YT YouTube2Text

YouTube2Text → Transcripts → Kallaway’s EXACT System For Going Viral (Over and Over)

Kallaway’s EXACT System For Going Viral (Over and Over) — Transcript

by Jay Clouse · 8,838 words · 1,204 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00I'm trying to derisk my effort slashin
0:02increase my hit rate. So I was like, how
0:03do I use data but not make myself a
0:06clone that's copy pasting everyone
0:08else's stuff. And that's basically why I
0:09built Sand Castle's movie. That's
0:11Callaway. He started four businesses
0:13that were all killed by the same
0:16problem. No one cared. He had no
0:18audience, which meant he had no
0:19customers. So he made it his mission to
0:22crack audience growth. and now he has
0:24over a billion views, a 50% viral hit
0:27rate, and he created a tool called Sand
0:29Castles that helps other creators do the
0:32same. In this episode, you'll learn the
0:34exact system he uses to find and
0:36validate viral video topics before he
0:38ever hits record.
0:39>> My framework for content is I call it
0:41the modern content stack.
0:42>> The metrics you need to obsess over so
0:44that your content grows your business.
0:46>> Views are not the right metric to study.
0:49Use our A metric to study and I would
0:51say the third most important when I look
0:53at the stack.
0:54>> And if you stick around long enough,
0:55Callaway will show you exactly how you
0:57can build your own content research
0:58system from scratch. Thank you to one of
1:0110 for sponsoring this video.
1:03>> There's kind of a fallacy between like
1:05views equals money. It's really like
1:07views does not equal money. You want
1:09density around one ICP as many times as
1:13you can possibly get it in a given week.
1:15So like my framework on this is called
1:16content minutes. So essentially like you
1:18could imagine your offer ladder or
1:20whatever your offer stack at certain
1:22price points. So I don't know what
1:24you're like running right now, but I
1:25know you have the lab which is probably
1:26one of your higher ticket things. And
1:27then as you go down you've got like
1:29course bundles, individual courses.
1:31Maybe you have something at $47. So
1:33essentially like imagine you took that
1:35ladder from low to high and you overlaid
1:37an amount of minutes that someone would
1:39have to watch your content and trust you
1:41for those minutes. That's like pretty
1:43critical. And at the $47, maybe it's
1:45like 20 minutes worth of content. And at
1:4720 minutes, they'd trust you enough to
1:49buy it if it fit their problem. And then
1:50one one rung upg
1:54like to get into the lab, they probably
1:56would have to have watched hundreds of
1:57minutes. But like once they cross that
1:59threshold, then they'll convert. So if
2:01you think about it, most people are
2:02like, "Well, I should just do YouTube."
2:04Because every YouTube video, if it's 20,
2:06and yours or podcasts, like 60 minutes
2:08long, the average watch time, let's say,
2:09is 20 to 30 minutes. You're getting a
2:11lot of those blocks in one shot. And so
2:13you can scale someone up that ladder of
2:15trust pretty quick, which is why you've
2:17had so much success with YouTube, right?
2:19Because somebody watches five episodes,
2:20that might happen in a single night.
2:22Boom, you hit 200 minutes, and then
2:24they'll they'll be willing to buy
2:25anything. My framework for content is I
2:27call it the modern content stack, which
2:29is a basically a triangle. You've got
2:30short form, long form, both going into
2:33email. That's like the golden goose. If
2:35you if you could do anything, it would
2:36be one short form platform. I think
2:38Instagram is the most valuable or
2:39LinkedIn depending on the audience. for
2:41you is probably Instagram, YouTube for
2:43long form, both into email and you you
2:45pretty much can't lose. Now, from a
2:47content man perspective, you may think
2:49if I had to make a tradeoff of YouTube
2:51verse short form, you're going to want
2:52YouTube. And I agree with that, right?
2:54The reason most people want to start
2:55with short form is because YouTube's so
2:57difficult, like so many moving parts,
2:58you've already cracked it. However, once
3:00you're getting what you're getting from
3:02YouTube and you kind of have a formula,
3:03that's when you look to expand the
3:05universe into short form to kind of get
3:07more uh breadth of the minutes to bring
3:10it all back to what I was starting with
3:11in terms of the narrow ICP. The minutes
3:14don't acrue unless the person who
3:16watched the video trust you and actually
3:18have the problem you're talking about.
3:20So if you broadened or you know made
3:23more generic the topic you're talking
3:25about yeah maybe it gets 100,000 views
3:27instead of 20 but actually none of those
3:29100,000 blocks hit because no trust was
3:31acred because the topic you had to talk
3:33about was so generic and like surface
3:35level that it didn't actually acrue
3:36trust. So for me it's like what I try to
3:38get people to do is very narrowly define
3:41your ICP. I like to think of it as an
3:44audience of one. So like if there was
3:45just one person in a chair in front of
3:46you that's like probably somebody in the
3:48lab. You can just picture them right
3:49now. It's like if I could give value on
3:51every single piece of content to that
3:53person, there's a lot more people like
3:55that and it would scale accordingly. The
3:57algorithms in the last like one to two
3:59years have gotten so good at targeting
4:02the exact type of person you want that
4:04you don't actually need more than like
4:0510 to 20,000 views per post on Instagram
4:07other than for ego reasons. You can
4:09drive a ton of email lead genen from
4:11that. So I would say like the first
4:13thing is narrow is good. You can there's
4:16things we can talk about in terms of how
4:17to expand the saturation within that
4:20narrow slice like the way you frame the
4:22hooks and the topics but like you want
4:24narrow. If our ICP are other creators
4:27who are educational in nature for their
4:30audiences,
4:32my perspective has been that education
4:34has been harder to do well on uh
4:38especially Instagram and Tik Tok. But
4:40that might be a limiting belief too when
4:42you see people who are doing education
4:44in short form video and doing it well.
