast production process, um, but, but really across the company? Because I, I know that, uh, Steve and the whole Flight Story team are pretty obsessed with learning and implementing AI?

Yeah, there's honestly— it's across every area. I don't think anyone doesn't use AI in Flight Story. Um, and we had an AI agent competition last year, which was really fun because I'd never thought about building an AI agent. It's not really my area, but stopping my work to then go and try and challenge myself to build it was super interesting. And the things you can build if you take a little bit of time to learn it when I'm not a developer, I'm not a coder, was super valuable. So I built a tool that could translate our social media clips and then upload them themselves as well. Okay, so not— I didn't code anything, like I didn't build an ElevenLabs, but it was able to use software like ElevenLabs to then go and almost replace a social media manager. And we're seeing a lot of different AI tools starting to replace certain areas of jobs, but also speed up the process.

Yeah.

So being able to help speed up production, being able to help speed up even designing thumbnails, especially if it's something that's the same thumbnail every time, but maybe swapping out a guest's face or text makes it really easy for AI to do descriptions, titles, even inputting. Say you were scripting for a social clip and inputting your voice into GPT, or your tone of voice, sorry, into GPT so it understands you, and then it's coming out straight away perfect in your tone of voice. And the whole team use AI across the board to try and help speed up, but also improve the quality of what we're doing.

Well, it's a great example from what we were talking earlier about dubbing the clips or the episodes. And you're saying, well, that would be nice to be able to do, but I don't have a dedicated team member that I could put on that. You're like, well, you can build an agent that will take the file that was uploaded here and go and use the transcript, create this, and You know, use the off-the-shelf tools and run that as an automation so that, you know, now we're just into quality control and you can reach this new audience without actually having to, you know, do 2 hours or 4 hours of work per episode on that.

Yeah. And it can be basic things. It could be like you want to stay across all of the new features on social media channels and you set up a GPT or an agent to do that for you.

And you keep your— it's like giving you a digest of this is what's changed in Instagram or— yeah.

And it's super basic. But it's like, then you don't have to have someone— because previously for the last 4 years, we've had channels called, um, Industry News, Staying Ahead, where we share— where we go and share them. And I'm like, why am I still doing it? Personally, I do like it because it teaches you something of going to look for—

and you see what's interesting, what other team members found interesting.

Yeah, yeah. So I see the value in the human side still, but I actually think to not miss something, it can be really valuable to make sure that no one has missed a specific major update. And so that's a tiny simple one of being sent news updates, but could speed up the process of needing to take 10 minutes out of your day to go and find those things.

Right. Were there any other agents from the agent challenge that stood out that you were like, ooh, this was particularly excellent?

Ooh, so many. There was some production ones. One of them that wasn't in the AI agent competition, but has been interesting to see progress is We'd be testing AI podcasts for a little while. So we tested a version with Steve's voice, um, and it was— it's still progressing, the test, to be honest, because it's evolved into something else that's launching soon. Um, but even testing like the design and getting imagery and things, and it's interesting how far it can be pushed.

Okay.

With AI. And I know there was a content creator, um, at the summit who has now got quite a few channels of AI avatars that he's got millions of subscribers on, of these avatars on YouTube. And I just thought it's interesting, that use of AI. Obviously AI can also help us speed up our work or improve our work, but still the AI tech side of the production is, yeah, interesting to see where that future will go. And will people still be interested in adopting it and watching content like that, or do they want the human people?

Right. I mean, do you have a perspective on where the creator economy goes in this world of AI?

I think it will go down the AI route, definitely of translations. I would be intrigued to see. I think there will definitely be AI avatars or AI podcasts. How quickly they get adopted, I'm not sure, purely because I think people are still now converting to wanting humans and human opinions. And even there's the whole debate online right now of is social media too perfect now with AI? And people are now putting in spelling errors on purpose to then make it seem more human. I'm like, gosh, how sad that we were trying to put spelling errors to make it seem like actually it's me. Whereas most of my content I write myself. I don't really even use ChatGPT to come up with ideas. I'm like, this has come into my head, I'm just going to write it and post it on LinkedIn. But I know obviously it's amazing.

