How to Start, Scale and Sell a Software Business with Rael Cline

How to Start, Scale and Sell a Software Business with Rael Cline Unknown Speaker: Thank you for listening. Speaker 1: Hello and welcome to the 50th episode of The New Frontier. Super exciting milestone and thank you all for listening and being part of the journey and continuing to give us feedback. Super grateful for it and we have a super special guest joining us for this 50th episode. He is someone who has been an incredible mentor to me personally throughout our early days of e-content and I'm excited to kind of have him on the podcast and share his His wisdom with everyone here listening. So Rael started his career in private equity, but he soon realized he would have more fun building companies. And I guess we'll see if that still holds true with all the experiences he's gone through. He has built and sold multiple businesses, including most recently Nozzle. Which started in 2018 with a mission to both simplify and maximize profits and advertising investments returns for brands on Amazon. And before we bring him out, I want to mention this episode is brought to you by e-content and we help you to optimize for AI-powered search. And go check out our latest feature which we launched last week on helping you see the Ruthless questions on your product listings and answer them in one click. But without further ado, Rael, let's bring you out. How are you? Speaker 2: All good. Thanks so much for having me. And very kind words. Speaker 1: Yeah, no, I've been excited. I mean, it's true. I've been very excited to make this episode happen. Because I remember, I'm not sure how we first connected, but we kind of we first connected Basically, three weeks, three or four weeks after I started the content. Amazing. It was early. And because I remember we, we started it on this like entrepreneurship program. And I was in I remember our first meeting was in that in the building where we were when we were still working there. So it would have been in the first weeks and everything. Speaker 2: It's always the best time to just chat with someone is like literally sort of as close to day zero as possible, right? Speaker 1: Yeah. But um, yeah, excited to talk to you today about I think I call this start scaling and selling a software business. I think you know a lot about that. I guess maybe we can start at the beginning for you. How did you come up with the idea for Nozzle and how do you think broadly about ideation for a software business? Speaker 2: Yeah, I think probably to talk about Nozzle, which is the sort of Amazon focused business, probably its roots are actually in my previous business. So probably we're starting there. That was a company called MediaGamma. This is 2013, I want to say. So 2013, 14, somewhere on there. And this was a company that's focused on what's called programmatic advertising. So nothing to do with Amazon, just sort of, you know, general advertising across the web. And this is a world where algorithms in real time are buying and selling in space, right, by auctions. And I don't come from, I guess, the advertising world or the, you know, any of that stuff, as you mentioned, I come from the finance world. But what I do know, and the idea was, The finance world have been doing buying and selling stuff via algorithms in real time for decades. And so there's actually quite interesting parallels around how you could apply some of those, like the infrastructure or the financial techniques. Don't want to get into too much detail unless you want to. But it could be apply this in the advertising world. And the initial idea was some financial instruments around futures and options and all these sorts of things. And I found somebody else looking into the exact same thing, which was a computer science professor at UCL, University College London, who had written a whole bunch of papers. So long story short is that we wanted to see if we could transfer some of these ideas into the advertising world. Turned out that was not such a great idea after founding a company called MediaGamma. But what we did do is focus on just the algorithmic side. And so we would have much better buying or selling algorithms in these auctions. And so we would have, let's say, large mobile gaming companies that were interested in, you know, showing an ad, getting someone to download their app and obviously spend money ultimately, I guess. And we were just much better at doing that via like customized algorithms than anybody else. And one of the most important things that differentiated them, which enabled us to do our job a lot better was this idea of, we're not stopping at an app download, like here's what the lifetime value looks like, right? Like here's, it takes them three days to spend some money and then 10 days and six months or, you know, tracking that whole trajectory. So that we could go build better algorithms. And so when I sold that business, yeah, kind of 2018, I think it was somewhere on there. It was almost variation on a theme, which is I looked was looking at the Amazon world to say, like, actually, is there something transferable here around this whole lifetime value idea, and the Amazon ecosystem It's pretty sophisticated in some areas, but at least in this LTV, it was definitely a gap, right? And so this was really interesting where we could have the actual data to go construct a very, very important metric that certain types of brands would value very highly and, you know, kind of run their business according to this sort of metric. So the ideation part answers a very long-winded way of answering your question. The ideation part was sort of germinated in the previous business, but I think like to answer your The question more broadly around ideation in general is that I think what's very, very underrated is this idea of taking ideas or methods that have worked in other domains and then sort of seeing whether you can transfer them, right? And so there's like what we did in the beginning here is to say in the finance world, it's sort of quite a lot of precedence here. It might translate, it might not translate or certain things might translate, but that's sort of go investigate that, right? And so getting and kind of also go like, I think generalists in general, in general, but are kind of underestimated, right? Like having that expertise, and so I've never placed a really massive emphasis on, you know, hiring and maybe we can talk about that later, like having specific domain expertise, right? Like in some cases it's useful and you need to have perhaps a balance in the overall team, but bringing sort of fresh ideas, fresh perspectives from other industries and domains is massively underrated. Speaker 1: It's a really interesting one because a lot of people, when you ask them how they form the business, it will be some variation of, oh, I was experiencing this pain and I didn't find a solution anywhere. So now I'm going to go off and build it. And that's kind of maybe the most standard kind of path of creating something. But I, yeah, I haven't really come across this so much where you go, okay, this is like a domain expertise I have. And now I'm going to go and pick a different, like a completely new industry and try and put that into there. Speaker 2: Yeah, it's a non traditional route. It's maybe lower probability, I guess, in the bigger scheme of things. But there are two routes to kind of get to the same thing, which is customer empathy, like really understanding their pain point. And so, you know, you can understand their pain point because you've personally lived it, I guess, or to some degree or some portion of it. There are other ways to get there, right? Like I will always be an enormous advocate of basically customer empathy is probably the most important skill set in building products. And so, you know, setting yourself up to do that where you might have some ideas around how you can solve certain things, but really distancing yourself from, you know, the emotional side or being very passionate. I'm a very passionate person. But you've got to channel that passion properly, right? You've got to establish that customer sort of pain points and empathy and we can chat about how you want to do that. But you've got to establish that first and be pretty objective about it. Ultimately, you've got to be passionate about finding the solution to a customer problem, not the actual problem, the solution itself necessarily. And so, yeah, just different ways to get there, I think. Speaker 1: So if you if you start with this process of, I have an edge or like a unique insight in X, and therefore I'm going to find industry Y and apply it. Do you just playing like devil's advocate? Like, or, you know, just asking the question, how do you Then get to understand the customer because by default of that approach, maybe you're saying that you don't know that customer as well. What you know really well is your solution, your new technology, the new process, whatever. How do you then learn an industry that you're going to go and apply that in? Speaker 2: I mean, learning the industry, to answer the last part of that first, learning the industry, there's no substitute for curiosity and grit, right? At the end of the day, you've got to just, it's up to you to surround yourself with all sorts, and it's easier than ever right now, I guess, sort of getting up to speed. You can do maybe quite a bit of it quite quickly, like yourself online, but there's no substitute for talking to people in the industry, not even customers, just people being around and all that. You might have a sense of like the problem space, like a very wide area of problems or solutions. That's definitely not wedded to any one of them. And you've got to like be really honest with yourself around a solution if there's such a problem or there's a better way to do things. Like I've never ever hold a very strong conviction about the way which would go and solve a problem. But I think in general, this technology should Be able to help in some way that's the hypothesis I guess and then it's your kind of job to figure out