The FREE SubmitHub Tools Most Artists Don’t Know About

Jason, if you only had one minute to give music artists the best advice you possibly could, what would you say? 1 minute on the clock. >> Oh, Andrew, thanks for having me on here. I feel like I don't even need a whole minute to do this. Um, but I love these pop quizzes. Uh, if I were to give artist advice, I would say it's much more important to be a big fish in a little pond than it is to be a minnow in the ocean. And so what I mean by that is if you are beginning in your career, focus on local because if you jump into the international circuit, you're suddenly competing against millions of other artists who are doing it as well. And you're going to start connecting with people who you'll never actually be able to engage with. Whereas if you start local, you play your church, you play the coffee shop, you play the convention center, whatever it is that you're playing locally, you're way more likely to create proper connections with artists. you're going to get way more reward from that. Um, and you're going to be a big fish in a little pond and that's kind of cool. So, I would do that. Forget about going international and focus on local. And, uh, I've got Oh, that's a minute. Done. Getting negative 3 seconds. Nailed it though. >> Yeah. Yeah, I mean I feel like that also applies for uh if if being in a subg genre and not trying to be like have like mass commercial appeal like I think before in the in the olden days pre pre where everything was an algorithm um it was you kind of needed to maybe change your sound to maybe get more mainstream appeal and now like everyone's on their own little algorithm on every platform you go on YouTube like my YouTube is completely different from your YouTube and Tik Tok and Spotify. And so I think it's actually better to be like a alternative Viking pop artist than it is to be like a pop artist. You know, >> find your niche and most importantly enjoy what you're doing because you're probably never going to make it. So like music should be a channel for for engaging and discovering and finding and learning new things. Uh, and at the end of the day, I like to say that that uh, even for artists who feel like it's been a let down and they haven't succeeded, they've done so much more than the average person who just sits in doom scrolls or watches Netflix all day, like they've actually done something proactive and there's reward and value in that. Um, so yeah, that's my take. >> Yeah, man. Well, you've you've been on here. I mean, this this video is going to go live on both the My4 Sense podcast, but also my channel. um depending on when people >> one of the best channels on YouTube. >> Oh, I'm gonna take that quote and everywhere. >> Um but you you've been on you've been on my channel a bunch of times and so a lot of people here probably already know who you are, but >> uh could you give people just like a little minute backtory on who is Jason Gishkot? >> Oh, sure. Uh I am the founder of SubmitHub and Indie Shuffle. I've been music blogging since 2007 and I've been running SubmitHub since 2015. I created SubmitHub to solve a problem that I had with the music blog which was that there was a very high demand of musicians who wanted to be featured and I didn't have the bandwidth or capacity to go through all of that. I was getting about 300 emails a day. And so for me to sit down and actually listen to all of those songs, it was impossible. And what it meant was that I often just went and regurgitated whatever was popular at the time. Uh so Submitub was created in 2015 to solve that problem and make my inbox manageable and the idea caught on. So it went from blogging in 2015. Around 2017 we introduced Spotify. We attempted Tik Tok and Instagram but that's a that's a different story. I feel like you're you're better at that. Um, and then, um, for the last almost 3 years, we've been stomping on your toes a little bit with the meta ads, but I I think we're in different spaces there again. And I and we learned a lot from you because you've got the best music marketing channel on YouTube. So, that's it. That's that's a quick one minute. >> Yeah, man. Well, thanks. Yeah. I mean, so, uh, Submit's changed a lot over the last few years, right? And and like for example, you brought up the whole Tik Tok Instagram thing. Um, like there was a point when when you guys were going pretty hard on let's have make let's add a the ability for Sventum to kind of be like an influencer outreach thing because because that was a was and kind of still is a big part of the music industry. Um, and then over the years you eventually added marketplace and you you added your smart links feature and then you added an an automated ads tool and um now I feel like I almost don't get through a consulting call without mentioning SubmitHub to someone when they're like oh like what should I use this this no matter what you end up coming up for for a few reasons. One you have the best free smart link tool on the internet that I'm aware of. Um, there's no like alternative. Like it's kind of like you and Tonen and >> that's Tonen's kind of laughable right now. >> Um, so yeah, you're winning that race, but also like you have the AI detector, you have your playlist detector, you you have all these like little free tools that are helpful. Um, and then you also have, of course, like kind of classic Submit Hub where like someone can go in for like 20 bucks and get get a handful of playlists or get someone to check out their song or so you kind of inevitably come up uh no matter what you're what what I'm talking about with with someone. >> Cool. Uh, look, oftentimes I'm just building things for what people are requesting or talking about or using. So, uh, the AI detector is one of our most used features now. And I built that because curators were complaining that they were receiving AI songs. So I started by asking artists to disclose did you use AI or not and about a third of them lied about it. So that's where I had to go and build this detector. I started Yeah. I started um on on the backs of Daer if you will. So Daer actually published a lot of their research publicly and so I was able to learn about their dict detection techniques and integrate them into my own model and I built it from there. So there were certain things about theirs that were effective and other things that were not as effective. Um and so yeah it's it's been uh two and a half years I think of the AI detector being present. It's on version four now and I'm I'm constantly improving. It is not perfect. Um, many artists are quick to point that out, but we did have a third party come through and run a thorough analysis on it. They tested 20,000 songs, and it came back with a 99.4% accuracy for them. So, it's pretty good. It's not perfect, but it's actually the basis of a lot of the decisions that I've been making in the last couple of weeks, which um are are slightly controversial, shall we say. >> Yeah. I I I think that's be a great like kind of first big main topic here is you guys had a big announcement I think a week ago at the time of filming this or early September or I don't know when this video is going to go live. >> Early September. Yeah. >> The you guys announced that um essentially you're not allowing AI generated artists to to pitch to curators in the platform. So, and um you're using auto detection for that, right? Someone adds their song, you detect if it's AI generated, and it's not like a partial like you're looking for 100% AI generated tracks. You're not like necessarily detecting >> I'm looking for confidence in the detection. It's it's a very nuance one. If if you go into the announcement, there are I think now 480 comments as of this video. A lot of it is people debating the nuance of the detector and whether it's accurate and what constitutes uh use of AI. Is it uh you know if I write the lyrics but then use surely that still means it's my song and it's not AI right? To which I say wrong. Um, >> and and so when I made this announcement, I I I sort of tiptoed around the subject. I I've been very careful to not share my personal opinion about AI. It's a topical thing and people on both sides argue a lot. They feel very strongly. Musicians feel very strongly that AI is destroying music and AI creators feel very strongly that uh it's the future and that musicians are just scared. Uh so over the last two years my policy has been one of of transparency and disclosure. So the idea being look if you want to create AI music and share it that's fine. Uh I'm going to connect that group with each other and if you are opposed to AI I'll connect you guys together. So at least everyone's in their bucket. But I I tell you I cracked. Uh I was I cracked man. Look at me. I cracked. Uh, I was shopping at my grocery store and they replaced all the music with AI slop pure like sununo 3.5. >> They used to have really good music in that grocery store. >> Not even a quality uh uh five plus algorithm or now >> not even. >> So I I emailed them and I went straight to corporate. It took it took like 10 emails to get, you know, it's like it's like dealing with Spotify support, right? You just have to keep hammering until you talk to a human. I did. I finally got a response and they said, "Look, it's an effort to showcase South African talent." And I was like, "Wait, sorry, what? It's no, it's AI slop. It's clearly you have replaced your whatever radio you had with AI slop so that you don't have to pay royalties anymore. Some hot shot startup that sold you this package is taking their money, but it's not going to the artist, is not championing this." And that was my cracking moment where I was like, "You know what? We have hundreds of thousands of songs released every day. Like people always quote this 120,000 number from Spotify. That was four years ago. That was before AI was huge. I bet you it's like 500,000 songs easy per day. We don't need more slop. Like if you >> you got good lyrics. You put a lot of effort into songwriting. You've got great ideas for melodies you were dreaming about in the morning. And you had to go and like work six hours to push the button until it came back with what you wanted. Like I don't need that on the internet. For me, for me, like music is so much about the joy and connecting with the work that's gone into it, the creative process and and looking for some like what's a like something original, a spark. And AI itself cannot be an original spark. It is it is derivative of what it was trained on without permission. And so, yeah, I snapped. Um I snapped and that's it. No more AI slop on SubmitHub. I don't want the curators sharing it. There's a lot of curators who feel like they've lost out on free money. I'm sorry to you guys. It's for the good of society. >> And you can go you can go to Groover if you want that. Go to go. They don't they don't have an AI filter. Just go use Groover. >> If you want slot, go use G. >> Sorry, Dorian. Um, but yes, interestingly enough, the the last two Uber rides I've had or one was a one was a Uber and one was like a like a private limo or whatever. We we got