4:47What is true about those folks that
4:49they're being successful in a world
4:52where there's a lot of just like great
4:54pure entertainment that is probably more
4:57engaging?
4:58>> Yeah. How do you define doing it well?
5:00Are you talking about views?
5:02>> That's a good question. I would say
5:03doing it well in terms of uh they are
5:08making Instagram worthwhile for them.
5:10[laughter] Guess is like the ultimate
5:11answer. I would say that probably in the
5:13back of my mind when I asked that
5:14question was like okay yes they're
5:16they're doing good views. They're doing
5:18views on par with people who aren't
5:20doing education. You know the target
5:22market for this is actually fairly
5:24large. So if if views are in the
5:29singledigit thousands, it feels like
5:31there's just a lot being left on the
5:32table. So views as a metric doesn't
5:35matter in an absolute sense, but it
5:37feels like we are really barely
5:39scratching the surface
5:40>> 100%. So I would say views are not the
5:43right metric to study. Views are a
5:45metric to study and I would say the
5:47third most important when I look at the
5:49stack. I also think not all views are
5:51created equal, right? So an
5:53entertainment video that has a million
5:55views is not getting the trust
5:56accumulation the same way an education
5:58video it's 40,000 views. It's super
6:00dense. So that that does make it hard.
6:02The hard part is this decoupling from
6:04like the thing you can see from other
6:06people which is views and your own like
6:10all of our brains are like see number
6:12num want number bigger compare number
6:14right and so it can cause a lot of
6:16people to stop or to go down the wrong
6:18path when they're anchoring views as the
6:20number one metric. And here here's what
6:21I mean by that. If you know professional
6:23creators are the number one target
6:25audience and you know all you have to do
6:27is get them to trust you and they will
6:28ascend down the funnel because you have
6:29all that dialed. You just got to get
6:30their email and you know that you can
6:32monetize them. If you know that's true
6:34and you know that education content is
6:36the best way to get trust at scale.
6:37Let's just hold those things constant.
6:39Then you studying an entertainment video
6:42from someone else is actually
6:44distracting you from like the right
6:46benchmark. Right? the right benchmark is
6:48am I making a a high trust high value
6:51video educationally to get one more
6:55content minute on the stack like in a
6:57vacuum right looking me looking at Logan
6:59Paul does nothing for me right me
7:00looking at I'm just thinking of like
7:02random entertainment names and so the
7:04number one thing I think to to go
7:05through is what what are the best
7:07metrics to track and maybe I could just
7:08share like how I'm approaching this on
7:10my channels cuz I basically solve short
7:13form in like a pseudo entertainment
7:15niche you could say like I was doing
7:16tech, which sounds educational, but
7:18really it's like entertainment. And I
7:20was doing that for a brand deal
7:21strategy. So like that was a game to
7:23play. The game you're playing and the
7:24game I'm now playing on my Callway
7:26marketing, completely different game.
7:27And so even me, like I can't really look
7:30at tech as a proxy for marketing because
7:32it's completely different. So maybe I
7:33can explain like what I'm doing on
7:35Callaway Marketing and how I'm looking
7:36at it.
7:37>> The three metrics that I think matter
7:38most, number one is conversions per
7:41video. And conversions measured as
7:43emails collected per video. That means
7:46every video I have a CTA. Every CTA is
7:48driving to a lead magnet where I can
7:50capture emails and I'm measuring all of
7:52those in many chat and I can go through
7:53like this batch process that I run.
7:55That's number one. Number two is
7:57followers gained per video. That is more
8:00valuable than views because a follow
8:02when someone goes to follow it's like
8:03five to 10x more powerful because
8:05they're like that was so good I can't
8:06miss it again. So before I looked at
8:08total views I would look at followers
8:10gained and just say which one is driving
8:12more followers. Last I would look at
8:14total views. Now, the reason why total
8:16views is helpful when you're looking at
8:17competition is because you don't you
8:19don't have insight really into how many
8:20email subscribers their videos are
8:22driving or how many followers gain.
8:23Those are like private metrics behind
8:25the wall. But you could look at someone
8:26else's and see views. So, if you're
8:28looking at like what's working for
8:30others, you have to settle for views for
8:32others cuz you can't see the other two.
8:34Something I admire about Callaway is his
8:36willingness to invest in experts to help
8:38him skip the learning curve. That's what
8:40he did when he hired the one of 10 team
8:42to help with his YouTube channel. You
8:43probably know the one of 10 team as the
8:45leading tool for researching viral
8:47ideas, titles, and thumbnails. It's what
8:49I use the most on my channel. The reason
8:51the tool works so well is that the team
8:53is full of YouTube strategists who
8:55really know their stuff. What you may
8:57not know is that they also have a
8:59consulting arm that's not only worked
9:00with Callaway, but Olly Abdal, Will
9:02Tennyson, Mike Shake, and a bunch of
9:04others, too. Collectively, they've
9:06helped generate over 2 billion views for
9:08their clients. The one of 10 team helped
9:10Callaway grow a new channel from zero to
9:12over 250,000 subscribers in just one
9:16year. They helped with this video, this
9:18video, and this video, too. One video
9:21alone brought in over 2 million views
9:23and generated more than $100,000 in
9:25revenue. If you want to work with the
9:26same team that took Callaway from zero
9:28subscribers to where he is now, that's
9:30the one of 10 strategy team. They are
9:32very selective with who they work with
9:34and only take on about five new clients
9:36per month. They focus on the things that
9:37actually move the needle, ideas, titles,
9:41thumbnails, and positioning. You get a
9:43dedicated Slack channel and regular
9:45one-on-one calls to communicate with
9:47their team to ensure each video is set
9:49up for success. This is not a course,
9:51and it's not a one-time solution. It's
9:53[music] direct ongoing support from a
9:55real YouTube expert who has done this
9:57for some of YouTube's largest creators.