You know, is this like a hair-on-fire problem for certain types of people etc, etc And like I mean, I've gone through I mean ideas are the easy part frankly, right? Like yeah, I'm through loads of ideas and you throw away 99.9% of them. So like absolutely. Speaker 1: well, I definitely agree with you there like ideas and the easy part, execution is definitely the challenging bit. So, okay, so you have this kind of thesis about like LTV and you kind of, I kind of half agree with you, which is, I think it is relatively easy to get to know an industry. I mean, if you just sign up to 10 events in the industry or five events in the industry, you would pretty quickly get, you can understand the lay of the land pretty quickly. Whereas maybe arguably, you could say, having some unique expertise and insight, sure, like you did, that takes years to build where you could go and. Speaker 2: Agreed. But you would also, I mean, I'm not building any of these companies like solo, right? I mean, there's an opportunity to build teams and co-founders and the rest of it, right? It's like, it's important that as I said, you need to sort of balance between that. And so you're sure somebody else with that kind of domain expertise or experience is going to be super helpful. But I think also, what's interesting to me is, like, you can get up to speed pretty quickly, but you're almost presented with a consensus view, right? Pretty early on. And so what you really want to understand is like, I'm not really interested. I mean, it's good to know the consensus view, but it's not the thing that interests me most. And it's almost like, we'll get onto this, I'm sure a little bit later, but if you're using any of the LLMs, they're pretty good at presenting consensus views, unless you know them properly. So, you know, if you're going to get up to speed by doing sort of vague prompting, you're going to understand that very well, but they're not going to really help you build anything differentiated, right? It's like you need to. Quite quickly get a hold like understand their consensus view and then find the I guess more interesting rabbit holes to go around and find the people who know a lot about those rabbit holes. Speaker 1: Yeah so you have the idea for nozzle which was basically kind of taking expense from the other previous business did that. Did that idea like, stay true for the whole time? Or did it shift around? Speaker 2: No, it's, you know, it's a strong conviction loosely held kind of thing. I guess the idea, the central idea of like, LTV is lifetime value is something pretty valuable for a certain subset of brands on Amazon. I suppose that's held quite consistently over time, right? And it's probably something only in the last maybe six months or so that Amazon's starting to release some data. But I mean, I can chat about that as well. But like, for a long period of time, it's, you know, nobody could really do this. Or if they did it, they weren't sure if they're doing it correctly. Like, you know, it's a pain point. But like you'll go to market and the monetization and how do you capture sort of the value that you're creating, the actions. Like I've always said about LTV, it's an output metric, right? Lifetime value is the result of all the other inputs and stuff that you're doing right or wrong in the business. So quoting somebody a specific number is nice, but you're not helping them really to improve their business if they don't know what to do about it or how to move it. Right? So you're quite quickly getting to like, Those sort of discussions and features and roadmaps around what are the most important levers to pull to move in a certain, you know, direction or whatever that is. And so like that shifts around, you know, daily in the beginning. Speaker 1: I'm going to come back to that question. But just to go with the story, so you have this hypothesis. How did you start the initial custom discovery then? Like what were those initial conversations look like to see if that was actually a pain that people cared about? Speaker 2: Yeah, I mean, so some of it is just naturally thinking about like, what sort of brands would be interested in this. And again, you don't want to rule out everybody, but like, you have a sense that, look, if you don't have repeat orders, this isn't really useful. So there's certainly certain types of products or companies we can rule out on that basis, like electronics or something, you know, stuff like that is probably not the most useful. But you land on, okay, CPG is probably like a core part of the business. And I say this with like good hindsight, but in the beginning, you don't really know these things, which is Not by design, but we noticed that a lot of the early adopters were D2C brands that were running on Amazon. So they're D2C native and maybe begrudgingly opened up an Amazon store as it was at the time. But this was kind of the easiest sell to us because they're used to these metrics. This is well understood in the DTC world, running a business according to lifetime value and customer acquisition cost and payback period and all those kind of things and thinking about churn and all those kind of things. And so the picture was pretty simple, which is... You know, we give you like inside parody to your business, but for the Amazon channel, yeah, so you can really think about like, you know, investments in the rest of it. So that was just, you know, just sort of happened. That's not like something sort of planned, but narrowing down sort of particular like industries. So, you know, anything clearly supplements or health and beauty or food and drink and that kind of stuff. It's like what and it helped me in the early days because It allows you to go deep into like a few areas instead of just crossing yourself incredibly wide and thinking like anyone can use this tool. Like there's just a natural focus point. And it's also quite good because those categories are fairly large on Amazon as well. Speaker 1: Right. Speaker 2: And so we're not talking like a We're going to max out in terms of the TAM or something like that pretty early on because these are pretty, you know, massive brands, massive categories and enough of a wedge for us to kind of, you know, get in there. So, but then, you know, again, there's a real art doing that customer discovery side because it's very easy when you would know this from maybe jobs to be done type frameworks, etc. Like how to ask the questions in a non-leading way so that you get honest answers because people are always You know, for good or for worse, they're always trying to like not hurt your feelings. And you've covered this in previous podcasts and stuff. So maybe like, don't don't have to go into too much detail here. But you know, really trying to like Let them talk as much as they can at the end of the day. And to be a good listener, I think is probably the most important skill set there. And to not lead them into like answers, like, you know, as desperate as I was for them to say something like, God, I wish we had, you know, lifetime value metrics or whatever, you know, it's like, you know, how do you, you know, tell me about the last time you, you know, measured success in the business, right? Or something like that, like, how does that, you know, and you start to understand how the process work and what metrics they're using. Who's bought into it and, you know, all those sort of things. And so, you know, doing it more sort of in that kind of discovery, almost Socratic type way of doing it, right? And so, yeah, I mean, I think that's, you know, knowing that there was enough there in the early days for that to happen. And clearly the biggest thing is someone being able to write a check, right? Words are cheap. Getting someone to pay for it when, you know, you probably don't have anything to show, right? Like that's kind of the biggest article of faith. So we managed to do that, but it's, I mean, we still completely over-engineered The first, I guess, iteration of it, which is you're so desperate to get your first set of customers where you kind of, I wouldn't say over promise so much, but it's, you just give them much more than they even need, right? So like for us, we had a whole Tableau dashboard and, you know, log in for the clients where they can go check all these fancy charts and whatever. We're literally the first, I'd say, 3 out of the first 5 customers, all they cared about was like, Rael, I don't care if you just email me or text me the value, like the LTV and the cap. Speaker 1: Just tell me the number. Speaker 2: Yeah, just give me the freaking number, man. Honestly, I'll never forget that. Yeah. And we were like, you know, all these dashboards never logged in, right? Didn't want to log in. I didn't know what the freaking number was. And so, you know, you could avoid like a whole bunch of stuff like that. But by Yeah, yeah, I totally redo things that way. But yeah, so that's kind of like, you know, the very, the very early days and you kind of equate this We can talk about pricing, I suppose, as well. We were charging a fair amount. It was probably like $800 to $1,000 a month or something like that in those early days. It was a fairly big customer as well. But, you know, you kind of mistakenly think that you have to provide all these things, right, to justify their price. Like here's like a kitchen sink of like all certain ways of slicing and dicing the data, where it's like, no, actually, it turns out an LTV number is worth $1,000 to me. Speaker 1: Yeah, I want to break down all of these one by one, because you covered a lot of interesting things there. So firstly, initial customer discovery, and I agree with you, the temptation And even when we launch a new product or new content, I still have to stop myself from doing this. The temptation is to go, oh, I've got a great idea. Amazon sellers need to know their LTV. And you get on the call and you pitch them. Don't you wish you were on TV? We could do this. And as you say, the natural reaction is to be like, Oh, brilliant, especially if it's a friend of yours, or it's someone who, yeah, been introduced via a