Because the people who use AI for their LinkedIn posts in the most basic way, you're like, I, I can see through it a mile away and I'm just not— I'm going to scroll right past it.

Yeah. And it's like that. I do still like the human side, and I think, and I think most people will be on that for a little longer. What do you think though? Where do you think it will go?

I mean, it's obviously going to augment every single workflow, allow people to build like substantial reach with small teams. Um, you know, I saw a creator, her name's, um, Allie, and she does a lot of painting and art And she ran a whole series of, um, of ads, like, so short-form reels and all that run as, as ads on Instagram and Facebook. And it's like, hey, this is Allie's AI clone. And, you know, and she's able to do these different things and, you know, now she's on a boat painting and then she's over here and, you know, it looks like her, but it's clearly AI and she names that it's, you know, it's her AI clone. And she did it in an interesting way, and it was about retention. And, and then I got one of her other ads, and it was definitely her, you know, an actual video of her. Um, and so the, the dynamic, like, going back and forth— she stopped the scroll by saying, yeah, you know, and she kept retention up with— in these different— with her AI avatar. So that's all really interesting. I also see creators, especially on YouTube, there's a wave of people who are— it's kind of like putting the spelling errors in, but going to much more analog formats, um, like creators who are shooting on camcorders where you're like, wait, you're going to 1997 technology? I know, you know. Yeah, because— and you're, you're recording on actual, you know, the film cassettes or whatever. And, and because they're saying, oh, we're going the opposite direction. Yeah. Um, and so it's less about polish and it's back to storytelling and human connection.

Yeah.

Um, so I'm obviously making, you know, similar to you, making a huge investment in the creator economy and, and all of that. I don't think it'll go anywhere. Um, but I think that people will continue to crave human connection and human stories.

Yeah.

Um, over the like polished, perfect.

Definitely. And I think as well, what will then happen with events like this?

Right.

Like, where people who were so excited to see MrBeast or other creators, is then like, if they're an AI avatar, who are you excited to see? Like, who do you run up to? Where's the human side of being excited to see your favorite creator in real life?

This event has been fascinating because if you just, you know, you hang out in the hotel lobby where all the speakers are or whatever else, and and you just meet tons and tons of people. And right, and we're, you know, flying. My flight was almost 15 hours, right, to get here to Dubai. And it's so worth it. And not just because you can see, you know, MrBeast walk by and go, oh, he's taller than I thought he was, you know, or connect with all these people. But it's about the human connection. And I think that's going to continue to ramp up.

Yeah, it's spot on. I agree.

One other question that I want to ask, and this is entirely a selfish question, and so I typically save it towards the end. This is the, um, a question about getting the right audience. So something that I do for my show is I care not about views, I care about, uh, attention from the right people. And so, uh, with our podcast team, they'll often say, hey, do an episode for beginners and title it like this and all of that. I'm like, I'm never gonna do an episode for beginners. Mm-hmm. Because there's tons and tons of content out there.

Yeah.

And so on one hand we're like, hey, we wanna reach more people, we want more views. But I'm saying I only want to reach the creators who have, you know, who are making hundreds of thousands of dollars. Mm-hmm. I care about the right people, not the most people.

Yeah.

And that goes against most advice on YouTube, or like if someone's like, okay, here's how to get more views for your podcast.

Yeah.

How do you think about these, these shows or channels that have a very targeted audience that they're after and like seeking out the right views versus as many as possible?