9:59[music]
9:59book a free discovery call with the one
10:01of 10 team today with the link in the
10:03description and let them know that Jay
10:05sent you. But when you look at your own
10:07channel, the biggest trap I find is only
10:10relying on views from your own channel
10:12because obviously if the TAM is bigger
10:14on a topic, you're going to get more
10:15views. And so if you expand the TAM
10:17beyond the trust zone, those views are
10:20actually empty calories. And so that's
10:22like just one anchoring methodology that
10:24I think about is can we have enough lead
10:27magnets in the ecosystem that we can
10:29drive to a lead magnet on every video
10:30and then evaluate emails collected per
10:32video. So I the Callaway marketing
10:34channel I've been running for I think
10:36I'm 35 videos in and the goal is to get
10:40to 2x per day. We're doing one per day,
10:41seven per week right now. Every single
10:43one of them I have a CTA. Every single
10:44one of them I'm measuring. And what's
10:46funny is not always are the highest view
10:49ones driving the most emails, which
10:51which is like you would you would
10:52probably accept that to be true and same
10:54as me, but the average person would be
10:55like, "Wa, I'd expect views equal
10:57emails." It's not the case, right? It's
10:59all about what's acrewing trust, what's
11:01what's targeting a painoint people
11:02really have a need for, and then you
11:04have a lead magnet complimenting.
11:05>> So, when you're doing these CTAs with
11:07these lead magnets, are you reverse
11:09engineering a video topic based on a
11:10lead magnet you made? Do you have an
11:12idea for a video you love and then you
11:13create a lead magnet for that and now
11:14that's just in the bank and maybe we'll
11:16use it again when it makes sense in the
11:17future. What's your order of operations?
11:19>> Yeah, so I've done it both ways. So I
11:21have an academy where I like teach my
11:23whole system to people and what I've
11:24done is I've taken pieces of the
11:26academy, extracted them out and turned
11:28it into a lead magnet. So then I'm just
11:30like I have that lead magnet made. If
11:32the topic suits it, I'll pull from it.
11:34So I guess like step one is you have a
11:37couple that you believe in. Right now,
11:39when if I were going to like what videos
11:40do you make, I wouldn't make them based
11:42on the lead magnet. I would make them
11:43based on what I can talk about next,
11:45which is like how you actually identify
11:47which topics are working. I would say
11:49it's probably maybe we can get into this
11:50now in terms of like topics because one
11:51of the questions you asked is like
11:52what's the most important thing to to to
11:55look for to to focus on? By far and
11:57away, it's the idea of the video. Like
11:59you can use almost any hook, any
12:00storytelling structure, any editing, any
12:02format. If you have a bad topic, and I
12:05can talk about what bad and good looks
12:06like, the video will fail. it'll it'll
12:08be medium at best, but if you have a
12:10great topic, you can plug in almost any
12:12of those other things and it will almost
12:13always work. So all like the 8020 9010
12:16is can I pick topics that are validated,
12:20relevant to pain points they want and
12:21that I have a non-obvious contrarian
12:23take about. So for me like if I had a
12:26pie chart of where topics come from, I
12:28would say like at the beginning 70 to
12:3180% comes from data for what other
12:34creators topically have talked about
12:36that have worked. and I'm looking at
12:37views because that's all I have access
12:38to. And so I'm informing my first couple
12:40batches of content based on that 20%
12:43comes one out of every five comes from
12:44like an original idea where I'm just
12:45like let's just test this. That first
12:47couple batches and I do batches of 10.
12:50So let's say like 20 to 30 videos is all
12:52about testing. I'm basically just trying
12:54to validate which topics work best for
12:56me. Once I then know that we we could
12:58talk about like downstream, but you know
13:00topics are the most important thing. And
13:02what I'll do is based on the topic that
13:04I choose, I'll then plug in whatever the
13:07closest lead magnet would be to that
13:09topic. The way I think of it is I don't
13:11really want to spend every hour making
13:12content. Like I'd rather live my life,
13:14right? And so I'm trying to like
13:16increase I'm trying to derisk my effort
13:18slash increase my hit rate. So every
13:20time I sit down, I want it to bang so
13:22that I get the result so I don't have to
13:24spend more hours doing it. That's kind
13:25of like how I think of it. So it's the
13:27measure twice, cut once. Like what would
13:29be the way to derisk? The way to derisk
13:31is to use data on the topics and then
13:34creativity on the contrarian insights.
13:36That combo is like that's what all the
13:38goat creators are doing. I mean on I
13:40would say on YouTube and that's what I
13:42do on YouTube too on YouTube and short
13:43form. Uh and it makes it so much easier
13:45and that that's basically why I built
13:46sand castles, right? So like for anyone
13:48who knows Sandcastle.ai that's like my
13:50software tool AI tool that does a lot of
13:52this for me. The reason why I built that
13:54is because I went data first from the
13:57beginning because I really did not enjoy
13:58the process of spending 10 hours on a
14:00video and it failing. I like hated that
14:02a lot. So I was which I'm sure everyone
14:03does obviously but I was like I can't do
14:05this forever. Like I can't do the
14:07guessing. So I was like how do I use
14:09data but not make myself a clone that's
14:12copy pasting everyone else's stuff
14:13because I also think that's where you
14:15lose too. If you're a derivative you're
14:16going to lose. So you have to be
14:18original but you have to fish in the
14:20right ponds. That's that's the key
14:21thing. If you're fishing in every pond,
14:23some ponds don't have fish. So, it's
14:24like, how do we figure out where to fish
14:26and then use our own technique, our own
14:28contrarian insight, our own, you know,
14:30whatever the metaphor, our own boat to
14:32go get more fish, right? So, the way I
14:34do it with sand castles, which I think
14:35is the best way right now is,
14:37>> can you show us this?