friend. So they get just going to be polite. Yeah, great idea. Like, this is amazing. Love it. But you as you say, when you're doing like trying to search for a real problem, you need to, you need them to do it. And you need them to do it without you asking, oh, what's one metric that you really miss? Because like, oh, maybe that is a good question. I know, but you need to, you need to do it in a way that is like, what's one metric which is LT and which is next best in this. You need to actually feel that they care about it. So you're having these conversations and people were saying to you, oh, You know, you're asking, you know, how do you, I guess, like, what metrics are you using to look at your Amazon business? And they're saying, oh, like, we've got sales and revenue, but I wish we had LTV, like Amazon, make sure this is really annoying. And that was happening. Speaker 2: Yeah, pretty much like that. You know, there might be a bit of prodding about how do you think about repeat sales and their impacts on your business or, you know, things like that. But yeah, you generally want to start like very, very, yeah. Speaker 1: High level. Speaker 2: Yeah, high level around. How do you judge success for your business? Let's break, I don't know, profit growth or sales growth or whatever sort of mode they're in at the time. And it's like, okay, then how do you, you know, how do you think like, what are the drivers behind that? How do you increase your profits? And then they get into discussions about repeat versus New customers. Okay, that's interesting. So what do you know about your repeat customers on Amazon? Speaker 1: Right? Speaker 2: Have you ever tried to look into that? And it's like, well, turns out not much. And you know, like, it sort of evolves in that. Speaker 1: That's the thing. You also want you also need to be looking for behavior where they try to do something. Speaker 2: Yes. Speaker 1: Right. You need to be listening. What what have we tried to do and failed to do? Because if that if you're, doing something completely new, and they've never even thought of the problem, then it's a lot harder. Speaker 2: Yeah, so I mean, we would In the beginning, yes, absolutely, right? So like this is almost the equivalent of bottom-of-the-funnel. It's like customer acquisition, which is people actively trying to solve the problem, but are struggling somehow or not confident in like their results and you can help them out. Then there's, as you work your way sort of higher up the funnel into, you know, people who aren't even problem aware. Like we would call it a problem aware and kind of marketing speak. And this is, these are people maybe running much smaller businesses doing a million or something like that on Amazon. And they're doing very well, and they've sort of cracked Amazon in some way because they're doing quite well. But then it's like, okay, how do I get from 1 to 10 million kind of thing? And yeah, I mean, now you've got to start breaking up your growth into these separate components and pulling different levers kind of thing. So that's kind of the educational piece and problem aware kind of stuff, which is a slow burner. But like when you're just starting off, you want to be solving the problems for the bottom of the funnel, people who are trying, you know, actively trying but failing to solve it. Speaker 1: So, how many of those conversations did you have before you built any product? Speaker 2: That's a good question. Probably about, I suppose, like 10, something like that. Speaker 1: 10 conversations. And people are meant to bring this up. So you're like, right, my Harkoff's, this is right, people care about this. I mean, it's kind of an obvious thing. You know, if you're a serious business, like you want to know that number. Yeah. So then, then, yeah, how, how did the first, you kind of mentioned it, the Tableau dashboards, but like, how did you build that first product? Did you have orders, pre orders? Or like, how did you, how did you do that? Speaker 2: Yeah, so I didn't do pre-orders. I mean, this is, you know, it's a B2B proposition, you know, it's harder to do the, I don't know, waitlist pre-order stuff. So, yeah, I mean, this is but I was very conscious that I needed effectively a signed contract. Yeah, you know For revenue. I mean, I wasn't fixated whether it was I don't know $200 or $1,000 depending on the site like that's you know, those sort of monetization issues I sort of kick the can down the road, but it was very important for somebody to demonstrate that we have to say So I wasn't fixated on the amount necessarily But again, you get the usual, you know, the usual things like, you know people drop out and you know the rest of it so You know in the beginning and also, you know, it was a very small team to begin with and So you don't want to be running sort of 10 to 15 of these. We set them up as pilots effectively, right? And so, you know, because we wanted to like really understand the customer properly and have regular calls and feedback and time to fix things and add things and you know, that kind of stuff. So 10 to 15 is like, I guess way too much. So we started off with I think it was three to five, you know, to begin with there. And yeah, I mean, some of those I think I'd like to think most of the I think most of them did eventually sign up at some point. And one one was an agency as well. So like there's different dynamics there. And then you're getting into discussions like, I have no idea how agency pricing works here. Speaker 1: Like, what were you selling? What was that early product? So you got the first Yeah, you know, you've got people that they said they have the problem, then you sign five pilots, you're charging them $200 to $1,000. Like, yeah, I agree with you. There's always they're paying this stage. It doesn't matter, really. What what are you what are you giving them? Speaker 2: Yes. So the first of the most valuable thing that we've always done is the cohort analysis again I realize we're somewhere into this conversation says about 24 minutes I haven't really described what nozzle does properly but anyway. So. It was effectively a cohort analysis and frankly, in my view, remains the most important thing. And just to back up a little bit, what is an LTV cohort analysis? It's a way of isolating people and tracking what they do over time in the simplest form. And so what does that mean in kind of the Amazon world is you can say, okay, everybody who bought from me in December 24, let's isolate them and track what happens to them in terms of repeat orders. And, you know, their profit over time, like a customer, customer lifetime value I use interchangeably with profit over time. So people who first converted in December, how profitable are they after three months, six months, 12 months, whatever it is? And you want to be comparing that to let's say people who first converted in November or October or whatever other sort of cohorts they get segmented into. And the reason why this is so, so important is if you make basically any change on Amazon, you change your price, you're doing a remarketing campaign, you're doing coupons, like anything, How do you know whether it works? Well, it's not really an A-B test because an A-B test is like a fixed period of time. You kind of see what happens. The only way to figure out for brands that have a high number of repeat orders, the only way to really figure out whether these changes are working is to track people who were exposed to them, who bought on that coupon or bought again via their remarketing, whatever it is, and track what happens to them. Are they more profitable over time than your baseline? And that's the cohort analysis. And so like a lot of the big questions that this LTV cohort analysis is answering are number one, if you know how profitable a customer is over time, you can answer the question of how much you can afford to spend to acquire a new customer. So that's an enormous, in and of itself is an enormous thing because it has all to do with your ad budgets and how you want to split them and how focused you want to be on sort of net new versus returning. So there's a whole sort of decision tree around that. But the other thing is this idea of understanding what changes you're making by actually working and driving growth. Like this LTB and cohort stuff is the thing that we landed on. I think quite early and as I say, to this day, I think it's probably the most valuable. Look, it requires quite a lot from the customer to understand what it is, how to use it, all that kind of stuff. So I think one of the things we've always struggled with was that education piece. And so outside of that very early adopter of We're a D2C brand, we live and breathe this kind of stuff every day on D2C, like we know what a co-analysis is. You know, there's always like, a bit of, certainly a bit of friction on outside of that, like early adopter, I don't know how to use it. Speaker 1: But so, but this, this data isn't available in Amazon, hence why Nozzle existed. But how did you surface that to the customer? Like, in an early, like in an early pilot of the tool? Speaker 2: Yeah, so it was a Tableau dashboard. They would grant us API access and so we would get the data directly from Amazon via the APIs and we would go build, you know, you would ingest the data, you store it in the database, you write the SQL queries and you produce it in a Tableau dashboard. Speaker 1: I see, so you're building, so the product is like this analysis of looking at when the purchase come through and doing like some algorithm to decide to give them like cohort analysis. Speaker 2: Yeah, I mean, it's algorithms probably being a little bit generous. Honestly, it's a SQL query, which splits up, which does the code analysis. So like, you know, we can talk about where you would add some sort of machine learning or whatever. But in the first place, no, like it's complete overkill. Like, again, it's very easy to throw a kitchen sink. But when someone wants an LTV, analysis, you can do the SQL. And so we just throw that in the Tableau dashboard. And the other thing they wanted