I am actually in that camp. I am in the idea of creating a community that are super highly engaged because that will grow, and there's usually more people like that community. So even if it was only 1,000 people at the start, or even 100 people that are really highly engaged and love your content and are subscribed and do everything you say because they take it as biblical, they are going to know other people like them and they're going to say, hey, you need to come watch this, right? Or I really liked this, and share it to their community online or on their Instagram page, and people like them will see it. So I actually think even though people want a viral video especially on short form. I think viral and short form can absolutely mean nothing in a lot— like, the context of your career, because you can go viral on a video every week and it might not actually help progress anything. Whereas when you look at long form, it's actually where you're building that community. People are willing to sit here and listen to you for an hour. They actually are interested in you, and like, you provide really good information, I want to keep coming back. And that I think is better quality than something like a challenge video where people might watch for 10 minutes, have an affiliation or know your face, but not actually be willing to say you did a course, come on your course, or buy your tech if you were selling, or whatever, like products. And that real community side, I think that is so much more valuable long term than a view or a viral video could ever get. And having a niche specifically, because I know what I'm coming to you for. Like what we said before, that podcast is known for X, right? Whereas you could be like, oh, the podcast where they talk about this, this, and this, and this, and I don't even remember what the name is. Whereas it's like, I know I'm going to that creator for this. Like, yeah, Colin and Samir, you know what you go to their content.

It's very focused.

Yeah, yeah. Which is amazing. And I'm like, I actually thought it was cooler to see Samir than MrBeast. Yeah, I obviously love what MrBeast does, but I was Samir. I was like, that's so cool, he's, he's so amazing. But it's obviously in the industry, but I think being in that industry like can be so much more valuable to the people there, and it will grow. Um, and it's a long game anyway, no matter what industry you're in, whether it's doing MrBeast-esque videos or Colin and Samir. Um, it's still gonna take a while to grow, no matter whether it's growing to 100,000 or 100 million.

Yeah, that was, I had the same experience as you did. Um, so at the speaker's dinner last night, right, there was this, I mean, I say speaker's dinner, there's like, there's hundreds of speakers at this event. So it's wild who they've flown in and all of that. But there was this group and it was, uh, MrBeast and the Yes Theory folks. And, you know, I was looking at this like, uh, group of maybe 12 people or whatever. And you're like, oh, there's hundreds of millions of YouTube subscribers represented by this this little circle. And, but I had the exact same thought, right? Like Samir and Jimmy were sitting next to each other and I was like, ooh, I really want to say hi to Samir. You know? And I know, like, we've talked on the phone and we met before and all that. But it was one of those things where I waited until he got up. Cause I was like, there's no way I'm going to come and interrupt. Also, like Jimmy has his security standing around and all of that. But once Samir left that conversation, then I like made sure to seek him out.

Yeah, and I think it's cool. Like, I just thought, well, and I think that's nice, is creators that are— that have— but I think it's dedication. Yeah, it's like dedicated, really niche, interesting topics to talk about, rather than— I don't even know if I met MrBeast, I don't know what I'd talk to him about. Maybe YouTube, but it's not as specific as like the conversation you could have with Samir. Yeah, it'd be so interesting. So yeah, I agree, it's so funny.

Okay, so the last category of experiments that I want to know about is the experiments that you're running personally. I see you building your personal brand, posting a lot. Like, what's something that you're experimenting with, either that you've recently learned or the experiments that you have out there that you're still waiting to see the results from?

Yeah, I had previously for the last few years only focused on LinkedIn, and now I thought actually I need to diversify, and I'm experimenting with Instagram now. So pushing out the boat a bit, doing more video content. But also, there's so much to experiment with on Reels, even on carousels alone. So how do I make sure the retention stays up? What sort of hook do I add? Is the font— because there's so many videos with fonts out there that you can't even read— how do you make sure the font's engaging? What's the ideal length? So there's honestly, it's a challenge. And especially while having a full-time job and trying to do content creation on the side, it's um, a challenge, but the best way possible, because not only am I learning from creating content, I'm then able to take those learnings and experiments and apply them back to my work at Flight Story, which is really interesting to have kind of a whole full flywheel of testing, learning, and then applying it back to my full-time work.