14:38>> So, this is sand castles in the upper
14:40left, you can build like separate
14:41workspaces for different niches. So,
14:42like you might want to say, okay,
14:44professional creators is like my main
14:46one, but you might have other like
14:47verticals you want to test if you want
14:48to if you want to partition them
14:49separately, right? So, this is my
14:51marketing workspace. Basically, the
14:53first thing you want to do is build this
14:54channel list. So, if you go to the
14:55channels tab, I've got this list here of
14:57like all the top creators that I've come
15:00across or that I know of in my category.
15:03And like you, like I said, you can
15:04either build this one time or like over
15:06time. But how you would do it is maybe
15:09let's just delete all of these so I can
15:11show you uh kind of the process from
15:13scratch. So, typically you would come in
15:15here and you'd be like, I don't know who
15:16to add at all. So you go to describe and
15:18let's just say you're looking for like
15:19for me social media marketing and growth
15:22content and I can say I only want people
15:25from Instagram or I only want people
15:27from micro and small cat which is like
15:2910k followers to 250k. Now I'll go up to
15:32medium. The reason you want this micro
15:34to medium range is because typically,
15:37and this is like not intuitive, but if
15:39you track videos from massive people
15:41like Horosi, let's say, who has like
15:42four or five million followers,
15:44typically when he has an outlier, it's
15:46not necessarily just because the idea
15:48did so well. It's because he's so famous
15:50and has so much uh built-in like brand
15:53equity that people will like anything he
15:55does. So, it's actually a counter signal
15:56to study the huge people. You just won't
15:58find the best ideas. Conversely, let's
16:00say you're not wellknown at all. If you
16:02have a banger, that means the idea, the
16:04concept, the substance was the reason
16:06why. It's not because of your face
16:07because nobody knows you, right? So, you
16:09want to study people that are slightly
16:11smaller in the that's where really the
16:12golden ideas come from. So, let's just
16:14do like these three and we'll press
16:16search. So, what Sand Castle does is it
16:18breaks whatever you gave it into three
16:20different buckets. So, in this case,
16:21social media marketing, social media
16:22growth, marketing and growth strategies,
16:24like they're all kind of derivatives of
16:25this. And this will give you the best
16:27ranked channels that we have on our in
16:30our database. We have like 100,000
16:31channels or so. And if you want to see
16:33more, you can just load more. And the
16:35goal is you just highlight that or you
16:37you you hover over these. And you if you
16:38want to investigate these, you can just
16:40click and open it in a new tab. Boom.
16:42This person's in a new tab. I don't know
16:43if it's sh it's probably not sharing,
16:44but like I open it. Boom. I see her
16:46channel. Here it is. And what you want
16:48to do is build out your list. So I could
16:50say like social with Philip. Boom. Press
16:52the plus and I build him out here. Now
16:54you can do this quickly by just pressing
16:56add all. And what we do in the
16:57onboarding is like automatically build a
16:59list of 20 for you. But I find like most
17:02people want to tailor this pretty
17:03tightly. Now the cool the cool thing
17:05about a few like hidden features we have
17:07here is like let's say so uh Jaden is
17:09really dope. Like he's dope in the
17:11space. I've seen his stuff. So like I'm
17:12I'm going to add him. Now what's cool is
17:14I can hit this little double arrows
17:16here. And what we'll do is make him the
17:18anchor. And now these are all the best
17:20fits based on like closest to him. So if
17:23you let's say you're not good at
17:24describing what you want to do, but you
17:26f you know there's one anchor person
17:27that you love, you can just go to him,
17:29press this, boom, now this list is like
17:31the closest possible thing to him. So I
17:33would just like add all these people in.
17:35Right? So that's how that's one way to
17:37do it. The other way is if you know
17:38someone's good, like this girl Ava is
17:40really good in the space. I'll just
17:41search her name, personal brand launch,
17:43and if you just wait a second, it will
17:45find her. So she's right here. Now, like
17:48I said, we only have 100,000 channels.
17:49like these are top curated channels and
17:51our users every time a user adds one it
17:53gets added to the database. But if
17:54there's somebody that exists that you
17:55know should be there and they're not, it
17:58just means nobody has ever added them to
18:00Sandcastle. So you can actually do it
18:01for free. You just go to add URL and you
18:03can type it in. So we have this. I'll go
18:04to parse. If we have you, the picture
18:06shows up. Boom. I add you. No problem.
18:08If we didn't have you, it would be like
18:10new channel and you can import it. And
18:12then when you import it, we'll take in
18:13your full backlog, right? And so th this
18:16process of building out your list is
18:17super valuable because now you can focus
18:20your attention, you can focus your
18:22research attention on these channels
18:23which you know to be good. So what I
18:25recommend for people is like almost
18:27everybody knows at least one name of
18:29someone that's pretty pretty good in
18:30their niche. Start with that person,
18:32search them or add them and then just go
18:34to this kind of refactor. And then once
18:36you hit this refactor, you can just be
18:37like, "Okay, boom. These are all the
18:39people that are really close to this
18:40person." So we'll add a bunch in there.
18:43One thing that's interesting about this
18:45approach that's a little bit counter to
18:46like the YouTube advice in a similar
18:49fashion is a lot of times the advice on
18:50YouTube is like get outside your niche.
18:52Find somebody who's doing something in
18:53finance or gaming that's interesting and
18:55try to pull from their video ideas which
18:59I think makes sense but I don't know it
19:01doesn't seem like it would translate
19:03exactly to Instagram but I having a hard
19:05time articulating why that's true.