like a scheduled car ride. >> Andrew Southworth is doing well. >> Private limo. >> It was actually it was actually about the same price as doing the Uber because it was like, you know, 40minute Uber ride. So, it actually say one of your clients was trying to impress you and I got you a limo. >> It was actually a family trip. It was uh it was just we uh Yeah, but it was about the same to to Uber. Like it was like$100 to Uber. >> Anyone can hire a limo. It's not elitist. >> Yeah, it's not it's not that elitist. Little little bit privileged, but you know. Anyways, ramble story. Back to the point. Um on um actually that and I think the the next two Uber rides. Either way, the last several car rides I've been in where there's been a person driving me somewhere, um, they picked the music that place, and in the last two of those I've been on, the music has been AI generated. And one of them was a YouTube playlist a person put on called like something blues or whatever, and it was like a 2-hour long YouTube video of just like AI generated blues rock. Um and then the second one was yesterday because um I had to go get a haircut and my wife had her car and we're just driving there and there's just an AI generated song playing and it was like a EDM track just a random thing on Spotify and I was like this sounds AI generated. So I you know I go and grab the Spotify link I jump over to uh well actually in my case I I have your API on my website. So it it's you, but I put it through the ads detector on my website, which is just submit hub and a rapper. Um, and it uh it analyzes it and tells me it's AI generated, you know. And I was like, I wonder if the person knows like either one the Uber driver is playing their own music, which I've also had happen with human artists, like I've been in the back of an Uber. I'm like, what is this like what is this awful song up here? You should have it. And then I look at the profile picture like that's our board driver. you know, he's like playing his own music as he drives out. >> Oh. Oh. >> Very clever and awkward marketing strategy. >> Yeah. Um either there's Uber drivers making it or two, they don't know that it is or what I think is is probably happening is yes, they don't know. But there's probably someone marketing playlists that are for Uber drivers. I like I guarantee it because like I've I've run playlists where we're marketing them at like people who are Twitch streamers or we're marketing them at people who like own coffee shops, right? Because those people just play music all day. So if you get one person who plays your playlist and you own a significant portion of the catalog in their playlist, you get a lot of streaming volume. And that seems like a ripe opportunity for for an AI business person, whatever, to like let's stuff a bunch of Uber driver playlist and market them. There's millions of Uber drivers. >> Yeah, it's kind of brilliant, but >> it's hard. I I'm doing my piece here to try and combat it. And I know there's going to be some uh collateral damage and and possibly some people who who are falsely accused of AI. Uh I'm there. I'm listening. I'm responding. I'm I'm watching the reports of false positives. I'm trying to teach my model to be able to recognize these better. But at the end of the day, like this, it has to stop. It just has to stop. Um, it's it's not good for art. It's not good for music. So, that's how that's how I feel about I feel strongly about it. Like, I'm becoming an anti- AI ambassador, but someone has to freaking say it. And and we are like we're taking a big hit in revenue from this. It's 10% of our revenue is coming from this pure AI. These guys, like if you think about the profile of a lot of AI creators, uh many of them are retired. They have a lot of money. They wished they could get into music. They have a lot of ideas. They have a lot of stories in their head. AI is a fantastic tool for them to bring those to life and to share it. But it doesn't need to go on Spotify. It doesn't need to be promoted. Like, keep it to >> your circle. Keep it to yourself. have it be something special for you, but like it's not special for music and and yeah, that's it. It's done. We're trying. >> Yeah, that that's an accurate description of what a lot of the people who promote AI music are. Um like you know my my agency we we ba like I thought a lot about this too and my process going ever since then was like I feel like I shouldn't play god on what's like allowed or not allowed to be marketed. So the same goes for AI music versus human music, but then also for example if I don't agree with its like political beliefs or I don't agree with its religious stance or anything like my approach is like if I think there's an audience for it and it won't get me in trouble with meta or the law to promote it >> um then we don't do it. So for example, if someone has like a political sign and don't agree with it, that's fine. But if it's like hate speech, it's not fine. So that's kind of how I draw the line. And if it's AI music and I think it it there's an audience for and I think it could do well in ads then we'll do it. Now the the you guys >> challenge yourself Andrew challenge yourself. This is your chance to rethink this cuz >> well the change is >> how much do we want to help >> considerably. >> It does right. So so it means you did a great video about this recently. You've done two videos now about it. But um Spotify is supposedly cracking down on AI personas. There are examples of them doing it. There are examples of them not doing it. There are examples of what seems like other artists getting caught up in the coattails of some change to policy here. Um and and initially I thought this would be you know this was this was like the reason I could say I could I could let my personal side which I've been hiding come out and be like oh well you know it feels unfair to be selling promotion that's going to fall flat in the end. Um, and and halfway through I was like, "No, it. I just got to say it. I just have to say it and honestly give it up." But I mean, that is it's it's it's potentially going to fall flat. They're not enforcing this policy wild scale yet. They haven't said like, "Can an AI a real person have AI tracks because they're not an AI persona, which is nonsense and I hope they don't do that." Um, but I would encourage you to like rethink this because it's what I've been doing. So, like we took 10% hit in revenue. It's a big chunk. Um, yeah, it's been great growth for us. >> Probably roughly where we are. I would think it's probably 5 to 10%. >> So, it's a small chunk. >> You know what? 50% of our customer support was coming from these AI creators because >> for them, this is their first time diving into music. They don't know how it works. They're not used to rejection. They're wondering why they're not getting streams. They were supposed to like blow up. >> And so, there's a lot of new users here. So yes, we're losing 10% of our revenue, but we're gaining back 50% of our sanity. So in a way that helps. >> There's one interesting case that shows that is is uh we we had a song, this was 6 months ago, you know, before all the algorithmic changes. U and it it was a, you know, AI generated track. We're running campaign for it. I think they had like two or $3,000 budgets. Pretty hefty, you know, pretty nice budget. And um it goes great. It's getting like 15 cents a conversion in tier one countries. phenomenal performance and um the client is pissed at this result and it's like we're like this is like a this it really does not get much better than this. Like is it technically possible for it to be better? Yes. Uh but you're in like the top 1% of performance in that case. And the algorithm is looking good. Everything's looking good. And he's just like yeah but like I'm not it's been it's like two weeks into the campaign. He's like, "Yeah, but I'm I'm not profitable." I was like, "You expected to be profitable like two weeks into this campaign?" Edward is an AI artist with no. And he's like, "Yeah, why would I even bother releasing music if it wasn't going to make money?" I'm like, "Oh, >> that's and and that is right." And that's why um >> All right, this is my Andrew Southworth challenge. Just stop the AI slop. You can do the hybrid stuff, but like stop the AI slop. Let's work on this together, Andrew. We can fight it together. >> Therapy session. So I've been re reconsidering um reconsidering with the change in the algorithm and so maybe that'll change but that's been my my thought process up to now is like it shouldn't be my prerogative to say like what happens what is and then I'm like fighting people on like oh why you don't doing this why are you doing this and >> that's where I was I I was trying not to pick a battle um and there was good money in it >> and that's the wrong reason to do it. I'm I'm changing my tune. I'm done with it. And I and I hope that the tides start to shift now because it's getting out of hand and the general public cannot tell. We are curators. We are like in the music industry. Both you and I run curation as well and and we got to stop it, man. We got to stop it. >> The the I think a a lot of it's probably going to be throttled just from the changes. One, it's Spotify, so there's there's it's a lot harder for the financial incentive is not nearly as there. still possible, but it's kind of fallen into a lot of smaller cracks kind of. And honestly, a lot of the lower cracks it's fallen into are ones where human slop was already a thing, right? Like making just super simple like kind of sleep music and >> yeah, >> that's kind of what it's fallen into. Um, and that you could do with human slop and separately from AI slop, but also Sunno changing their policies where you can only download 20 to 60 tracks per month depending on your plan unless you pay more. >> And fortune, they're charging a lot of money for it. Yeah. >> $3 per song for for a download. And um, their new model is trained with permission and they they deleted all their old models from use. So V6 was fully consensually trained with Warner Music, I believe. Um, and I'm sure there's some gray area with the label not getting all the permissions it should have, but either way, it's not it's a far cry from scraping the entire internet against everyone's will. >> Can I just say Can I just say those guys? Not Suno, but Warner and the labels. Cuz cuz who's taking the money here? It's the major labels, not the artists that I got trained on. And many of them probably didn't have any consent for this. It's just that Warner owns it. >> It's bad. I mean, the good news is that >> Sunseo 6 model, here's the other thing. They say it's trained on this other music, but I think they they kept the underlying like model and and the underlying logic in place. I feel like they didn't have to replace that. Um, and so those those weights still exist because the Sunno 6 model comes back with the same characteristics as five and 5.5. So, I didn't even have to update my model when they announced it. I was like, "Oh I got to go get, you know, 10006 songs." And then I started testing them and like, "Okay, well, they