to be tracking is the CAC. So LTV over time, CAC, and the number of new brands, like those are the three key metrics that people want to care about. And one of the things I think we had to think about early on is, these are very slow moving metrics, at least LTV is, meaning you probably only need to check this maybe monthly, but probably quarterly. Yeah, right and so why should people come back right like what you sign up for a trial? Get your numbers. Thanks. Maybe I'll sign up again and whatever you know six months or something. Yeah, and so what's the reason to come back and use it? And it was basically the cohort analysis was the reason to come back and use it right because again How do you measure it's fine? You know what some baseline? LTV is, or CAC is, but it's the direction of travel that's important for your business, right? Like, how is it shaping up quarter on quarter or something like that? So, we certainly had a bunch of, like, sign-ups and instant cancels sort of behavior, but really had to think about retention in our own sense. Speaker 1: You build this like MVP, SQL query, got the customers, and I assume everything is going well, people are liking it, they're using it. So how did you kind of like ramp up the, you know, go to market from there? Speaker 2: Yeah, I mean, this is interesting. Again, it's easy for me to say with hindsight around a bit like what we were producing effectively are proprietary data sets. So whilst the data sets will always, you know, always belong to our customer and all that kind of stuff, the aggregated data sets are interesting. Meaning, one of the most common questions you would get from, again, a supplement company to pick on them is to say, cool, I understand all the stuff around my business or around these key metrics, but actually, how do I compare? Yeah, what is the category average here? And if something's gone up or something's gone down, is it something that's happening at a category level? Or is it like, you know, to do with my business kind of thing? And so we would effectively generating pretty interesting proprietary data sets there and being able to then benchmark performance on these metrics. So, you know, we could say that, you know, typically for a, yeah, sort of a drink company or whatever the sort of subcategory may be. This is kind of where you want to be, right? Like this is what maybe average looks like. This is what 90th percentile looks like, both in terms of lifetime value growth, like after three months, assuming your purchase cycles are, you know, 90 days roughly. And what's going on on the CAC side, right? Is it becoming just more expensive to acquire a customer in your category? So there have been loads of metrics that are like CPC driven, which is useful but limited. What you really want to understand is How expensive it is to acquire a new customer on Amazon. And we can provide an industry benchmark or index on these things. And so that helps with two things. That helps with retention. For our customer base, but it provides an incredible marketing opportunity. So you told me again to go that long way of answering your question, which is like how do you attract more customers? How does how do you get your sort of marketing machine going is that we could release some of these like in indices or yeah, I'd agree data and Then you know be the thought leaders in all of this stuff right and say like find out so it's a great lead method because I Now, find out what your CAC is, like this is what it is for your category, like sign up here and find out kind of thing, right? So, I think that was like quite a nice early, not by design, right? As I said, it's just something, you know, you kind of, yeah, one day think about it and like, okay, this is, these are interesting data sets, we can use them, we can use them both for retention and marketing purposes. Speaker 1: Yeah, and I actually remember we talked about this a year and a half ago. I don't know if you remember this conversation, but we were talking about this and it was literally like over the week after the GPT store launched and you were kind of describing like, oh, we were doing these, you know, like PDFs and whatever. I was like, and I was like, oh my gosh, I should go and build a GPT of how people are doing for Cosmo, like a free thing. Yeah, like, as you're saying, kind of like, if you have an area which you are an authority on, and you have some kind of proprietary data, which we do, like, you know, chuck it into a GPT, Send it out to everyone and got in all of the newsletters, like Kevin King talked about it, like John Depp. Speaker 2: I remember that. Speaker 1: Yeah, it was so new, right? Like, it was like, everyone was talking, you know, like, it was when like, the GPT store hype was like, at its top. And it was like, this is gonna kill the iPhone and whatever. And, and yeah, like, it was a really good way of getting people to just like, give them value for free and easy. And also like have a little link back to the content at the end, like, okay, your customer score, For like 180 if you want to generate like all the optimized visual written content and do it at scale. There you go. Speaker 2: So yeah, like, yeah, I mean, I'm, I think, you know, I, my DNA is more, I guess, data driven. It's just kind of the way I look at the world kind of thing. So I'm always looking for like data. Way Yeah, to, you know, to differentiate and create sort of lead magnets and stuff. It's just one way of doing it. But I think I think your app ranked quite nicely as well or something. Speaker 1: We are. Yeah, we're like, you know, if you type in Amazon, we're in the top 10, at least last time I checked. Yeah, it was on the GPT score, which is, which is great. So you built these, you know, you've got this data. And I think this is yeah, you're bringing, you know, fundamentally, you're bringing value to the customers and the community and the industry because you've got this insight. You're giving it away basically for free, but then also people understand that they can come and use nozzle and and get like their own data and that's, that's helpful. So kind of like that's a broad marketing way like, like, were there any other ways that you're going to market in these? I don't know where we are in the story now. Pre raising money, right? I'm talking pre pre any race. Speaker 2: There was no there was there was a raise. I don't remember the exact amount there was. I can't remember, was it maybe a million in sterling? Something like that, somewhere around there. So there was no, there certainly was raised just because you need an engineering team. Speaker 1: You raised, what point did you raise the money? Speaker 2: Um, well, it was 2020, I would say, somewhere on like, maybe mid year 2020, something like that. Speaker 1: So like two years in, basically. Speaker 2: Yeah, yeah. And somewhere on there. And, yeah, so in terms of The other sort of way to accelerate would be via agencies. Yeah, which is pretty complicated, I guess, because it's just a very different set of requirements and incentives, I guess, from sort of the product point of view. I mean, honestly, it's only in the next month or two that we're allowing multi-user access, which kind of sounds nuts. So someone signs up to Nozzle and says, hey, can we just add another user? We haven't really built the infrastructure for that. So it's kind of like being a single player for quite a long time on that. But I think that also speaks to some of the You know, knowing who your customers are and all that kind of stuff and like trying to just make sure you provide value, more and more value to like your ICP at the end of the day. Speaker 1: You're saying agencies is a go to market because you would approach agencies and they would manage 10, 20 brands. So basically... Speaker 2: Yeah, that's a quicker way to scale. Speaker 1: You're going to go and sell to these agencies and they will sell to the ecommerce brands. Speaker 2: Yeah, but then... I think there are quite a few false starts with the agencies, right? Just because, um, you know, a few things in some cases, like LTV is most valuable when you're doing a profit per customer, not, uh, sales per customer. And so we had a few instances where brands weren't giving the agencies the cost of goods sold, uh, for whatever reason, right? It's not that great. Uh, but then there's also, um, in some cases it could look, make the agency look not so great with what they're doing and stuff sort of declining. And so like, and I would, I'd always be flexible to say, like, if you want me on the call, the client's happy to do it, otherwise do it on your own, like, whatever, that's fine. But I think being too far removed wasn't good. And so a lot of it from the customer wasn't good. Yeah, it's ultimately on me, right? Because I think there needed to be more training on the agency side around, you know, again, like, what are these metrics? Why are they important? And, you know, all that kind of stuff. So we certainly had a few false starts with that. We do have, we did, for sure, had a couple agencies that I mean, to what we were talking about earlier, we're actively trying to do this themselves. And again, weren't confident on the accuracy or the methodology and whatever. And like, you know, that's like an easy thing. But in general, agencies, I think have been sort of, you know, tougher for us. Nozzle always works best when there's talking about sort of our ICP. Clearly, it has like one of these categories is really important for us. But also, it's ultimately somebody with the P&L. Responsibility so it's more like in smaller companies. It tends to still be kind of the founder I guess but in larger companies and we work with some you know enormous brands, you know doing 50 80 million or something on Amazon But it's generally somebody who's got P&L responsibility for the Amazon channel or for marketplaces or something like that, right? Because it's just it's a strategic tool and I think that's one of the other challenges we faced early on is It's not, at least in those days, it was quite rare to pay money for a strategic tool. Yeah. And