19:07>> Yeah. So I think there's there's like
19:08different buckets right. So when you do
19:09the YouTube research, if you're looking
19:11for topic inspiration or interest topics
19:13on YouTube, you need to have people that
19:15are in your space or like very close
19:17adjacent. But if you're looking for
19:19title formats or thumbnail inspiration
19:22or series ideas, you can go broader,
19:24right? For example, if you were doing
19:25creator stuff, you typically wouldn't
19:27look in the ecom niche for like interest
19:29topics, but you could find a title
19:30format that's crushing and then
19:32translate it over. It's kind of similar
19:33here where if you're looking for topics,
19:36you need the topics to be within niche.
19:38But if you want to look for formats or
19:39you want to look at editing styles or
19:41you want to look at hook hook, you know,
19:43hook formats, you could do that outside
19:45the niche as well. So what we found is
19:47like most people have trouble building a
19:50list outside of their core niche. Like
19:52most people are not aware of like who's
19:54good, what's good, even like this
19:55databacked process. So we try to keep it
19:57in niche and then you can I'll show you
19:59a couple things we have where you can
20:00stumble across new stuff that's like
20:02outside the niche. But this is this is
20:04what I like to start with. So this is my
20:05list. You can see I just I had a list on
20:07Excel. I just pasted all the URLs in. I
20:10added them super quick. So this is the
20:11list that like I I like. So once you
20:13have this, then you go to the videos
20:15tab. And so this is the to to zoom out.
20:18What we're doing here is we're breaking
20:20down how do you find the best performing
20:22topics in your niche so that you know
20:24which ponds to fish in. So if you go to
20:26the videos tab, what you now see are
20:28these are all the videos from just those
20:30channels. And you know that because if
20:32you click the channels filter, you can
20:33see the channels in here. So, if I
20:35wanted to look at like just my channel,
20:37I would click my channel and then I've
20:38got boom, I got all the videos from just
20:39my channel in here. Now, the difference
20:41between this and like Instagram or Tik
20:43Tok proper is that I can filter and cut
20:45this data however I want. So, what I
20:47like to do is I'll go posted in the last
20:50two months. I like to set engagement
20:53rate at 2% because that filters out the
20:55brand deals. Typically, if you have
20:57videos with super high views, but they
20:58have 0% engagement, it's because people
21:00paid to boost the views. I don't want to
21:01look at the brand deals because those
21:02are that give me a that gives me a false
21:04sense of what topics worked. So I just
21:06set this bottom engagement at 2%.
21:08Anything below that filters out. And
21:10then what I do is I go to sort and I'll
21:12go either outlier score or views.
21:14Slightly different ways of calculating
21:15it, but let's just do outlier score. So
21:18the outlier score, which is this green
21:19badge here, you can see like this one is
21:218.2x. This means that the views of this
21:24video were 8.2x 2x higher or above the
21:28average views in the rolling 3 months
21:30before this video was posted. We do
21:32rolling three months so that if the
21:34channel's dying, the outlier score kind
21:36of gets recalculated. You don't get
21:37punished for like super fast or super
21:38slow growth. But so an outlier score is
21:41like a better apples to apples
21:42comparison across multiple channels. So
21:45the beauty of this is like let's say I
21:46clear uh just this and I look at the
21:49whole cohort. Now, this is in the last 2
21:52months sorted by outlier score across
21:54all these channels, the top performing
21:56videos. So, like this is the the wise
21:58investor. I don't even know maybe that
22:00maybe he's on the channel, but there's
22:01Ganon, uh, Peter Visuals, etc. So, you
22:04go through. So, now what you can do,
22:06there's a few things that are like super
22:08helpful. Let me let me go by newest so I
22:10can show you this. When you see these
22:12videos, the very first thing you're
22:14you're probably wanting to do is either
22:16play it in a new tab, so you can just
22:17watch it. So you can just hover over and
22:19play a new tab like this. Or you want to
22:21get deep analysis on this video, like
22:22what is the transcript? How can I
22:24understand like what this video was
22:25about? Right? And so if you just press
22:27this analyze button, we will actually do
22:29a deep analysis on this video. And if
22:32I'll just go back to one that I've
22:33already analyzed.
22:35So like this one by Ganon, I can press
22:37into it. You get the transcript, the
22:40caption if you wanted is here. But then
22:42all this like deep analysis, right? So
22:44what is the topic? What was the the seed
22:46of the idea? The topic is like a 3 to
22:47five word kind of description. This is
22:49more like a single sentence description.
22:51What is the unique angle of this video?
22:53What is the common belief versus
22:55contrarian reality? This is like really
22:56where curiosity comes from right here.
22:58What kind of like core facts or evidence
23:00did he put in this video? Then we go to
23:01the hook. What was the formula for the
23:03hook? What was the exact hook? What was
23:04the analysis of why this hooked work?
23:06What is the category of the hook?
23:07Storytelling, same thing. Category, what
23:09what the storytelling structure was?
23:11Visuals as well. So the whole point is
23:13like now you're getting all of the deep
23:15analysis on this video. We're both
23:17taking the transcript and literally
23:19watching it with a vision model and
23:20giving you insights on this. Now at an
23:23individual level, you could be like,
23:25"Oh, cool." Like I'm just going to A lot
23:27of people just use sand castles like
23:28this. They come in, they filter, they
23:30sort by outlier score, and literally
23:32they just watch each one. They're just
23:34like, "Watch, watch, watch." And and
23:35they like watching it themselves. They
23:36don't even want the detailed analysis.
23:38This alone prevents you from having to
23:41scroll to find ideas and now you can see
23:43the best performers. I'll I'll expand on
23:45what I would look what I would take from
23:47this verse not to keep it original, but
23:49like that's that's like the lightest
23:51weight way to do it. The more advanced
23:52way is once you analyze a bunch of these
23:54and you can go one at a time or you can
23:56press bulk analyze like you know let's
23:58say you wanted to analyze my channel.