all get picked up." They they mostly come back as hybrid AI in my detector. And the reason is because for it to come back as pure AI, it has to be like literally the same as what was downloaded from Sunno. And so, it's not literally the same. I'll maybe update it that at some point, but like how many points am I going to like soon? >> Yeah. >> Six 6.1 6 point like I don't know. It's not they're not changing it yet. They're >> that that's an interesting point. So it's so similar where your ad detector barely had to be changed which is wild. >> I haven't changed it yet. One one particularly annoying thing with with AI music has been like and I have an account on label radar you know because periodic I'm like let me go see if there's any like cool music like sign it promote it you know like human music and you go on there and it's just all pseudo tracks and it's like why would I ever sign someone else sell someone else's pseudo track when I could just go generate the same thing like it makes no sense right >> because of the thoughts and the the lyrics that they had to put through chat GPT to like refine fine. Think about think about how hard it is to be a songwriter. Sorry, I don't want to diss songwriters. Like, there are good songwriters out there, but >> yeah, there are a lot of people doing cool music things with AI. Like, I I've I've met people where they had these full instrumentals that they they did and um they did them like years ago and now they're they kind of like brought them to life. like they did them 20 years ago and they went and had a family and then and they have these really cool recordings they wrote but like they they don't really play music anymore but they want to hear what those are like in a final form and it's like cool yeah put that through it you wrote it distribute it make money from it I don't care like you actually put a lot of effort into it the thing is in that the final output's 100% AI generated so like stuff like that gets completely even if you record >> since we've announced for like mixing and mastering the final output will be considered 100% AI generated. >> Exactly. I I've had people point out that like I took my human song, I literally uploaded it to Sunno and downloaded it and now you say it's AI. >> Yeah. >> And I was like >> this is a limitation of the detector. It is like whatever's happening is stamping it with stuff. They're processing it. They're putting it through their system and those are the signals that I can pick up on. So >> don't do that. Um, >> yeah, the irony is like that's that's the easiest thing to not use Sununo for, right? That's the cheapest part of the whole music production process to outsource is finding like a mixing mastering engineer. There's plenty of fantastic ones that would love to to get some early work for like, you know. >> All right. All right. I've vented enough on AI. >> Let's move on. So, you you sent me another cool one that is a different type of AI that you guys have added, which is your uh song analysis playlist finder. >> So, you can upload your song to submit hub and it gives you a genre breakdown, which is cool. But the cooler part is gives you a list of curators that your song might be a good fit on, which is also pretty handy. And I tested it with my song and I gave it a pretty hard one. I gave it an al alternative metal metal core type track. Um, and it nailed it. Yeah. And when when we we when we met up a couple months ago, I remember telling you about the your one of my songs in your genre detector and we were like seeing how accurate it was. Um, because you you I think you even admitted that like metal was probably the one with like the lower amount of data. So, it wouldn't be as accurate for, but actually nailed it. The genres are pretty good. and I listened to some of the playlists and it was like actually like pretty damn good fit. Um, and I'm guessing you made this because of a thing that you've been dealing with for years and I know you've been dealing with for years is people pitch to the wrong playlists, get rejected and then get mad at you. So, yeah. Yeah, a little bit. Uh, uh, okay. The history here is that there's this really cool group of researchers in Barcelona that receeded uh what they call the essentia tool set, but in that they've got all sorts of really cool analysis tools that can extract things like BPM. That's fairly basic, but they could also extract moods. And they released a very early model of genre detection based on discog. So Discogs is a website that people have been tagging music on for decades now, but they have this huge library of human tagged music and it's tagged by genre. So they were able then to take, you know, a million songs from discs that were categorized, put that into a machine learning algorithm. So So what they basically do is for each of those songs, they extract a fingerprint using their own technology that they came up with. And that's that fingerprint is looking at things like, you know, how how loud it is or what the tempo is or are there drums or are the these things there. So they they create this fingerprint. They've got all these buckets. And so when you give them a song, they go, "Oh yeah, looking at this fingerprint, it fits into this bucket." So that was their really basic one. Then about 3 years ago, they released a new model where they they expanded big time. They went up to 3 million songs across 500 plus genres. And they called this their uh maest model m- aes t and mist was the fingerprint that they came up with. So they they came up with this way of fingerprinting songs and then again they did this bucketing thing with 3 million songs across 500 genres and it worked really well. So we have been running that on submit for a couple years. But there's one problem with it. Discogs their genres don't match submit genres. So I had to do this mapping where they'd be like okay you know it's this kind of music goes here or if they say flamnco that's easy we have flamco but if they say old school house do they mean deep house Detroit house or Chicago house I was like well that must mean all of them so the mapping wasn't great a year ago I tried to solve it I was like well they're missing genres they're missing solo piano or funk or all these let me add those in on top so like I I I created my own little buckets and I used this fingerprinting to to try and organize it all and I rolled it out and we've had that on there for that latest model was on there for a year and it was really it was good. So it solved the problem that you're pointing out where people are sending completely the wrong genre. Like most artists don't know which genre their music listen belongs to. Like curators we know we listen to enough music we know what we like and we and we learn over time what these buckets are. Like what is alternative metal, death metal, doom metal, sludge? Like what are all these different sounds? Post rock, black gaze, all these crazy things. Like everyone I've said, you're probably like playing a little bit back in your head because you kind of know what they are. The average artist doesn't. >> What would you call something like, you know, deaf tones? Like they're kind of like alternative. Toast, >> toast grunge or but like they fit like there's this >> they have a lot of different tags that would apply to them because they they kind of in a way carve their own lane that now a lot of people are copying and calling like deaf tones core because of how like define their so it's it genres are a tricky thing because it it >> you know you have a very clear definition of what you call them, right? >> Um but so does everyone else and it's it's not always the same. So, you kind of just have to pick one and be consistent with it. Um, that way as people like at least use your platform more and more, they'll be like, "Okay, cool." Like what I would call alternative metal, maybe Submit Hub thinks is metal core, for example. You know, >> right? Right. Right. >> So, it's a bit of a problem. >> Right. >> There's a really cool website called every noise.com. Uh, and it was this guy about 10 years ago who who came up with this incredible model that mapped 6,000 different genres like by proximity and style and sound. Spotify ended up acquiring the company and they actually sadly let him go a year ago. Um, one reason is that Spotify has learned that that putting things into genre buckets is too difficult. Like so what I'm trying to do here is like almost impossible. Uh, YouTube also learned this. I've got some friends at YouTube. I used to work at Google. So, I've got some connections there who still work at YouTube Music. And they like I was talking to them about this genre detector I've been working on now for the new model for six months. I'll get into it or this is the segue, whatever. Uh, and they were saying, "Look, we gave up on genres and buckets and we're looking at at like associations instead, clusters, like these people listen to this and these listen to that, and so we cluster things that way rather than putting them into genre buckets." So, Spotify ended up firing this every noise guy. I I don't know what happened internally, but what a cool website. Check it out. Everyise.com. >> Yeah. >> Um, and every single one has a playlist on Spotify called The Sound of And then you insert the genre name. So, what I started to do is I said, "Look, you know, this model I've had up for a year is great, but it keeps on giving the wrong genres confidently. Like, every house track was coming back as old school house. It's not. It's like tech house or melodic house or this, but it was all coming back as old school house. So there was a flaw in the model that was constantly repeating these same genres, old pop, dance pop, old school house, and trap. And I'm like, no, but there are subg genres within these we should be working with. So I set out and I said, you know, here's SubmitHub's 260 genres. Let's go and find 300 songs that represent each one. And so for the last six months, I've been dealing with like this back and forth of trusting Spotify too much. Like their soundoff playlists, they're not as cohesive as as they sound. Um, so I've been like rebuilding these folders meticulously along the way, taking fingerprints of songs. Um, and I've got about 50,000 songs in there, most of which I have now listened to and actually sorted. So, it has been like an all day, everyday project since May. Uh I I'm driving my wife insane because I'll be like every time we hear a song I'll be like oh yeah that's uh that's postgrunge or that's sambas or that's this or did you know that like sca actually comes from reggae? Uh and I know that because the sound of sca playlist has a bunch of reggae in it which was really confusing to me when I was sifting through this. I was like why is there reggae in here? And that's because scar comes from reggae. Um >> h long story short the new model went out like genres. >> The the new model went out two weeks ago and the goal here of course is to make sure that artists songs get to the right curators so that curators are sifting through music they want and artists have