when we were searching, I should have brought this up earlier, like when we were searching for, you know, early adopters and ICPs and all that sort of stuff, we found people who valued their time over money. Right. And so like, even if this thing is manual and like, yeah, I I just didn't want to like I wouldn't they weren't willing to I guess. Sorry, the other way around, they value the money more. So they weren't willing to part with this, even though I can save them time or whatever. So those tend to be like smaller brands. And that's fine. I could, you know, I totally get that sort of decision. But that was something we realized quite early on, where trying to convince somebody to pay for a strategic tool, and tying the amount, like, I guess, in some ways, not recognizing the value. So they would say that it's far too expensive. And that's fair, right? Like, it could be that the business We're not, if you're just starting out on Amazon, it's not the tool for you anyway, right? Like we need to be more established users. So like figuring out those sort of things early on was incredible. Speaker 1: This is a really interesting point. And it's a point which we've recently faced. So when you're, unless you're kind of like Sam Altman, or someone of this level, and you can just like start a company and immediately you're going to be talking to all the CEOs of the top 50, 150. I think for many of us kind of normal people, The way you start out, you do your ideation, you're talking to small companies, as you probably were, right? The founder really cares about LTV. Maybe you have a 10, 20 person company. And then at a certain point, you want to start to be like, okay, this is a great tool. And I know, actually, L'Oreal or whoever, you know, like big shampoo brand, they're going to want to know this or, you know, the biggest hymns or whatever, like the big, you know, the big supplements brands, this is obviously going to be useful to them. How do you go about working out like how that ICP individual transitions? Because yeah, early company is the founder, easy. And then now you're going to a big organization, so many stakeholders, you know, how do you work out who needs the product? Speaker 2: Yeah, that's a really good question. I mean, I think a lot of that would come from if I'm running webinars and things like that. It's understanding who's joining those webinars. So who's the internal champion? Who's the internal champion there? Like, you know, the buyer is not the same as the user in most B2B, you know, and so like understanding that distinction where the buyer is the user and For the smaller companies, right, it's the CEO that's kind of logging in and checking the LTV numbers or whatever the things are. But that splits. And so like really splitting that into who the people logging in from our tool and I'm building sort of product for them, and how do I make them look good, you know, and sort of help them, you know, progress in their organization or whatever in their careers. But also being extremely mindful of like, who's actually signing off on this? And like, what metrics are being used there? So understanding those, like, as I said, it was like, who's joining, who's joining the webinars, often as well for the sort of mid market, maybe going a little bit enterprisey type customers. We would also do services. So this is another interesting thing, which is like, let's prove the value by doing basically a consulting project, something custom to them, that will use the data and, you know, it's something customer related or LTV related, but it's ultimately custom. And that's kind of a good way of getting in just via like, you know, consultancy project. But it does mean that, you know, as a sales process or sales skills as me, you know, I need to adapt. You're talking to a very sort of different person than, you know, the rest of it. But I think like doing a service type Project is often a good way. Speaker 1: And this is something else we've been dragged into. And it's an interesting one, which is, yeah, often they will want a service or consulting type business. But you're building a software product. And you actually you don't, you know, you don't want to, I don't know about you, but you kind of don't want to, you know, you want to build something that scales that, you know, you can launch anywhere, it's going to become hopefully a billion dollar company, you don't want to start doing like consultancy, which is going to take A lot of the founders time and have a lot of like manual work, which doesn't necessarily scale. How do you manage that kind of? Speaker 2: Yeah, I think every like, you know, if you ask me towards the beginning of the company, I mean, every sort of SaaS founder will almost be ideological about it. Speaker 1: Yeah. Speaker 2: No, it's a distraction. And, you know, it doesn't help with fundraising or valuation or exits and I don't know all those kind of things as well and all that kind of stuff. As you mature as a company and as a person, frankly, the more of an almost dirty secret it is in the SaaS world. For me, most SaaS companies or successful SaaS companies will have a service component to it. As long as it's very related to the stuff that you're doing. For us, it's always something to do with You know, the core cohort analysis that we're doing. Then I think, you know, I think it's great. It's a way of like strengthening. Yeah. You're growing the LTV, frankly, of our own customer base and strengthening the bonds. And, you know, when these people in these organizations, your sponsor in that organization moves on into another company, like the amount of like calls I've gotten saying like, cool, let's do this for this new company that I'm at. And having those personal relationships has been awesome for me. So as long as it doesn't feel too far away from what you're doing, and honestly, it's helped us refine the product as well, right? So like we're getting almost... In some ways paid to test out new product features and the rest of it. So you know what? Speaker 1: Yeah, I don't know. It's fine. I know the business has been acquired. So you can't really talk about some stuff. But as a percentage of revenue, if you can, if you can answer what percentage was coming from like enterprise services, and what is coming from like people just signing up and using it as like a pure SaaS? Speaker 2: Yeah, the majority still The majority is still going to be on the SaaS side. It was always in the majority. Most of the time, it always started off as a SaaS product and then layering on service stuff. For the larger customers, it's a bit of a different onboarding process. I would reach out to them. I'd have a call with them, all that kind of stuff. I'd let it be known right in the beginning that we do offer services. It's not really on the website kind of thing. You know, let's do that. And then I would check in with them every, I don't know, six months. Speaker 1: Services being, we'll build an analysis, present this to you with some like specific ideas for your specific company. Speaker 2: Yeah, often this comes from them though, right? So like often it's so one of the like two common things we would do would be forecasting. Yeah, and I don't want to get in. Well, if you want me to, I can. But like, there's a pretty unique approach to forecasting you can have for businesses that have repeat orders, meaning you can use those cohorts, you project out every individual cohort based on the retention rate and LTV and whatever, that kind of stuff. So you can actually do very neat type forecasts off the back of that. Very difficult to productize forecasts, right? Like ultimately, all forecasts end up in the spreadsheet again. Yeah, um, and everyone's got kind of the unique requirements. So it's like very hard to like, standard, like you can have a template, I guess, but it's very hard to like completely set. And so like, that was super valuable, right? These forecasters are massive, right? Like, it's basically, it's like, this is like board level stuff. Speaker 1: Well, it's like another product. I mean, in a sense, it's like, that is a product in itself, right? For I mean, obviously, it's a custom built one. Speaker 2: Sure. Speaker 1: Okay. So at some point, You're doing all this stuff. You're selling to agencies. You're doing some services to big customers. At some point, you raise the venture money. Why do you decide to raise it? Speaker 2: Yeah, that's that's a very good question. And I think it's like a very timely one just because I think the environment is so different now, right? But like, you know, drinking, I guess, the SAS Kool-Aid and all that sort of stuff. Look, I do I do certainly think that the very traditional SAS, you know, business model is to a lot of like upfront development, right, costly development, but then your marginal cost to serve an incremental customer is near zero kind of thing. And so it's like kind of a classic. Rooted like that. But I would say, you know, it's very hard to get off the VC train, if you've decided to go on the VC train. And so it's kind of a one way door in that sense. Yeah, if you make the decision that you're going to raise venture capital money, they are doing this. Yeah, it's like, you're on a particular path, and it's maybe binary, maybe not binary, but Not many hours. And so it's a very careful consideration that you need to make. I still think it was the right decision for Nozzle. I have no regrets about all of that sort of stuff, but I think in today's environment... There are probably like many other ways to kind of get going, but yeah. Speaker 1: So you raise the money on like just getting an understanding of time wise on an MVP or like you're making revenue? Speaker 2: Yeah, so we had committed, we didn't have the product. I think it was like letters of intent. Speaker 1: Got it. So you raised a million pre product and everything. Speaker 2: Yeah, and it's, um, you know, like in the UK as well, you've got those, you know, SEIS and, you know, EIS things as well. So we also, yeah, uh, this may or may not be the case. I might be misremembering, but, um, we had R&D