24:00You would just search my channel. You
24:02could say bulk analyze 50 videos or you
24:04know you could look and say okay that
24:06looks like there's 25 in the last three
24:08weeks. I'm just going to bulk analyze
24:0925. Once you have all these analyzed you
24:12can either export the data as a CSV and
24:14feed that into Claude or you can use our
24:16MCP which connects Sand Castles directly
24:19to Claude. You can just install it in
24:20like 30 seconds. I can I can link a
24:22video if people want. And then in Claude
24:24I can say look at the last 30 videos
24:26from Callaway Marketing's channel. Make
24:28sure they're all deep analyzed. give me
24:30a full report of like a breakdown of the
24:32topics, the hooks, etc. Right? So,
24:34there's so much you can do with this
24:35data individually and as a cohort.
24:37>> So, you're coming in here for Callaway
24:39Marketing and you're doing your uh your
24:43filtering, your sorting, you have a
24:45short list by outlier score. When do you
24:48stop and say like, "Yep, that's one I'm
24:51going to do." Is it something that like
24:52inspiration hits or are you doing a
24:54minimum 10 and then saying, "Okay, which
24:56one of these did did I feel best?"
24:58What's your process from here? So
24:59there's so there's two paths. There's
25:01looking at what competitors are doing to
25:02get new ideas and then there's looking
25:04at your own channel to understand all
25:06the other stuff that's working really
25:08well. So like which videos did I make
25:09that crushed? What hooks are working?
25:11What formats are working, right? So I I
25:13kind of like do both separately, per se.
25:16So I've got a guy on my team um that
25:19literally just does the idea stuff all
25:22day long. So, he's looking at all of
25:24these competitor channels and he's like
25:27trying to like he looks at this this
25:29view right here and sometimes he'll go
25:31uh last, you know, 30 days. Sometimes
25:35he'll go further out. It depends on like
25:36for your niche and my niche. This is
25:38pretty evergreen. So, you can go back
25:39like four four six months and still get
25:41good ideas. For newsbased niches where
25:43you need more like real-time stuff,
25:45he'll do 14 days or seven days like for
25:47tech. So, he'll look at this and he'll
25:48open this video and he'll be like,
25:49"Okay, is the topic of this interesting
25:53to him?" First of all, like because I've
25:54kind of trained him on like what
25:55interesting looks like just like are you
25:56curious about this period? Okay. Number
25:59two, do we have something relevant and
26:03non-obvious to say about this general
26:06topic? So, like let's just look at this
26:07one right here. So, we'll open this. So,
26:09let's just watch this.
26:12This video got nearly 25 million views.
26:15[music] That's over 200 times more views
26:17than the creator's regular content. And
26:19the reason comes down to one thing. Your
26:20brain is always trying to guess what
26:22happens [music] next in a So this is
26:24basically like a a hook case study where
26:26he's explaining the hook or psychology
26:29of this, right? So my my guy or you or
26:32anyone who's doing this, right, would be
26:33like, okay, this is a breakdown of an
26:36example video that had a viral hook that
26:38had traction that he explained how to do
26:40it. So, the close clone, which I don't
26:42really like doing close clones, would be
26:44like, let's take this exact same video
26:45and make a like this exact same example
26:47and make our own spin on it where we
26:49break down the hook, but in a different
26:50way. We say different stuff. To me,
26:52that's still too close. I want to be
26:54more unique. So, I would say, okay,
26:56interesting. The style that he went
26:58through to explain this is he started
26:59with the hook itself of of the example.
27:02He played it without saying anything.
27:03Then, he introduced what it was and
27:05broke it down. Could we run a similar
27:08flow like that, a similar story
27:09structure, but using a different
27:10example? Do we have a different example?
27:13Do I have unique insights to say on a
27:14different example? That would like start
27:16the process. So, if if my uh my like
27:18head of content pulled this and he he
27:20thought we would have something good, he
27:21would put it in the list for us to
27:23review. If he thought it was like too
27:24junk foody, he would pass, right? But
27:26the reason why like that's helpful is
27:28all these are on target and they're the
27:30best performers by the people in my
27:31niche. So it just focuses your yes no
27:34process super dialed. So he'll do this
27:37uh daily or every other day and he
27:39builds a Google doc that has tabs on the
27:41side and every tab is like a separate
27:43idea. He's only putting the ones on
27:45there that he thinks we have something
27:46good to say, right? And it's based on
27:48the whole corp like I've just trained
27:49him over time like the whole corpus of
27:51everything I believe. He's seen all my
27:52YouTube videos. He knows kind of like
27:53our isms, let's say. Then we get on a
27:56call. I like to do it twice a week where
27:58he'll pitch me each one and be like,
28:00"Here's what I found. This crushed.
28:02Here's what I think we could do. What do
28:03you think?" If I don't have something
28:05unique to say that's like contrarian and
28:07interesting, we kill it. If I do have
28:09something unique to say and I'm like,
28:10"Ooh, that's cool. We should use this
28:11other video as the example." Or, "Ooh,
28:14we could bring in this one framework
28:15that I talk about and like, you know,
28:16I'll I'll construct it." We we green
28:18light it and we put it into like
28:20approved bucket. We'll go through that
28:22call like we just did one today for an
28:24hour or so. And what that gives us is
28:26like eight to 10 approved ideas where
28:28they're validated by data. I know we
28:30have something interesting. We went back
28:32and forth for five minutes on the call
28:33for each idea and like kind of
28:35brainstorm the best option. And then
28:36from there I know I can go into the
28:38scripting phase. So that's kind of like
28:41the research flow. Now he's most for
28:44short form mostly he's using sand
28:45castles. Like he doesn't really scroll
28:47that much. Uh for YouTube he's looking
28:49at uh or like for for tech ideas he's
28:51looking at Twitter and other other
28:52things like that like email newsletters
28:53around tech. But for like Callaway
28:55marketing, the marketing ideas, he's
28:57pretty much just looking at this. So
29:00that's the how do we get new ideas just
29:02on the topic front.