a higher chance of success. Um it has been do I want to call it like a disappointing flop maybe. Uh people have not been very receptive to it. Uh, the previous model had a 93% thumbs up rate and I'm currently cruising at 72%. Um, which is it's disappointing because I put a lot of time into this, but but here's the good news. It's getting better. I'm still going through like yesterday I fixed noise rock. There was way too much post rock in the noise rock folder. So, I've now cleaned that up. Um, and then what I've learned along the way is these these mice fingerprints, which is what I'm using. So, like every song gets a fingerprint, goes in the bucket, and then when you give me a song, I try to figure out based on this fingerprint, like where does it belong? Um, that's actually really cool if you take not just the 50,000 songs in my bucket, but take the 500,000 songs currently on SubmitHub playlists. So, I've actually taken fingerprints of every single song currently in a Submit Hub curator's blog or playlist. And when you give me your song in what's my genre, that's what we call like the what's my genre part of submit hub. There's two new sections underneath. One is similar songs which is just a stack rank of songs with matching fingerprints. It's flipping good. Like try it out. And you'll be surprised. You might find music you like, too. I think I need to add a little button that lets you export it into a playlist on Spotify cuz it is it's shockingly good. Um, and then of course what I can do with that is I can find playlisters or bloggers who have a similar taste or a history of sharing that. And then the last step in that equation which rolled out this week is if you submit your song and you say I think it is acoustic folk but I'm picking up on similarity that it actually falls into a different bucket. I can correct your course a little bit even if my genre detector is wrong. Even if my genre detector is saying acoustic folk, I can actually correct course by looking at the songs that curators have shared and giving those guys a little boost. So our recommendation algorithm has been changed as of Sunday. That's when I rolled it out. Um, and now genres are still the most important thing driving the genre match score, but this similarity in songs and presence on playlists, which is on the submit hub playlist finder, uh, is now becoming a strong signal in correcting that, which means that if someone like Dax, Dax is a really popular hip-hop guy, notorious on Submit Hub for often coming along and just like spending a big budget across every genre, he'll send you his song Tagged Metal Folk, whatever. He's a curator on cement hub. I got a bunch of his songs and I don't have any hip-hop playlist. So >> I I think I've rejected him a bunch. Um but his stuff's really good. It's just it would just be the wrong fit for my playlist. >> It's the wrong fit. But we were all talking about it in the curators chat too. Like everyone knows who Dax is now. So >> uh in a way it worked. But but now the model that I have the recommendation engine kind of corrects course and makes sure that you're on track. Um so I think it's pretty cool. One interesting thing with like uh you know this particular song I ran through it. It's it said this song is metal core electronic core progressive metal gent which I don't think you knew what gent was when when I when I uh mentioned it to you last time. I think you you were like what is gent? I think I >> a very kind gentleman who DJs DJ >> maybe was already in here. I don't know if you added it because of that but um it's uh it's f like so all those like seem cool to me. like that that actually makes sense what the song is. Um, but when I look at like the playlist has been added to, I see like, oh, this is like a hard rock songs playlist and this is a metal core, an alt rock and metal core curator and this one says alternative rock and this one says folk rock alchemist. And so like some of the ones where like I I probably would have never clicked to pitch to these people because of how they describe what they want and then I listen to what they have in their playlist and it's like oh it's actually a good fit. Um which kind of helps. That was really the most tedious part about trying to pitch to a curator is like you don't you could just go off of what they say but you won't want to look at their playlist because maybe they say they want new metal and then you look and like all their playlists are more hard rock or something, you know, and and it's not actually even though they say they want it, they don't actually really have >> the best playlist for that. It's just something that they like they've added to their genre list. So this is a nice handy fix for it, I think. Now, Andrew, let me pitch this to you. You have said in the past that you would uh generally carve out 90% of your budget to meta ads. Those are great for active listeners on Spotify, but you still want to get those streams up there. You have to pass that 1,000 streams. Threshold, popularity, and stream count are tightly intertwined. We can get into that later. I can share some of the data I've been analyzing there. Um, but playlists are a good way. Now, what if I could tell you you don't even have to spend time choosing the playlists. We'll do it for you. submit hub packages. Soh >> yeah, >> so I rolled this out earlier this year, but basically the idea is that I'm using this what's my genre and similar song mapping to automate submissions for people who don't want to spend time doing it. So if you want to spend 50 credits doing playlist submissions, but you don't actually want to spend 30 minutes sampling every playlister, I'm going to go ahead and submit this automatically for you with a submit package. And the other thing about packages is that we strongly believe that meta ads are an essential part of marketing yourself on Spotify. And so a submit package actually comes with both of those with no thinking involved. So you don't actually have to get involved. You can set it, forget it, focus on the rest of the stuff. So for me, that's why it's been so important to get what's my genre in a better place. Even though people aren't happy with the results, I'm looking at it and I'm going like, yes, that like those subg genres are right. they're going to get to the right curators because curators care. Um, and then the similar song matching is going to also do a good job of getting that. So, we're we're kind of getting to this point where like we can do we can we can offer the same service that someone like Playlist Push has been doing for for a decade now where you give them $500 and they take care of everything for you. We can actually do that at scale >> and we charge five times more than Submit have charged. >> They charge $20 per curator. They take the majority of it and they own half of their playlists. Um, >> shots fired. >> No, I mean, George is a great guy. They they uh they've had a very successful model doing it. And I think a lot of the success and the reason that they can charge $10 or $20 per curator is a you don't know what's going on behind the scenes. And b, they take care of it for you and deliver those results. So, if you're a record label or publicist, you've got $10,000. you're focused on like your Tik Tok campaign just like here's $500 take care of it for me. >> That's kind of the the niche it fills. Yeah. So >> submit a package >> being a curator on you know years ago when I was act on both and behind the scenes it literally basically is just submit up behind the scenes. It's just they're pitching it to the individual curators instead of the person submitting it to it. At the end of the day it's just a curator listens to it, approves or rejects it. Um, and and they have a game system to incentivize you in certain ways. And if anything, I found their game system to be kind of like not the best reward-wise because it was like the more points you get, the more you get paid per submission. So you kind of arbitrarily increase your pay rate by just like saying really really nice things to artists in replies because the the person would get to review your feedback >> and if they said that you were like the best feedback you get a ton of points >> which then meant you were going to get paid more in future submissions. >> So the happier the artist >> and the best feedback is not the constructive feedback it's the it's the most flattering feedback. Right. >> Exactly. So I I always felt like their their reward system was just misaligned with what the actual goal should should be, which is like having a song get an accurate fit, you know. Um >> yeah, >> but anyways, yeah, I mean, I did a video on on comparing you you guys versus Playlist Bush years ago, and um the moral of the story was like, hey, these both will get your songs on playlist, so one's going to charge you five to 10 times more, but you don't have to spend. On the flip side, the curators on Playlist Push are much happier with their pay rate per song, cuz who doesn't want $10 or $20 to uh copy paste your feedback without having to listen? Um, more shots fired. Did you know that SubmitHub, to my knowledge, is the only submission platform as of 2026, which is ridiculous, that uh actually requires curators to listen to a song >> and does not let you copy paste feedback. >> Did you know that? Isn't that crazy? >> I did not. >> Yeah. Groover has no way to double check if you actually hit play. There's no there's no mechanism for them to tell that you listen to the song. You can still copy paste. So, if you want to just put your chat GBT text in there, go for it. Uh, no recourse for that. Uh, and I believe all the other platforms are the same. It's so simple, dude. Like, it's so simple. Like, you build your player in here, you track the duration. >> Anyway, it's like I've had it for 10 years, but these guys go around pitching this. >> I'm in a feisty mood tonight, aren't I? But like, >> yeah, >> come on. >> I guess so. Yeah, that does seem like a particularly uh you know, you're you're one you're a si you're a single developer and you've been working on this. >> That's maybe the that's maybe the secret, right? Because if you've got Groover has five full-time developers last time I heard uh and and virtually no features get rolled out and part of the problem is likely that they're constantly discussing or planning features and and also they're disconnected from the customer support. So like I'm the only developer but I'm right there on all the customer support. I've got these chat rooms on Submit Hub where people can >> raise concerns, make suggestions, >> uh get angry at me, whatever they want to do, like I'm there and it it helps grow the platform and and I can learn from that. So, >> I will just just because um it just overweigh the uh negativity on Groover. I will say one thing they do nice is they've managed to kind of enter into like different local markets and have kind of built like little communities in certain places and there's certain like they the they there is a lot of overlap for curism too but there is a a difference in the two. So like there's certain types of