I like that. What would you say is the, you know, the 80/20 or the 90/10 of the most effective things for you as a creator building your personal brand when you have like such a, a narrow window of time.

Yeah.

Because I, I don't think that, uh, Flight Story is a, a 4-hour work week, you know, type. The vibes that I get is everyone is very committed, working very, very hard.

We have a call at 11:00 PM tonight on a Saturday, so that's not often, but it's the times I've been here. Um, but yeah, it's— I, when I come up with ideas, I just write them in my notes and put them away for later. Because if I'm on the train or something, I'm obviously not filming it then. And a lot of it is done on my weekend. So a lot of it is trying to fit as much in, in the evenings or the weekends. I have a train journey to and from work, so I also try and do a bit of editing then, but also realizing when is the time to delegate and pass things off and say, actually, I need someone else to edit this for me or do it. But I really want to challenge myself right now to do it, to learn more.

Yeah.

About what works, what doesn't on socials, why. And it's a good challenge. Like, I want to try and see how do you get a video that's super engaging, or how do I increase shares on a video that people go, actually, I want to share that with a friend, that was really good, or that really resonated. Um, and learning and putting— I— it's about putting the time into it though, because that's when it really happens. Is it's not going to happen if I post once a month. The growth will happen when it's once a day.

Right. What's a, a metric that you're targeting in your personal brand right now?

I— not really a specific number of followers or anything. It's more how often can I post. I would love to get to posting once a day, specifically on Instagram, and I just need to get the time to film more content. And I have videos edit, uh, recorded. I need to edit them. So it's that time and dedicating the time to put into it. But it's, yeah, output right now rather than any specific number.

Well, I think it's such an important thing of— well, and the takeaway that I want people to have from what you just said is focusing on lead indicators versus lagging indicators, right? And you said what I can control is the effort that I put in.

Yeah.

You know, okay, we're posting once a day. I can control that. I can't control how many views something gets.

No.

I can have the inputs and the learnings, and if I keep iterating, then presumably given enough time, we'll get to the output.

And it's the taking more shots. Like, the more— even though I know there's the whole debate of quality versus quantity, and I agree, I still think even if you're going for quantity, you still need a high quality video, right? If I'm choosing to post every day. But I still think the more shots I'm taking, maybe 1 in 7 will do okay. The other 6 might fail. And I'm sure you'll— people will look at my channel and be like, oh God, she's failed so many times. But it's the taking the chance to put it out there, seeing what works and what doesn't, and what can I learn to grow from that.

And then I think there's also something of being hands-on yourself, of seeing exactly— like, now, because I've studied a little bit more and starting to learn Instagram more, I go through and I realize, oh, this video is designed to get comments, or this video is designed to get shares. And you can see the little things that if I wasn't like really trying to study it myself, I wouldn't have picked up on. If I was just like, oh yeah, hire a team and they'll— and delegate, we wouldn't have picked up on any of that.

Yeah, I love, I love the whole strategies, and I think love LinkedIn, but what you can do on Instagram and what you see other content creators doing, styles or tiny little hacks, right? Um, it's interesting of how it can develop and grow.

Yeah, that sounds good. Grace, this has been amazing. Where should people go to follow you as you build your personal brand and, and learn from your experiments and your failures and all of that.

My Instagram is @graceexperiments, and then my LinkedIn is just my name, Grace Miller. And then obviously watch all of the Flight Story shows. I don't have a podcast, but I work on so many amazing podcasts. And obviously this podcast they should listen to. So go for a second time and listen to it again.

Sounds good. Thank you so much for coming on.

No, thank you for having me.

If you like this episode, go check out episode 103 with Nicole Burke. She shares the simple short-form systems that helped her go from zero to $6 million in annual revenue with Gardenary. Be sure to like the video and subscribe, and I'll see you next week.