29:03>> How are people doing this before they
29:05there was a tool like this? Were they
29:06just like
29:07>> manually?
29:07>> Manually with like uh bookmarks and
29:11folders
29:12>> and spending a lot of time on their
29:14phone.
29:14>> Yeah. like they either had a person like
29:16a social media strategist or them and
29:17they were scrolling for like two to
29:18three hours a day and they were having
29:20to go through all the noise, all the
29:21nonsense is not on target and then
29:23they'd have to see all the ones that are
29:24like mid- performers. Then finally
29:25they'd get to some bangers and the ratio
29:27from my research was like one out of 10
29:30or one out of 20 would be like even
29:32close to potentially relevant and then
29:34you'd have to see it. You'd you'd
29:36bookmark it or you'd save the link into
29:38a notes file. You'd have a bunch and
29:40then you'd go through this process of
29:41like is it relevant? Is it not? So, this
29:43just speeds up a lot of the research
29:46time. So, you don't have to spend all
29:47that time doing the the finding. Right
29:50now, we also have like this automations
29:52feature, which I think is pretty sick.
29:54You can set up these rules that will
29:55automatically deep analyze whichever
29:57videos you want based on the conditions.
29:59Right? So, what I'll do is we have this
30:00one on that says for any of the channels
30:02in our list, as long as there's a 5x
30:05outlier score on the video and it has
30:07above 2% engagement, we want you to
30:09autoprocess it. So it'll automatically
30:11be analyzed. And what that means is when
30:13it's analyzed, it's ready to be pulled
30:15from in Claude, right? Because we have
30:17the MCP set up. So we could just set up
30:19a daily scrape that's like, give me the
30:21best performing videos of the day today,
30:24every day. And we could set that up as a
30:25scheduled task. It can do that. I mean,
30:27it could it could just give you the
30:29video to like link out without analyzing
30:31it. But by analyzing it, we can also
30:33have it say, "What are the common topics
30:35that are working?" And it can just know
30:36because it's already done all this
30:38analysis for you automatically. So, this
30:40is just like one way to speed it up cuz
30:42you you can tell like my whole thing is
30:43like how do we speed this up as fast as
30:45possible and get as much signal as we
30:46can in the workflow.
30:47>> This clearly solves my topic ideation
30:51and even to a degree hook problem
30:54because I can see how this is working.
30:56The next place where I feel like I would
30:58fall down is if the video is too long or
31:01in some way like the structure of the
31:03whole video goes off the rails. Is the
31:06answer to just use this analyze feature
31:09and look at the structure of these and
31:11make sure I'm following fairly closely
31:13to it and then just be relentless in the
31:16edit.
31:16>> Yeah. So, there's two ways. So, when
31:18you're when you're first starting and
31:19you don't have a ton of your own data,
31:20like you don't have a ton of your own
31:22videos, the best way would be find a
31:23creator that you like, like let's say
31:25Ganon, find the a creator where you like
31:27the flow of. It's not like you like the
31:29look of his stuff cuz that's the
31:30visuals. It's more like you like the way
31:31he explains stuff, you like the flow of
31:32his storytelling. You can use this
31:34transcript and you can either look just
31:36look at it and be like, "Okay, roughly
31:38speaking, these are the sections." We
31:39also have a feature right here that's
31:41show sections. You just press this. So,
31:43it breaks it down into like the five
31:44parts or they're not always five, but
31:46this in this case there's five parts.
31:47So, you could then say, "Okay, cool."
31:49Like, if I'm going to storytell, but I
31:50don't want to I don't want to copy his
31:51stuff. We need a hook. We need an intro
31:53to the series. Then, in this case, he's
31:55doing a setup tutorial. There's a live
31:57demo and a payoff. So, like at the at
32:00the roughest strokes, you could copy
32:01that structure. At the most detailed
32:03strokes, you could take his exact script
32:05and then try to translate it into your
32:06own. Now, you could do this manually or
32:08of course you can take a hundred scripts
32:10from Ganon, put it into Claude and be
32:12like, I want to understand what the best
32:13performing structures are storytelling
32:16wise from the scripts. Make me a skill
32:18file template and then when I jump in a
32:20topic, write a script for me in that
32:22way. Now, I likely would not just just
32:25go with that. Like, I would take that
32:26and then I'd tweak it. That's going to
32:27be 90% of the way there. But that gets
32:29you again like it puts guard rails on
32:31what you should do when you don't know
32:33what I recommend doing once you have a
32:35minimum amount of videos for yourself
32:37like 10 to 20 videos you know let's go
32:39to mine I have all of my videos analyzed
32:42so of course I can look one by one like
32:44if I sort by outlier score my eye is
32:46trained on content now where I can see
32:48like let me go 60 days so we have all of
32:50them in there I can see like okay this
32:53was about kind of content strategy this
32:55was about AI tools this was about
32:57storytelling this was about formats This
32:59is about strategy. Strategy. This one I
33:02think is storytelling. Storytelling. So
33:04I'm starting to see like common buckets
33:05like I can just look at it and see. But
33:07on the on the story structure side, the
33:09storytelling like with the script, you
33:10could just take these and once they're
33:12analyzed, you can again like export the
33:14data or use claude and just say, "Hey,
33:16I've made 20 videos. Can you look at the
33:18top the top five performers and just
33:20help me understand like what is the
33:22common structure or flow of the script
33:25that's working well for me? let's turn
33:27that into a template or turn that into a
33:28skill file and then when I feed you a
33:30new topic, can you write the script or
33:32at least guide me into what those
33:33buckets should be? To your point, like
33:35that'll that'll speed up a lot of the
33:37confusion. I still think you need like
33:39writing is the number one creative tool.