things where Groover makes sense. Um, not to say that all the negative things we've said about them aren't true, but just to give you some positives of Gruber as well. >> I think they've done a very good job with their community development and championing music. Uh, they they Yeah. Um, I'll tell you one thing, we've got quite different from them. We have systematically been reducing and getting rid of radio stations on SubmitHub. Uh >> oh, interesting. >> Yeah. So, >> I've always found >> and we've been doing this for years now. >> Yeah, you can go on Groover and get 30 radio stations in Brazil to approve your song, but when you dig into it, none of those are actual radio stations. They're just people who opened up a like a a channel on a free radio website and they cycle songs in the background that no one's listening to. >> And that became super pervasive. I don't know if Groover's taken steps to crack down on it. It's a really good way to hand out dopamine hits cuz these guys literally will just say yes to everything, put it into rotation, and never think about it. But, uh, like I've had customers come to me and be like, "Yo, Submit Hub is I got rejected by everyone, but Groover, I got 50 approvals and I'm like all over radio in Brazil." And I'm like, >> "Good luck, dude. I'm glad you had success because that that's actually not going to to do anything." And so I mean this gets to another point with submit hub is it's really important for us >> to deliver something of value to artists because if we don't it means our company's not valuable. Uh so so that's we've evolved over time like the meta ads doing that stepping away from the influencer stuff because it wasn't delivering value. Um, I don't know what's going to be next, but like the point of what of everything I'm trying to build is to help deliver value. And I'm not doing this altruistically necessarily. I just know that like from a business sense, if my company delivers value to artists, my company is valuable. Like no Um, but the proof is in the pudding. You you actually have to do it. Yeah. And so so that's what I >> that's overall what I've found over the years too is like if if you just like if you actually just try to do a good job and genuine like genuinely try to do the right thing it typically leads to getting more business I found. So >> sure >> you know if >> eventually you get some haters too but >> yeah. Yeah. Of course no matter how perfect you are you're going to get you're going to have that people where campaigns just don't work out. You're going to have people >> and we're both perfect Andrew. Let's be honest we're great. You're both perfect specimens of beauty. >> It's not us. It's your music. >> No, I'm joking. I'm joking. I'm joking. >> But it's it's that's a it's never the music. It's never ever the music, of course. Um but that, you know, that's what it is. If you do the right thing, people come back, right? If if if your product is valuable, then people will see that it's valuable and come back. Even if they leave and try a bunch of other stuff, they will eventually then come back if you actually have something that's cool. So, your incentive is to actually make a product that's useful and and like one kind of uh proof of this over the years, I think the first time I talked to you, >> um I think the average approval rate on Submit Hub was like 15%. And we I used to talk about how if you're getting a 10 to 20% approval rate, you're doing great. Like you're above average. >> Now it's 30, >> right? So all that's I mean this is years of progress. So, this isn't something that happened overnight. I'm sure there were jumps along the way, but like all the little things you've done over the years to make the transparency better, to make it easier to vet playlist, to kind of prevent fraud and and all this other stuff. Um, it's doubled the approval rate. I'm sure you have actual metrics for what it was back then, but I remember it being around 15 and on your site right now is it's 30. >> It was lower than that. Uh, I mean, blogs are still notoriously picky. If you're a blogger and you're getting 30 songs a day, you're you're not going to be able to cover more than two or three of them. Um, one of the bigger changes was the genre detector we touched on. So, like making sure that you get help artists tag their genre correctly means that >> curators receive songs that they like and that 30% watch it, watch it, it's going to go up quite a bunch over the next month. two reasons. This uh similar song fingerprinting thing we chatted about, but uh the AI slop is gone. And I've look, honestly, I've I've had I've had about five curators reach out who are quite concerned. I don't blame them. Like, it's a it's a part of their revenue, but they're concerned that they're going to lose 20% of their submission volume because of the AI songs no longer being present. And I pointed out to almost all of these people, they have like a two or 3% approval rate for AI and like a 40% approval rate for human songs. So >> watch that overall number go up. But it's actually I think and I hope it'll be better for these curators too because that 2% is pulling down their overall all approval percentage which makes them look less attractive. And when artists see that number start to come back up. Um, I think if you're an artist now, like there's a lot of awareness on Submit Hub that before you just select everyone, take a look at how often each curator is actually sharing. If you're going to spend five credits to submit to someone who only shares 5% of their songs, think about it like, is this really is it worth it? Sometimes it is like the potential upside if you do get that 5% is huge, but then hedge your bet and make your next nine selections all have an average rate of 30%. So, I mean, that's that's kind of one strategy to go for, >> right? Right. Yeah. Every time I I have used Submit Hub to pitch to playlisters and particularly I guess blogs do, but I I I usually start off by filtering and saying like I don't want the people who approve 80% plus and I don't want the people who reject um 80% plus. So I kind of want the middle. >> There's there's a filter in the sidebar for that. So it's a great way to do that. Yeah. >> So I just do a quick and I take the middle because like that way it's like I'm I'm not getting super picky people. I'm not getting this people who to take everything and so because I I don't want to spend, you know, like 20 credits pitching to a pipe dream, you know, and maybe sometimes I'll go back and I'll do a second pass and it kind of depends on the song and the budget and all that. >> Um, but like if if you're trying to pretty conservative, you you want to like not >> go for the impossible ones. >> I I usually suggest starting with five submissions just to dip your toes in the water and see what the reception's like. uh using your parameters of course like from 20% to you know 80% and not over in either direction but if those first five don't bite then you got to reassess and go okay you know maybe submit's not the right place for the song maybe I got the genres wrong maybe I targeted the wrong people if four out of five of them do then you're like oh sick all right let me just double down on this and we do this with our packages in the back end so if you give us 50 credits we don't blow all 50 on the first submission we take 25 of them >> and we contact six or seven curators on on the sort of what we call balanced level a mix as you describe it from that 20 to 80 and depending on what comes back in terms of results for those first seven or eight we will then decide whether we go back with a easier approach or a harder approach and when I say harder I mean like let's go for some of these big boys because this song has demonstrated that it's above average >> um so that's a strategy song is like getting an above average approval rating for like the middle chunk, then maybe this is one that can actually stand a chance in the pickier people. >> Exactly. Yep. >> Makes sense. >> There's a lot of pseudo science, but the end goal here is to like I want to get the right songs to the right curators. Uh so that that everyone feels good about it. Um but there's a lot to balance here too. Like so the curators now are worried that the targeting is going to be too exact and too precise and as a result they're not going to get as many submissions and they have to maintain their ad budgets to keep their meta ads running. So there's pressure from these guys uh and I understand it. I know where they're coming from but uh they feel like they are uh at risk of losing out on the submission volume because suddenly everything is too good. They can't say yes to everything and they're not getting as much. It's like there's so many dynamics to balance here. I mean, every day there are are around 20,000 unique registered accounts active on SubmitHub and I think we're doing over a million visits a day now. So, these people are like >> come in here all day. So, so like you're balancing all these different >> perspectives. Um, and it's uh I need a holiday. That's what I need cuz it's it's like start I think I'm losing more hair. You just >> I do this. I go I go like I go three months burnout. I live my life on my school's uh my kids school schedule. So like when they get school holidays then we're like, "All right, let's go." Um so either we do a road trip or we go like somewhere warm. >> Um yeah. So I've got uh two more weeks of burnout and then uh then I'll be on holiday. >> Nice. Where do you like I would imagine a big part of SubmitHub in terms of like credits that are coming in are going specifically to Spotify playlist and um I think you briefly mentioned earlier I don't know if it was before we started recording or after that you don't know if like is is Spotify going to keep this whole playlist thing going forever or is there going to be a time when they kind of sunset third party playlist because they've been pushing their algorithmic mixes and some of them are like crazy specific. It's like this is like your dinner core mix or whatever, you know, and this is like your specific progressive metal mix. It's not like an editorial, it's an algorithmic mix thing. And >> and um you know, there's been people saying about how they they own larger playlists and all of a sudden they playlist have been completely like just stopped all organic growth ever since a certain day. And um you know, do you how long like do you think Spotify is kind of trying to get away from this eventually? I mean, I think we might have even talked about this a few years ago and it's kept going, but what do you think? >> Your recent videos have pointed out that there's been a bit of a shakeup in the in the in the algorithm going on here. And I think I wouldn't be surprised if you have more videos about it coming up soon because things are changing quickly. Um, anyone who's delved into the Spotify for Artists dashboard has probably noticed that they're emphasizing active listeners more and more. So, and and they actually give you a nice breakdown of what this means. Interestingly, pretty much every playlist is