tax credits as well. So there's like another thing on the tech side, which is, um, for those who don't know, you have a specific, again, it's like quite a, I guess, stringent like application process, but you have a particular project and, Some of it can get funded by the government effectively. So if you couple those two things around, If the proposition to an investor or like an SEIS or EIS investor is every pound you put in has been matched by the government and you've for downside protection you can get what is it? I don't know any money back if you know You know, that's pretty easy Exactly make it easy make it easy for investors. Look, I know there's been quite a Maybe I've laid a little bit more criticism on the EIS and SEIS sort of incentives and all the government better off doing other things, but that's a whole nother discussion. Speaker 1: Well, I mean, yeah, it's a different discussion. I think it's one of the best things that the government does and a lot of the other stuff they supply them to be honest. But that's my personal view. You raise the money and then you're hiring the team. Do you want to talk about how you approach that? Speaker 2: Yeah, I mean, this is always sort of an engineering heavy thing, right? And so like understanding the Amazon datasets and APIs were not good. At that stage, it was the MWS, I think it was called before the selling partners, like documentation is not great, things don't work as expected. Yeah, things take longer to build as a result. And so there's a lot of like back and forth around all of that stuff. But yeah, I mean, hiring, you know, the data engineer and sort of engineering and SQL skills, that kind of thing, building a dashboard. So now we're moving to like a web app. Right no longer Tableau and so there's you know front-end skills that are needed there. So like yeah hiring hiring that team There's marketing that needs to be done. So, you know, although all those sort of things and I think one thing to maybe call out is thinking about the relationship between Machine learning and engineering right and so I think that I sort of put them on like a Horizontally, they're kind of the same, right? They're just different functions, but like, you know, that kind of thing. But, you know, ultimately, machine learning stuff has to fit into an engineering framework at the end of the day. You're deploying something like an app in a production environment, yeah? And so you can do all these fancy things with, I don't know, algorithms and machine learning, but at the end of the day, like, it's still subject to engineering constraints, right? And so thinking, for me, I think, wasted quite a lot of time on Maybe some of the machine learning stuff a little bit too early once, you know, you've got to have pretty good data infrastructure and machine learning like operations kind of infrastructure in place before you go and hire data scientists. So like I think, you know, in hindsight, some of the sequencing I think was was definitely out and which is Super valuable when you're very early stage company and runway matters. Speaker 1: Yeah. Speaker 2: Well, yeah, so that's only one thing I do differently. Speaker 1: I missed the point which I wanted to come back to. Like, you're raising money. So rewriting the story. And this is what you said about Amazon doing it. So now, like, The number one question that I think you get as an entrepreneur when you're raising money, and it's a very infuriating question is like, Oh, but Google could just do this or Mexico could just do this. Well, yeah, they can do anything, right? Yeah, of course they could do it like anyone could do this. You just need to start. It's not like you need an engineer. They decide this or anything. But how did you, like in those conversations and even in your own strategy, think about the risk of like Amazon doing what your business is doing? Speaker 2: Yeah, so there's two things there, two existential risks. The first is like Amazon basically doing what you're doing. The second is API related, like even if they're not doing what you're doing, they can just, yeah, they make a change to the API or they switch it off or do something and then you're screwed. So like those are the two big things. On the second one, which is the API one, we were I'm pretty convinced that at least that wasn't going to happen to the short and medium term, right, on the API. If anything, they were like investing a hell of a lot more into the APIs. The ad business is starting to take off. And Amazon's DNA is not an advertising business, right? They're not Google or Facebook in that sense. And Amazon's DNA, similar to like AWS, which is like we're good at the infrastructure side of it. So here's like some infrastructure. And we don't really know how to do ads all that well. And so we need help on that side. And so we're going to invest in API's and partners, etc, to go and grow us, you know, grow that sort of point out. And I think that's largely held true. I mean, their ad business is flying. Unknown Speaker: It's like that is. Speaker 2: Number one and number two in terms of profitability for Amazon is an entire business. So I think their bet was absolutely justified and fine. And, you know, at the end of the day, sure, every single investor asked that. But I don't know that it was a showstopper. Speaker 1: Yeah, it's an investor question. And I kind of like, I care less about it as an investor question almost as much as a strategy question. Because at the end of the day, you want to sell your business. You did sell your business. Speaker 2: Yes. Speaker 1: Like you, you don't want to build You don't want to build a business which then gets, you know, basically all the values just destroyed. Yeah. Speaker 2: So the other thing is like, what if Amazon starts doing LTV stuff? Yes. And then, so for me, it's always Amazon's been good on the infrastructure, or eventually good on the infrastructure, let's say, and they can provide some metrics. What they're not good at is telling you what the hell to do with the metrics. And so like if we're building all the context and the decision tree that you need to go through on how to move this, you know, this metric, that's a valuable thing. And Amazon's never going to be able to copy that. Like you just spend 10 minutes in Seller Central or Vendor Central or something. And you just, you know, you'll know that pretty quickly. Right. And so the mode there for sure, well, mostly in terms of like Amazon copying that kind of thing. Is not unlike the metric itself. It's has the metric and here's what you need to do about it. And so, again, I think that's pretty, you know, that's a that's like an hypothesis that's held true and I think still will hold true for some some amount of time like people. You know, it's the same sort of people that were like, would come up, sign up to Nozzle, get an LTV metric, and then just cancel on the first day sort of thing. Cool. Like, it's sort of the same, you know, people were just going to vendors. So I don't think we were losing out on customers, because those are the same people that are going to Senecentral, get an LTV metric, and then, and then carry on, right? So like, understanding like, yeah, what decisions can you make? What's the most, like, for instance, for driving you to brands, you know, is it coupons or ads that does a better job of like, You know, net customer acquisition or retention or, you know, all that kind of stuff to move the metric and what's, you know, the relationship between those things and LTV. So there's that. The other thing I do want to add, though, is like, what happens if there's a change in the API? Because this actually happened to us, right? And this one's much, much harder to manage. And it was pretty existential to us, which is, you know, we wake up one day and we just can't onboard any customers, literally like errors all around. And it took us the best part of like four to six weeks to resolve this with Amazon. And when you're an early stage company, and I can't really think what was the exact issue. It was honestly something so stupid and trivial. But the problem is, you know, Amazon is like so siloed and it's so niche as well. Like it's, I don't know, probably some person in a basement somewhere knows the answer in Amazon, right? Like what the hell the issue is. And so that hurts us massively. Thankfully, I think we didn't lose most of the customers. They eventually signed up, but the problem is the runway part of it. We lost out on revenue that reduced our runway by one or two months. Yeah, then you just super stress it right like on that side of it. So like that, that is the more material side. How do you mitigate something like that? Well, that like that's extremely difficult, right? Like that is, you know, you need to like, I would like to think we have pretty good contacts, like API contacts within Amazon that we can, you know, escalate things and whatever, whatever. But again, those people move on. And then what do you do? You know, we're not in Seattle, I can't go like, although I probably should have flown to Seattle at that point. But, you know, so that is a much harder one to manage. And that's like an absolute material risk if, you know, something happens there. Speaker 1: Yeah. Yeah, it's interesting. So you're building this beautiful company, you're growing the revenue, and at some point, You sell the company. So do you want to talk a little bit about kind of that side of it? Speaker 2: Yeah, I mean, I think for me, it's also, you know, the decision to sell is related to basically trying to give Nozzle the best chance to succeed. And what do I mean by that, right? Like, how do we access larger customers, enterprise customers, the level of investment that's needed in the product, so you could Go out and raise some money and it wasn't a great like funding environment as well. We all in 22 I Want to say 23, maybe 23. Sorry. Well, it's the end of 22 23 not a particularly great You know funding environment over here. See you basically needed more capital in the business for what I thought we needed to do right and so like to To really