33:41Like I believe writing is the most
33:42important skill. I think you believe
33:43that as well. So I never want to like
33:45fully outsource every word I say to
33:48this, but like as a as a 75 85 90% of
33:51the way their first draft automatic,
33:53it's pretty sick. And then you can tweak
33:54from there. Yeah, I agree. Um the the
33:58conflicting feelings that I get when I
34:01think about processes like this is if
34:03you remove all of your own creative work
34:06and writing from it, which you totally
34:08could. Like I could make this totally
34:09data driven and totally like automated
34:13formula
34:14>> from this and it would probably get
34:16results and I just know that would be
34:18like not satisfying for me either. Like
34:21what I like about these this analyze
34:23tools and these headers that you have in
34:24this transcript, I can teach myself to
34:27fish with this tool. I don't have to let
34:29the tool fish for me.
34:31>> Exactly. The way I like to think of it
34:32is so in in the content workflow, you've
34:34got topic, you've got research or like
34:37substance, you've got hook, script,
34:40record, edit, post. In each of those
34:43little little modules within the flow,
34:45you can decide to turn it all the way to
34:47100 on data or all the way to zero on
34:49manual. and you can pick where you want.
34:51So like for me, where I'm at right now
34:53is I really like using 100% data to
34:56guide me on the topic buckets so that I
34:58know where to start fishing. Then I look
35:00at my own my own data. Like the most
35:02valuable thing like just on YouTube, the
35:03most valuable data for you on YouTube is
35:05your own channel. Same thing here. But
35:06until you have your own data, you just
35:08guiding on the topics that make the most
35:10sense from other people is the best way
35:11to do it. For substance, the contrarian
35:13insight, I'm relying on my creativity.
35:15I'm not really like taking other
35:17people's insights cuz I find that's
35:19really like the sauce that makes it
35:20unique and that's the creative challenge
35:22for me is like okay how do I take this
35:2430 days of content topic and make it
35:26unique. So I came up with this new
35:28mechanism and like that did pretty well.
35:29So on the substance I'm doing it full
35:31creativity on the hooks and the
35:33scripting. What exactly what you said is
35:35where I'm at which is like the goal is
35:37to make your brain full of what works so
35:41that when you flow naturally creatively
35:43it's already infused in there, right? So
35:45like you just when I look at this I'm
35:48like okay let me study right here. Let
35:50me study what this is. When I make
35:50another one I don't copy paste it. I
35:52just read it just like copy work or like
35:54any form of training yourself on a
35:56skill. I'm reading it and I'm like, "Oh,
35:58okay. I see like how that flowed and
35:59that didn't. Okay, now I know." So, when
36:01I go to write from scratch, it's kind of
36:03like in my brain already. The way of
36:05doing this, that's the like most manual
36:07way to use data, but still be fully
36:09creative. You can of course dial that a
36:10little bit more where it's like, cool,
36:12I'm going to start copy pasting this one
36:13that worked. It's my own words, but like
36:15I'm going to start here and then edit
36:17around it. The the full automated would
36:19be like literally just rip this and
36:21change a couple words. So to your point,
36:23you can season to taste on this
36:25depending on how much creative
36:26expression you want. I don't think you
36:29can fully build an AI factory pumping it
36:32out with zero creativity and cut
36:33through. I just think there's too much
36:35supply and it's going to trend towards
36:36average because most people are doing
36:39that and they're coming up with the same
36:40stuff. So this is where like I call it
36:43data enabled creativity. That's my like
36:45big theme or thesis that I believe in is
36:47like using the data to guide you and
36:49then allowing basically freeing up all
36:51of your time to be creative and using
36:53data for everything else. What if I took
36:55one of your uh one of these videos that
36:57we're looking at on screen right now and
36:59I removed all visuals? The cut stayed
37:02there, but it was just you and it was
37:04just captions. How big of an impact do
37:06you think that would make on the
37:08performance of that video?
37:09>> A pure yap with zero graphics, which is
37:11kind of what you're describing, right?
37:12like full A-roll U and captions ba
37:15basically the way I the the framework I
37:17use to look at this is like a picture is
37:19worth a thousand words right so if you
37:20have no visuals on the screen you're
37:22relying on the viewer's imagination to
37:24to come up with if they're a visual
37:26learner like I am they have to imagine
37:28what you're saying and try to get
37:30comprehension from that more technical
37:33or the more complex or the more like out
37:35there ideas you're talking about the
37:36harder it is for people to imagine it's
37:38a spectrum right it's like the more work
37:40you the more frontend effort you apply
37:42to the production, the le the the easier
37:45it is to carve a moat visually because
37:47there's just less people putting in that
37:48time. However, to your point, the more
37:51resource intensive it is. So, you have
37:52to just experiment along that spectrum
37:54and figure out what you are happy with
37:57and then ultimately what converts into
37:58emails. That's that's a really critical
38:00thing too.
38:00>> Amazing. Well, I will link sand castles
38:03in the description for this. That's
38:05fantastic.
38:05>> If you're a business owner, hopefully
38:07this is helpful to you. Definitely free
38:08stuff below, so feel free to check it
38:10out. Now that you've seen exactly how
38:11Callaway structures his content
38:13research, you'll love seeing behind the
38:15scenes with Riley Brown, too.

About this transcript

This page contains the full transcript of Kallaway’s EXACT System For Going Viral (Over and Over) by Jay Clouse, generated from the public captions YouTube serves with the video. The transcript has 8,838 words across 1,204 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.