excluded from the active listeners number. The only playlists that are included are the ones that you create and listen to yourself. So, that's it. Even Spotify's their editorial playlists, their recommendations, their radio, none of that counts as an active listener. And logically, that makes sense. Most of that is passive. It's not people seeking out an artist. It's them listening to a playlist. They're looking for a theme and they're not actually engaging with that artist. And so Spotify is looking for these signals in their system about which artists are actually genuinely being searched for having their songs directly accessed and played, getting saved to libraries, getting added to people's playlists. Those are the signals that they're looking for now. And I think a lot of the shifts in algorithm are them emphasizing that. So we're hitting this point where playlists and active listeners are these two different groups. And I think there's two different ways that you can promote and market them and two different purposes for so traditional playlisting I think used to have a bigger role in the algorithm and uh I think that's slipping a little bit but that doesn't mean that playlists aren't worth something still. So playlists are still an excellent way to build up streams. You do get paid for it. It directly influences Spotify's popularity score and there's a little bit of this fake it till you make it type of thing. So, you asked me my one minute like what what's the magic thing? And I said like local is the way to go like be a be a big fish in a small pond. Um, you're not going to get booked for shows if your Spotify says you got 50 monthly listeners. But if you got 2,000 on there, which is not a crazy number to achieve using playlisting, it just looks a lot better optically, right? >> Yeah. >> So, that is one value of playlists. You still get paid for the streams. You are going to get some exposure, but you and I both know and you've done much on this. Like the save rates and engagement on playlists, be it third party or editorial, it's terrible. It's not great. And meta ads are a much better way to push the active listeners. So, we don't know right now what's going on with the algorithm. We don't know where the signals are, but my guess is that Spotify is putting more emphasis on active listeners. And the way to get this is through meta ads. And it's not meta ads to a playlist that you own. It's not that. Like traditionally, we've been saying, you know, a really good way as an artist to promote your music is to put it into a playlist that is themed, >> has a few of your songs, and mixes it with stuff that people are definitely going to like because it works especially well on songs that aren't as strong. Like, put it with a bunch of similar artists who are super famous that people want to listen to. and when your song comes up next, there's a better chance that they're going to enjoy that because they're anchored to something similar, but those playlists don't count towards active listeners. So, the whole idea of running meta ads to promote artist playlists, I'm going to say it's out the window. I think I think Spotify just killed that. Uh, and really what you want to do now is run meta ads to your tracks or your artist page because those are the active signals you're trying to send to Spotify. Yeah. >> I And And where does that leave Submitithub? I I mean, like, as I say, there's two buckets, right? Playlists are still great for streams, and I I I suspect that they still teach the algorithm the associations of music. So, like, if you get put into a b if you make metal and you get put into a bunch of pop playlists, that's really confusing for Spotify's algorithm and they don't know what's going on. But if you make metal and you get put into a bunch of metal playlists, that's a strong signal to Spotify. And and also don't forget that every time these third party playlisters add you to their playlist, that's an active signal. Like the literal I've added you to my playlist. It's just the streams and the listeners, those aren't active. They don't count towards that. So it's interesting. >> Still think they're valuable. It's just I try to keep the playlisting down to like less than 20% of the overall monthly streams. And that's like a generally a cap. I mean, there's exceptions. If someone's like scaling up a song and they want to get as big as possible, we might violate that rule, but we wait until the song's already getting pushed in the algorithm to go hard on the playlisting. And it's a very different goal. I I think like when I'll drop a song, I'll pitch it to creators and submit because I know it's not going to be like thousand streams a day. It's going to be pretty small, but it's going to be very accurate. >> It'll hit you. It'll hit that 1,000 in a month from an accurate group. So, you have to still get past the 10,000 streams for your song to be monetizable. >> Uh, and and I think there's still training in it. It It's a really quick way like playlisters are running good ads. You're you're not going to be a playlister unless you've basically figured out how to run an efficient cheap ad campaign. Um, otherwise it's not sustainable. And um so there's definitely value in there to get the quick streams, but 90% of your budget I think should be going to meta ads. >> And for artists running their own ads, um I do run some of my own playlist ads, too. Even though I'm not accepting any curate, like I'm not selling slots or whatever. I'm not like letting people pitch to the playlist. It's just a playlist my own stuff. And uh the the value of that is um it's cheaper per conversion. So >> it's way cheaper. you have a song and it's not doing well sometimes like okay let me let me put the song in a playlist that has a really clear title for it and then I'll make a really good ad for the playlist and then it cuts your cost per conversion in half. So, it's not as ideal because now it's like a passive listener of someone who's clicking on a playlist versus a song. But if your song's not doing well, it can be a really good trade-off that still helps you and it kind of crossromotes your whole catalog if you kind of, you know, my song's number one, then five, then 10, then 15, 20, etc. It's still valuable. It's just I I usually will do kind of the ratio you said where not only is 90% of my budget going to meta ads but 90% of the meta ad budgets going to songs and the other 10%'s going to playlist that I have >> I would agree with you. Um look if you got $1,000 you don't need to spend more than $100 on playlist. Um you don't need to spend more than $50. I again here I am giving counterintuitive business like advice but like I don't think people should come on submit and spend $300 on playlisting. It's It's going to cost a net that is too wide because we're we're way too picky about the curators we let on. We've got about 1,400 active playlisters. >> Thumbs up. And uh sorry, where's my apple gone? But but like our our >> fireworks. >> Oh yeah, our um our rules for who is allowed on the platform as a playlist are quite stringent. I think we've got one of the stricter rules. I know playlist push also has really strict rules. Um, but it means that you can't spend $300 running a playlist campaign. And I don't think you should. I don't think it's the right way. I think $50, fantastic. Put it in the bucket. Put it in background mode. Use the SubmitHub package. You like go that route. And then run ads. If you've got $200 or $300 for ads, do it through Submit Hub. If you got $1,000 for ads, do it through Andrew. Um, this is where we're and and and we're of course using a lot of Andrew's tips for how to run your ads. So, the Andrew Southworth methodology for running advertising is definitely one of the best in the industry. Uh, but but I mean like you and I sat down and the first thing I said I was like, Andrew, I'm sorry. I feel like I'm stepping on your toes here with SubmitHub ads, but you pointed out we're working in different >> spaces, right? Like you are a boutique. You get to work with Andrew. That's cool. And like you're actually going to get hands-on tailoring of your ad. The submit hub one is a lot more automated. we do pretty well. Um, and and we do actually sit there manually and work through it. So, so especially Tilly on my team, she's fantastic and she hand curates every single ad that gets published to make sure the texts are good, the videos make sense, and she publishes them. Um, she's done more than close to 30,000 ads now manually that she has reviewed. Um, or 30,000 campaigns. That's like 150,000 ads. Um, >> jeez. Wow. Yeah. I mean, when when you first came out with your your ads tool and and already some of these have existed over the years. There's hyped it >> and smart noise then they do a little bit differently and but they've been around for a bit longer. They do it through your your ad accounts and all that stuff which means you have to have some of that stuff set up. Um, and when Hyped it first came out with theirs years ago, initially I was quite worried. And then I, um, but I did, you know, talk to John Gold, who's the guy behind Hyped It, and he was like, honestly, I don't I don't think this affects like an ad agency business at all. I think if anything, this might give you more more customers. And and honestly, that's proven true over the years because people they people don't just magically have a,000, $2,000, $5,000 song. Some people do, but most people don't. They have music that they think is good and they want to see how it can do in marketing and I can't run a campaign for $100 or $200 or $300. It's just there's not enough money to make it worthwhile >> as a business, you know. And so the automated tools are a great way for artists to dip their toes in um either just to try it first or because they just don't have the budget to go bigger. And then what happens eventually when they have a bigger budget, it's like, well, now if I can find someone to make it 5% better, it's actually worth the cost um or make it 10% better or 20% better. And so there's it's kind of evolved into like a pretty clear um thing. If someone hits me up and they're like, I want to run ads, but I only have $300. And it's like, cool. Are you willing to learn this yourself? If you are, cool. Go do it. Right? That's great. I have a bunch of videos on my channel that show how to do that and all that. But if you're not, then it's like, cool. Go, go to SubmitHub, go to Hyped, go to Smart Noise, etc. >> No, no, no, no. Just Submit Hub. >> Just just go to Submit. >> Yeah. Come on. >> Well, there is there is an angle um there is a nuance here that Submit Hub does that the other ones don't do. And I think certain people will think it's um better and other people will say it's bad. And it just depends on what you're looking for, right? submit up one doesn't run through the client's ad account, the customer account, their socials, their pixel. It runs through yours, which >> you don't own your pixel. Yep. >> You don't own your pixel. So, that's the downside. The plus side is you don't need any of that stuff set up and your ads are running