grow Nozzle, to make it multichannel, right? So like go beyond Amazon. We never wanted to, and we still don't want to become like an actual bid management platform, but a lot of the actions clearly from Nozzle are going to be on the advertising side, right? So how do you acquire new customers? One of the, obviously the main lever on Amazon is going to be on ads. But so we don't want to do bid management, but we clearly needed a lot more investment on like the advertising side to be a lot more prescriptive around what we think you should do to decrease your capital, whatever, whatever the thing is. For me, it's like how do you set yourself up to succeed both from growing the revenue, easier access to, you know, customers and things like that, and then just like level of investment on, let's say, on the product side, allowing us to be multi-channel. And so, yeah, you know, Nozzle's found a great home in OptimizeOn, you know, accessing, you know, the customers who are multi-channel there and also like access the product development is interesting as well because Not only is it the Nozzle customer base, but it's the people within Optimizon who are using Nozzle as well that I can tap into, right? They're using Nozzle, so what are the insights there? What are the challenges? So that whole product cycle, the feedback and development, all that kind of stuff can happen a lot quicker too. Speaker 1: How big is Optimizon in terms of headcount or whatever you can share? Speaker 2: That's about 40 people now. Speaker 1: Okay, okay, cool. So, and how like, how do you manage? So you had a team of like nine, you said before we recorded? Speaker 2: Yes, maybe a little bit more, but yes, I'm on there. Speaker 1: So basically, you know, you're integrating 25% roughly of the purchasing business. How do you manage like an integration as a founder between your team and the new company? Speaker 2: Yeah, it's, um, you've got to make a, you've got to be culturally aligned beforehand, right? Like, sort of like M&A processes, for me, at least, and the other ways to go about it, you can be super transactional about it, I guess. But, you know, understanding, like, do you really click with the founders or the people you'll be working with is, like, just, I think, a really huge part of that. You've got to... Speaker 1: Did you know them before or do you only cover... Speaker 2: No, we were introduced by sort of a mutual acquaintance, yeah. But yeah, but James has been, you know, James has been great and clicked very early on. And it does sound maybe cliché to say it, but, you know, things like, you know, the values and all that kind of stuff, like, really does... Speaker 1: Yeah. I play a pretty big part. Speaker 2: But, you know, you also face a challenge where, you know, you're a SaaS company, a technology company, and like very skewed skill sets, let's say, into like, you know, effectively an agency business, right? And so like, that's a very different DNA of the business. And so, you know, there's definitely challenges around all those sort of things, right? So like, you know, this, this, I suppose I've been used to educating or training people on how to think about these metrics, if you've never heard of them before, and all that kind of stuff, that's fine. But, you know, the account manager or something, it's got like lots of other responsibilities as well, right? You're not like the center of their lives sort of thing anymore. Not that we were ever for some of these brands. But so like, you know, thinking about these trying to like, Upfront, think of these things and how you know how that will, that would be handled. But you know, until you're living through it, then you know, I think it's, you don't ever, you don't ever know. And I think a lot of the processes have a very high probability. Speaker 1: Sorry, integrations have a very high probability of going south very quickly. Speaker 2: Yeah, thankfully hasn't been the case. And to be fair, like also with With James, Optimizer has left Nozzle to be pretty independent. So you can sign up to Nozzle. You don't have to be an Optimizer customer to sign up to Nozzle kind of thing. So we kind of left our own devices in that way. Speaker 1: And like in terms of your role, are you basically CEO of Optimizer within this broader picture now? Or do you also take on stuff within the agency? How does that work? Speaker 2: Yeah, so I'm obviously heading all the Nozzle stuff still, but also Within the wider agency is sort of technology deployments and things like that, right? Speaker 1: So there could be, you know, it's used us to optimize and say, oh, that's right. Yeah. Speaker 2: So there's, you know, there's a whole bunch of like needs and constraints and stuff within the agency. And so understanding like what's bulldozers by decisions within the agency, how to really like be tech. I'm differentiated with technology. So like the stuff you really want to build, you need to make sure that like that fits into a very specific, like a, uh, optimizing philosophy way of doing things. Speaker 1: Cool. Speaker 2: So like, I don't want to go replicate other, um, I don't want to go up, replicate other, like really good, like SaaS software for the, for, you know, Amazon agencies. Um, I'll probably want to use them as a starting point and then build on the top of that for like specific workloads and, you know, things like that. So it's a, it's a different, very different way of thinking about things. It's kind of new to me, I guess, in that sense. But it's been great. Speaker 1: Have they made other acquisitions in the tech space? Speaker 2: Not in the tech space. Other agencies, yes, but not... So we're kind of the, yeah, I guess the first and only two days, I should say, tech acquisition. Speaker 1: But they've done it via agencies, so they kind of knew from the side. Yeah. Cool. I think that was like a super interesting kind of whistle-stop tour of all of it. A few final questions if it's okay. So number one, like, you kind of hinted at a few things you would have done differently throughout the discussion, but like, on reflection, on the whole journey, what what like key lessons or things would you take away? Speaker 2: Um, yeah, I think, I think in the early days, the hiring side is still probably the most important thing. Yeah, it's funny because like as a founder, everybody will tell you hiring is probably the most difficult thing, but it's also the most important thing. Speaker 1: Yeah. Speaker 2: And you probably know early when it's not working out. So it's best to part ways early. And so having read that, you know, 10 times, I still didn't do it. Meaning I kind of knew it wasn't working out for certain like senior people, but didn't do anything about it. Speaker 1: And how do you know if it's working? Speaker 2: It's not working? Um, I think you get a sense pretty, it depends on the roles. So like a sales role is probably a little bit easier. Speaker 1: That's the thing, right? In sales or in tech, like they're either delivering or they're not. And it's super, like, it's really transparent. We have a two weeks window on the tech side. If someone isn't building what they're supposed to run in two weeks, you know, like a founder who's a CTO, obviously feels like it's a reasonable thing that could be built, gone, right? But like, in some other roles, let's say custom success, or I don't know, like in other places, it's harder to It's hard to judge. Speaker 2: It is very hard to judge. I mean, I think like on the marketing side, it was always about prioritizing experiments and having a similar company wide attitude of like, you've got an hypothesis, you go test it, you know, like that sort of thing. And so I need you to be running, I don't know, three experiments a week. Kind of thing. But also just, you know, the general, I don't know, problem like problem solving attitude or new ideas on certain things like you can just tell pretty Pretty early with it, whether you know, you're drilling, but then, you know, as I said, I probably hesitated far too long on a few of those hires, just giving benefit, keep on giving the benefit of the doubt where I think you just, you just kind of know, right? And you're saying, yeah, yeah, I should have. And the guts always been something like very mysterious to me, being very, I don't know, data oriented in the rest of it. I always like, what is this thing called guts? But, um, But it's something I think I've learned over the past, you know, past few years of like how to tap into it, right? Speaker 1: I think so. Speaker 2: So like the hiring side is definitely one thing. The funding side is interesting because, um, I, today's environment is just so different, meaning a lot more can be accomplished earlier on, um, given a lot of the sort of tooling that's out there. Speaker 1: This is a question on how does AI change all of this? Speaker 2: So it's a big part. It's absolutely a big part of it, right? So it's absolutely a big part of it. Meaning, look, because of the toolings there and all these, I guess, possibilities now, expectations are also higher. So it's a little bit, you know, it's just because, you know, these things are out there, like expectations from like, let's say, customers and investors. Before, you know, it used to be maybe like a million run rates or something in revenue before, you know, you said these heuristics, right? So revenue run rate before you could raise a particular round or, you know, something like that. But now, with all this tooling, on the one hand, it makes it easier. But on the other hand, it's kind of like, well, that million is not no longer a million, it's 2 million or in kind of way. So, you know, just kind of corrects quickly. But yes, I mean, like, it's just so fundamentally different now, meaning I'm not a I don't write code, right? I can do like SQL, basic SQL stuff, but I'm not like, you know, production engineering or anything like that. But there are incredible tools now to get working, more than a prototype, like a very basic working product out there that you can charge money for without hiring any engineer, right? So using tools like