anonymously. >> They don't look like they're running through your artist. >> You can connect you can connect your account >> um but it won't run through your ad account. It'll show from your Facebook or your Instagram. >> So, lots of people do that. >> Okay. >> Yeah. Yeah, but you but but it's still going to run from our ad account on our pixels. >> Um, and we've got about >> 30 different to ads and they specifically they're like, I don't want this to run through any of our stuff even if they do have it set up because they're like, I don't want people to think that we're advertising. We want to make it look like a third party's advertising. >> And um, >> one, a lot of people don't have their own pixels or whatever. Or they're like, we we've never gotten good results in ours. want to see if like your ad account is better because like it is true that you get different results in different pixels and different accounts depending on the data. Like we're we just last week we had a campaign running from someone where we got it down to like 31 cents or something. So like pretty good not like fantastic but like a solid result. >> Um and we launched he wanted to like his next campaign run through his ad account so that he could like own his data and all that. So, we set up the account. We launched like a small campaign through his with the exact same best performing audience, best performing ad from the one that we're running in our account. The one in his account is doing a$110 a conversion. >> Same link, same adouch, same audience. The only difference is the account in the pixel. Um, and that's run that three times on your account and three time on on his account because what I've seen is that sometimes meta is completely unpredictable. >> Yeah. Um, and there's no recourse, of course, like they're not going to refund you or apologize. >> Um, >> that's what we've started doing sometimes is like just kind of roll the dice again, just like publish the campaign again. In this case, what we're going to do is we're we're adding on some of the other audiences because the other campaign is doing better, but it also has the benefit of the entire history of its learning >> and our pixel. So, like what we're going to try to do is like let's get some of the ads and audiences that didn't work best in that one and add them to this one. Another factor is just time. Ours has been running for over like $1,500 and his has been running for $100. So the amount of data there is >> it's just more info. >> But that's why sometimes running it through someone else's account is actually better because you you your account has probably hundreds and hundreds of thousands of conversions in your pixel. Um warm up process >> a day. I don't know what kind of volume >> per day. Um I don't know. Submitub links are somewhere around 2 million uh views a day. So >> god >> um yeah it's it's running it's running pretty pretty strong at scale. >> Where is submitHub like hosted? >> Excellent question or is it >> No, it's ultimately on on AWS. Um so a lot of the stuff is running through CloudFront which is their content delivery network. So, one of the reasons that there's that things are pretty quick is that when you access them, they're going to be local to you. Um, and then another one is just the efficiency of code. So, I've always like one of the reasons I created submit hub links was I I opened up toned in and it was bloated beyond belief. And then I opened up hyped at it hyped hyped it same problem. Like they're loading a Google Analytics library, the full Facebook one, the full Tik Tok one, the full YouTube one. They're loading a whole bunch of like embeds and iframes and and I remember one of them had like 18 megabytes to load a landing page and I was like hold I I bet I could do this in like 30 lines of HTML and so I did and now it's more like 60 um because we have added stuff on we we can do Tik Tok pixels we can do conversion >> cappy stuff on Facebook like it's all there but it's still quick as lightning because I'm not loading a float on it. Um, >> a couple years ago when when hyped it first launched conversion API, >> you could find the the conversion API token on the publish page >> which was like it's not there anymore as far as sorry, but um it was like when it was first new, I remember like looking through the source code of the site and I just like I searched for Cappy or token or something and it was like conversion API token like oh yep that's my token right there right in the public site which is like the biggest no no ever. Um so whoever they had developers >> there >> um >> great time to go take a while every hyped it link. >> Yeah. Yeah. >> I haven't checked in a while. I'm I'm pretty sure that they've they fixed it. But now I'm going to double check it or edit this section. >> I don't know what John's up to these days. It's actually I feel like it's been a little bit quiet. I haven't heard too much coming out of the hype that camp. >> Their recent their most recent thing I know of is they had a big kurfuffle with SoundCloud. SoundCloud changed their API where at least for them they like uh made it so they could no longer require repost, likes, comments in exchange for like a download gate, which is like a huge thing a lot of DJs use download gates for >> withity. So now it's turned into a promise where it's like, "Hey, repost this track to do it and you could just click repost and close the browser and it would think you did it." >> But reposting killed SoundCloud anyway. I I feel like that was that was the downhill path was reposting. It really was the end of SoundCloud. Reposting. Gosh, I remember that. That was like 2010. I remember those days. I used to I used to rely on SoundCloud. It was so cool. You could follow the artists you like and on the homepage you had a feed of new songs by the artists. And then they rolled out this reposting thing and suddenly your homepage feed just became repost after repost and you're like where are the songs by the artists that I like. >> Um but that's a different subject. >> So >> yeah I mean Soundcloud overall I I think is just a dying platform trying to stay relevant. >> Yeah. I I I was going to do a big push on Soundcloud last year or this this year. last year. I think it was last year um with my band kind of coming out with their album. Um and I was like, let's try let's try to figure out like what is a good like ad strategy for pushing SoundCloud and I was going to do a bunch of experiments, make a YouTube series. And so I I did a paid account on Soundcloud because SoundCloud makes you pay if you want to be able to upload so many songs and do all these features and artists um and get paid for your music as well. You have to pay a subscription to get monetized on SoundCloud, which I've always thought was And so I paid and then I upload my song and uh monetization rejected. You need like a license or whatever. And I'm like I own the rights to this. So I make my own document that's specifying that hey I am in the rights holder blah blah blah. And documentation denied. And I tried 10 times contact support. Um after support request three months later they get back to me and like do you still do you still need help with this? Reply yes if so. >> Um just like basically I replied yes. And then after 5 months go by, they send me another thing. Um, and uh, at that point I had already tried to cancel because I was like, I paid for this, you didn't give that to me. >> And so they they ignored that. So I just called my bank and I was like, >> you know, I paid for this thing and they're not letting me do this thing and they're not replying to my emails. And the bank was like, okay, cool. >> We'll refund it. And then um they finally got back to me five months into it and said, "We'll only look at your request if you reenable your paid account." I was like, "Fuck you guys." >> Okay, cool. >> There's no And if I'm getting this level of treatment, >> how bad is it if there's any because like I I literally know knew I had someone who I knew at the company who I even emailed directly to be like, "Hey, like anything way you could do this." And like no reply there and I even added them on social media. So like if I can't get any help and I have the audience I do what is random Joe Schmo artist gonna have as their experience you know so I just wrote it off entirely and they were already kind of dead in my book and now they're just dead dead in my book. I had a funny experience with some of their executive team who contacted me for a meeting and they said on that meeting we're thinking of launching a playlist submission platform but we wanted to know first um how much is submit worth >> and I remember >> I I was in a cheeky mood and I was like doesn't really matter you guys can't afford me. Um, and they that they ended the conversation quite abruptly. They never did launch their playlisting service. Um, but you know, bless Soundcloud. I owe them a lot. They uh were a fantastic platform 15 years ago. Um, and and the model just wasn't able to keep up. So, I feel like we're digressing. Who cares about SoundCloud anymore? Come on. >> What's coming after Spotify? What's next? >> That's what I was going to say is what's what's next for uh for Submit Hub? I don't know if you have any kind of road map or what you're thinking about the broader music industry. I mean, >> I have some thoughts on where things are going, but I'm curious to hear where where you think where things are going. >> I am still a big believer in local and live music and interacting with that. I don't believe in any of these uh startups that will let you pay to watch a live stream of someone playing. I think that's silly. Um I am not sure like Spotify is a commodity now and music discovery in a way is not what it was. It's different. I don't know what's coming next and and I'm not sure if you know back in the Soundcloud days there was like oh this the space is changing really quickly and and things are going to be different and and I'm not sure what's going to happen there. Uh there's like a dystopian future where you know your chat GBT headphones give you the perfect music for exactly what you want right now and it's all AI generated hodgepodge. Uh, and as we said and agreed and shook hands on, we're going to work collectively to end the AI slop. Um, um, don't worry, we didn't shake hands. Um, the >> like I don't know. Look, in in in some ways I sometimes feel a little bit like uh Zuckerberg here where you almost sit back and watch for the trends and then when they're emerging, you pounce on it and and do it. Um because I don't know what's coming. There's a lot of really smart people in this space right now running really cool products. Uh I mean you and I often mention Aaron from Artist Tools. He's young. He's smart. He's a good developer. He knows how to build products. Um there's a lot of other startups in this ad space that are doing some interesting stuff. Uh there's Songso. They're pushing. >> I want to chat to the guy. He seems interesting uh and genuinely passionate about it, but there's a little bit of fear-mongering going on with the whole like bot protection and whatnot. Um >> yeah. Yeah. I I um >> that's the point of a landing page to filter out. >> Yeah. Yeah. I I I get a little bit of testing with their platform, but not enough just time sake. And um >> but these guys these guys are actively on it. Like they are pushing the boundaries of this. I realize I'm I'm turning into a grandpa here. Um I'm in my 40s now. I'm not going to be able to keep this up forever and eventually these young bucks are going to come along and someone's going to come up with a cool new idea that actually that that takes it into the future. So, what's next for SubHub? I mean, I'm going to keep building cool things that I that I hope help artists. Um, you know, right now, as I said, is like the what's my genre, the AI detector, and the targeting and the playlist are a lot of my focus. Um, >> but I don't know what's coming next. Uh, we've got some ideas that we might pivot into, but I'm not going to Yeah, >> this is not the time right now. Um, but it's nothing exciting. It's more about like we have to start thinking about hedging or rounding out our business, if you will, because because we're in many ways at the whim of Spotify and their algorithm here, right? The rug could get pulled out. Do we become a record label? Do we start dabbling with that? Do we buy like uh cataloges? Do we start distribution? Do we like do all these things? And there's a lot of adjacent models that we are exploring and could explore and and we'll see where it goes. Um I think most importantly I still want to enjoy what I'm doing and I don't want it to turn into a schle or a slog. Um yeah, so >> keep building, man. >> We generally try to do the same thing is try to just diversify the things that we do so that we're not super tied up in only one thing. Um, so we've been doing like more social media services um, over the last like year or so. Offering >> potentially clipping campaigns. We've started offering um, we've been calling slightly different things like running a fan page or theme page for an artist, but also like social media branding help, working with an artist on like their social media strategy for their pages. >> Um, and and also I've been like dabbling more in like a lot more physical media and running ads for merch and and sales funnels for that. And I think as as AI music especially gets more and more widespread, which hopefully is have a little bit of a recession now, but it's was kind of like on this dominant trajectory to just make the internet music world very like unbearable. A lot of them wouldn't be pressing their music to vinyl or wouldn't be doing CDs at all, right? Because >> has a vinyl now, right? You can >> they do, but it's a print on demand. So, it's like you you buy one disc of like of a vinyl for $50. So, it's it's it's much different from an artist going and pressing $4,000 worth of vinyl and putting a lot into it. >> So, I think is like the internet gets more slopfilled. Um it like art it kind of gives artists like certain little safe havens. Like I've met artists who where they have songs they only play live so they never exist on the internet and other artists are doing vinyl exclusive songs as well and having more like stuff that's just for their fans behind gated communities like Patreon. So, I think the superfan thing, which is what a lot of major labels have been talking about too, like how can they get in on the super fan thing, um, I think that's just going to get more and more valuable, especially as like this slop just keeps going up and you stop knowing like what's real and what's not. >> If if you know this artist is real, if you've seen them live and you own their vinyl and you're in their community and you have more facetime with them, it's not just like a a couple of posts on social media and their Spotify playlist. There's the Kevin Kelly's 10,00 true fans where the idea is that if you have a thousand true fans, that's all you actually need to be able to earn enough to live. Um, you know, I do say local and live is great, but I'm in my 40s now with kids and I don't go out to live shows much anymore. So, like that's a very specific demographic you have to target. When I was in my 20s, I was going to shows every weekend. Um, and that was a huge part of it. I hope that's still going on for kids in their 20s, but like I'm out of touch. Um, Grandpa Gishkov over here doesn't know what's going on. And uh, so so like that only works for some people and and those online communities do make sense. Um, do you know what I should do? You know, I don't know if you've experimented about this, but uh like when someone clicks on a link, it opens in a new tab, but it doesn't, but it does. And then you navigate back and that link is still open, right? >> Yeah. >> No one's currently capturing that as like, did you enjoy that? If you'd like, here's more. >> I tried a test a couple years ago. So, I made a custom landing page where what it would do is when you click the Spotify button to go stream, it would open a new tab and that would do the deep link thing and then the previous tab would redirect to an email optin to do a free download for a song. >> Um, the thought was like, well, if they come back to their browser later, >> there's a there's an email opt-in waiting for them. And even if only 1% of people come back to their browser later and and do it, it's like, well, if I'm running this campaign for 10,000 conversions, that's still a 100 people, right? >> And that adds up over time. And so, uh, but problem was when I tested it was that a lot of people are converting inside of at least for ads inside of a social media browser, right? You click Instagram, it opens the Instagram browser, then that deep points to Spotify. they don't keep that internal Instagram browser open like their actual Chrome browser. >> Um, and two, I suspect that somehow maybe wasn't reliably working because there's a lot of different browsers handle like automatic redirects when they're they're not open differently. They don't all reliably they consider it like an automated action and I think a lot of browsers get spooked around like, oh, this browser is like doing two things at once. Like >> looks like a spammy behavior. Um, but moral of the story, no one was signing up. So, those are my two theories as to why. So, but maybe it'd be different with like a did you like a a softer entry or something, you know? >> Maybe the um yeah, the browser the inapp browser is a difficult one to crack and and the deep linking there's been difficult. Um, I think we do it pretty well on the Submit Hub links, but there have been times where it's it's changed and I've had to like panic and people are like, "Oh, the links don't open anymore." And I have to go research like, "What did Instagram change?" Um, and how do I get >> put the full Spotify link in the Spotify link? And I found that that works. So, I I know you guys do some thing where like >> Do you mean on SubmitHub? >> On SubmitHub. >> Yes. That's because I'm doing magic behind the scenes. Um, if I detect that the device is an Android, which it sends in the browser header, even inside Instagram, then I change it to what's called an intent URL. So, it's a specific Android one that Spotify and most major apps have. Anyone can register an intent. So, I could even do it for like the indie shuffle uh Android app. But basically what it means is that it's like Spotify dot dot slash track >> number whatever whatever it is. And then on on iOS it doesn't work the same. They actually tend to handle just plain URLs a lot better. So if I detect such an iOS device then I'll just give you the plain URL. >> Um and breaking out of those >> interesting >> Instagram and Facebook browsers is really difficult because they don't want you to break out. And uh at one point I was even testing where like Spotify's own website like if you opened up Spotify in the Instagram browser they would have buttons everywhere that says like open in app and when you clicked those none of them worked. So like they even broke >> Spotify. >> Anyway, it must be working now because no one's complaining. >> Tik Tok you have to have a business account, right? or uh >> all I know is like I'll I'll go to an artist's page and I'll they'll have a link tree which link tree like one of the biggest link platforms in the world. So if anyone's going to figure out it's going to be link tree >> and then if they have a Spotify link in there and you click it, it just says uh you can't open this link in this browser, go to your own browser and type in the link manually. They don't even let you copy the link. >> Yeah. So Tik Tok's rule I believe is that if you are listed as a business and it's not that you actually have to like pay anything but when you sign up you indicate whether you're an individual or a business and if you're listed as a business then your links work >> right well so you're talking about the account that's posting >> yes like a link in bio like if you want to put it in your in your Tik Tok bio for example >> but but so this isn't like this is people who are eligible for link in bio I mean, they either surpass the 10,000 follower requirement or they register as a business. So, they do have the link, >> but their internal browser still out of it, >> the link tree, >> but when you click the Spotify one, it doesn't just like take you to Spotify in their browser. It actually just says um >> Tik Tok can't open this link, period. Like, it doesn't even take >> one reason why no one's running Tik Tok ads. >> Yeah. And I maybe it works different with ads. I would hope. That'd be kind of egregious if like they just didn't let people click your ad. But like it it makes it so that if if you have a link in bio on Tik Tok, even if your link works, their platform chooses like what platforms they trust to actually like open in the internal browser or and I don't think they let anything leave the browser. And I wonder if that's maybe why they don't even open it is because they know that it it might break out of it. So they just say we're not even going to let you open that link because it's going to break out. Um, but it's particularly frustrating. So, I found like when you're promoting music on TikTok, you don't actually even need um the link in bio at all. Like, it really doesn't matter. I I I did a bunch of tests where I just promoted songs on different like theme pages that I did. Yeah. Thumbs up. Um, uh, it didn't matter if there was a link in bio in the bio at all. Basically like people on Tik Tok are used to either clicking the audio that's like tagged to the song the sound and then like saving it or just like typing it on Spotify and the link in no whatsoever. >> Okay. >> Hour and a half. Andrew, you can you can clip this, cut it, do what you want. >> This is this has been good. You have any uh final words of wisdom for the people? >> I gave the words of wisdom, didn't I? Number one is just enjoy what you're doing and be productive on the internet because there's way too much out there that's trying to capture your attention and turn you into a zombie. So, if you're doing something productive out here, two thumbs up.