Lovable or Cursor in some cases, which is an IDE. Unknown Speaker: You just look like, you just... Speaker 1: Or ChatGPT, right? You could, you could, you could build a project. Speaker 2: The ChatGPT is difficult because it doesn't have like a database and all that. Like if you're building a SaaS tool, you need... Yeah, you can do the website stuff, but like in terms of like taking a strike payment, you know, all that kind of, but I'm saying it's not that difficult. Like you could, you know, you could easily like, you know, I'm doing, and I think so much fun of the December holidays, just like doing some hobby sort of apps, but it's got a server and it's got a database and it's got a front end. They're saying like, you know, the best programming language to learn these days is English, right? Because if you learn good prompting, then you can do these things. Speaker 1: It's remarkable. Speaker 2: So you can get a lot more further down the road before thinking about, you know, funding and all that. I definitely think business models are probably like the one area where you have an enormous ability to disrupt and take on SaaS companies. So I mean, SAS has always been SAS is like either a perceived model or something like that. But it's always so a human can go and do the task at hand, you're enabling and empowering a human to go and do this thing. And we're going to charge you per seat or whatever it is. You know, the agent stuff is like super, super interesting. It's actually doing the task instead of the human. Yeah. So yeah, some of the like, I don't know more. You've had like, tasks on like, say, customer success and stuff like that. You've had like, Intercom, was it Intercom? I think come out and say that, you know, their agents are solving like some enormous amount of, of, of queries there and just cut down their, you know, half the team of the humans or, you know, that kind of thing. Being able to, like, change the business model and charge on a completely different, like, vector to what a SaaS company does, that is very, very difficult for any company to respond to. Right? And so, like, you've got an amazing opportunity to say, is there a different business model here that can really change something? What's not clear to me, though, and it's, like, super interesting, is to understand where the moats are if you're building an AI-based business. Because people would naturally, I don't know if you've got thoughts on this, because it's obviously something you probably get asked about a lot as well, right? Is it in the models? Is it in the data? Speaker 1: I love this question because when we first started raising money, everyone was saying, oh, like, you know, this is a thin, it wasn't a thin layer, but people say, oh, this is a thin wrapper on ChatGPT or stable diffusion. Anyone can build a And, you know, the people who are going to make money are the model builders. And kind of fast forward for years, what's actually happened is the rappers, so to speak, like perplexity, where you have a brand, and it's a recognized, you know, people come to and you have a recognizable brand, and you can charge people to use the product, that's where the values are curing. And actually the models you can swap out. It doesn't matter if I'm using R1 or, you know, O1 or Claude 3.5. Models are interchangeable and actually that's being commoditized. Speaker 2: Hundreds of billions of dollars later. Unknown Speaker: Yeah, exactly. Speaker 1: I think the application layer, which is where we sit, I mean, I hope, as you say, the agentic and the application layer, that's where You have the relationship with the customer, you can charge them, you know, as we try to. And we're going to do a human fee, like how much would a human cost to do this? Well, e-content is going to do like 100x any human. So, you know, give us the same, you know, higher one e-content and you're going to, yeah. Let's see how it all shakes up. But I think kind of like the thinking of this has really shifted in my view anyway. Speaker 2: Yeah. And it's probably going to shift again, honestly, because it's just like a pretty, it's definitely not an equilibrium anywhere close to that. A little bit in the weeds, but like, it's about the cost of inference. So traditionally with SAP, sorry, with SAS is, you know, you have all those upfront costs, but the marginal cost to serve like a customer is pretty close to zero. But now that's changing because now all of a sudden, yeah, so now it's like almost shift, almost the opposite, which is it's actually becoming expensive again to serve an incremental customer, which is why it's the most, you know, valuable customers. Speaker 1: Yeah, absolutely. Speaker 2: And so again, thinking about that, like, Is that a good or a bad thing? How does that work feeding to the business models? Like, especially if you're up against a SaaS company? So like, you know, all these sort of things. And then the other part of is on the data side, like if you have if you have access to I don't know private data sets or something like the data argument has always been very I mean, I remember like 10-15 years ago, even with the previous companies is always like, oh, but it's proprietary data sets. That's a different challenge. But actually, they're getting pretty good at making synthetic data sets to train models. Yeah. And so thinking about like, is it the cost of inference? Is it like the data sets? And then also the model building, right? Like you don't need At least to begin with, let's say, you don't need a super expensive machine learning scientist to go and build a custom-tuned model. You can literally feed inputs and outputs into Gemini or one of these models, and it figures out the relationship and everything in between. You don't need to know that it's using this method and this is blah, blah, blah. That's helpful, but at least to prove something out in the beginning stages, You can honestly, I've done it, right? Like I've fed these things into, into your models. You don't necessarily need to know everything under the hood there. So again, it's just interesting for me to think about like. Speaker 1: What is easier than ever to start a software company? Speaker 2: It's easier than ever, but it's enormously competitive. What I love about it, at least from a consumer point of view, especially on the ability to build these tools quite easily now, like I was doing over the holidays, like I said to you, is producing an app is not just the domain of some software engineer. Someone who has no coding experience, We can go test ideas over a weekend. And that for me is just incredible. That's absolutely insane. And so like a very, very bright future, you know, competitive in the rest of it. Absolutely. But just what are you thinking? Speaker 1: And I kind of feel that things like you won't have in the future, like these massive companies kind of dominating everything. I think my hope is anyway that The barriers to entry for entrepreneurship is severely lowered. As you say, you can be an individual person, no coding experience, build something amazing, probably even raise money, like if people are using it, people love it, then you raise money, build something proper. And therefore, hopefully, there'll be lots more entrepreneurs, and there will be a lot more competition. But at the end of the day, you know, fragmented marketplaces, and you won't just have like one Amazon, you have like many different places that you can, you know, like in search, for example, you have Perplexity, you have ChatGPT, you have Google, you have You have, you know, like things will start to become hopefully more fragmented and specialized and I think that's what we're here for. Speaker 2: Yeah, yeah, it's interesting. It's like to what extent does scale matter still? But but yeah, I mean, it's it's I don't have the answers. I don't actually have a strong conviction about these things. Yeah, we haven't even spoken about the marketing side. This is like building product side. But like on the marketing side, that's probably one of the first things to get disrupted in terms of like copywriting. And, you know, you get this incredibly personalized, you know, emails now of like these outbound, you know, agents that are just doing stuff. Speaker 1: Yeah. Speaker 2: Like just all around, right? It's like getting good on being able to like prototype these things pretty quickly. When I say prototype, it's still a proper app, right? Like you can take payments and stuff. It's not that difficult anymore, which is amazing. Plus on the marketing side, and then I'm just kind of like you can just go for it from there. Anyway, we've gone a little bit off topic here, but it does genuinely change everything. Speaker 1: Last question then. Would you do it again? Do you have another startup in you? Speaker 2: Possibly, circumstances are different. You know, I've got, you know. We have three kids now and everything else, so the risk-reward calculation is probably a little bit different. At this point, I'm just having so much fun building personal apps, which is probably going to become a thing, at least temporarily. A very short example of this, I needed to convert emails into markdown files to feed into an LLM. And I was trying to like look for an email converter and they all look a little bit like spammy, trying to get this stuff. I was like, you know what, it's literally easier for me to build this on reddit. And I did it in 20 minutes and I bought myself like solve my own problem by actually building an app like that. And it's just like that, you know, they're having that sort of default mentality, I think is, is, is inevitable. Speaker 1: Nice. Well, it's been an absolute pleasure to have this conversation. Thank you so much for coming. Do you want people to reach out to you if they want to find out more about Nozzle or what should they do? Speaker 2: Yeah, absolutely. You reach out to me LinkedIn. I'm pretty responsive over there or rael.cline.nozzle.ai. Probably the two best places to reach out. Speaker 1: Great. Well, thank you for coming on. Speaker 2: Brilliant. Thank you so much for having me. Cheers.