 Baiklah, maafkan saya sekejap. Saya hanya cuba melihat perkara ini berkelihatan. Bagaimana jika saya melihat perkara ini? Baiklah. Baiklah. Baiklah, semua. Sebelum saya teruskan. Saya hanya mahu periksa. Adakah ia terlalu panas dalam bilik? Baiklah, saya rasa beberapa anda tidak terlalu panas. Baiklah. Biar saya memutuskan ekon. Ia akan berjaya dengan cepat. Saya hanya membawa ke-26. Oleh itu, saya hanya akan memasak sedikit. Baiklah. Bersama saya, saya akan menerimanya. Ada sesiapa yang ingin berenang? Bersama saya, saya akan membawa ke-26. Baiklah, saya sudah menerimaput saya. Baiklah. Saya hanya mahu melihat perkara ini. Baiklah. Baiklah. Begitu yang perlu saya beri saja. Boleh saya menerimanya? Saya akan menerimanya. Saya akan meminjamkan perkara ini bagi orang lain. Err... My name is Alvin, my sharing won't be anything that is extremely technical. I can tell that a lot of work has gone in. She has prepared a lot of materials to share with us. But today I think my talk is a lot more about the way how we do things here at Digital Ventures, what I have learnt and this is the kind of methodology that we have used somewhat. For us to start off and also to explore the kind of potential venture businesses that we can build. And a result of this are all the wonderful businesses that you see that we have done. Some of which you might have seen the names like Oval Affinity. So it's the same methodology. So I'm just sharing with you, take it with you however you want it. Any questions, I will help to answer it. So let me just kick this off. So I think one thing is that far too many of us focus a lot on building the perfect product. Trying to perfect our code. Trying to make sure we know everything about the market. We know everything about the problem or the solution before we actually decide to launch it. And it will probably take a long time, one year, two years, how many number of years you need to get that done. And I think this is very true because people will always say that perfection is the enemy of progress. And I'm not too sure if you guys have seen this company that came up on the news not too long ago, maybe two weeks back. You know, this company is called Aqua, it's based in Singapore. I hope no one's from Aqua but they've been doing this product for the last four years and a couple of hundred plus of companies with I think total $20 million of funding and four years no product to show completely failed to launch. And I think it was in the news, that's why people were surprised. $20 million, no real product. So I think that's the lesson to learn here. So why is this happening? The complete failure can be avoided by trying to learn early on and also doing the cost corrections. That's what I'm trying to share with you, why we need to avoid all these complete failures. The reason for failures are typically because there's no product market fit which is very clear with Aqua. They're trying to figure out what's going on but as long as you never launch it, you'll never really learn from the market. And secondly, another top reason for failure is usually like poor understanding of the customers or the business. The moment you never launch it, you'll never learn anything. So always back to the same thing. So I think really what I'm trying to drive home is that the prototyping leads to a bit of testing as the moment you launch it and it's definitely going to fail along the way. But that whole idea is so that you can tweak that approach. Whatever you have figured out at the start, your assumptions might be wrong. You go back to the drawing board, you try it again. This is really the cycle that we always go through to help us to maximise our learning. So you take small little steps and make small little failures along the way. So on a very particular... Let me just maybe explain a little bit more just to add more context. But look at me. So I'll slowly explain to you. So on a very particular digital venture, we face a very specific challenge. So just for context, I think some of you might have heard, we have short projects and long projects. So the short projects are things like what we call the innovation and validation. So it's kind of like we take 10 weeks. One of our clients will come to us and say that, hey, we've got a big problem. We think that there might be an opportunity somewhere for us to start a business, but they don't really know how. So they come to us, right? They give us 10 weeks. We tell them, okay, we're going to help you figure out if there's an opportunity right here. So that is a validation project. That's the cups of it. Help them figure out if there's a real business in that area. So it takes us 10 weeks. This particular client, I can only say that they are an operator of job portals. They have many, many job portals. So that's kind of like the main business. So this client wants to expand market share in Asia, this part of the world, and they are keen to reuse what they have been doing back in their home country. So they think that it's a good idea. Just kind of transplant it into this part of the world. So there are three main complications with this problem. So firstly, the platform, what we see, is that it's not currently built for Asia. And for them to build the features to meet these needs, it is actually going to be a very big undertaking. That's number one. So number two, they're actually not too sure how big the market is. So for them, it's kind of like an internal thing. They've been trying to figure this thing out. They're doing surveys, but they're not 100% sure. And they don't know where to attack as well because when you talk about jobs, it's big. So when you talk about job portals, we've got white collar, blue collar, mixed color collar, so it's really kind of tricky. So it's impossible to make that real judgment online, okay, let's attack this and how much time we're going to spend to build that thing. I didn't know the complication number two. So number three, as you guys know, job portals is also kind of like a two-sided marketplace. You need to have the job post, and you need a person to job seeker to find a job. Then you can put them together. So they don't really know where to start first. Do we do this or... Oh, let me see. Do this first. Get the screen up. Yeah. So they're not too sure which one to start with first. So far, what we know is that it has taken the client more than one year to figure this out. We were surprised. We went in there. They were like, ask them. So how long have you guys been knocking at this problem? Say, last one year, no results. Okay, fine. This is probably going to be very difficult. So for us, the obvious way would have taken forever. So we will probably have to take a super long time to study, design, and then build something and launch it to get an outcome. So it's pretty standard. I think you guys probably have seen this. It's like, okay, we're going to spend the next three, four months. Let's do all the research, all the customer service. Let me call up customers. Or I'm going to put panel, customer, panel, sorry. Like focus groups. You guys know focus groups. Like, okay, focus group one, two, three, to all this. Okay. Then it's going to take, I don't know how many research. Then after that, we're going to do, okay, we've got all those results. Let's go and do the next round design. So like, okay, let's go and design all this. We've got the screens. We've got this and that. We go test the designs. Okay, not super good. I keep doing it again. Right. So it's going to be here. We don't really have the actual stuff. So once you have the actual stuff, right, then someone will like, okay, let us go and like translate all this into real actual features. We're like, okay, we're going to have feature one. Okay, just explain. Hi-Rail is the job. Hi-Rail, Seeker is the job Seeker. So like Hi-Rail, feature, we have this, then Seeker, we have this, then we're like, okay, let's build all this. You know, we don't know how long it's going to take. It's going to take one year, two years. All right. We're going to build all this. All we know, big corporate's work. Right. But actually, right, the idea is that we realize it is possible to provide a very convincing experience, right. We've way fewer stories than we thought. So, one way to do that, actually, right, you remember what I said, right, we just do small prototypes. Try to figure out what exactly is the real important features that we need to put out there. Right. So, just do the first base level damage. You know, like, do the minimum and try to learn 80% like the 80-20 rule. Like, do a little bit to try to find out as much as you can because everything else is going to be diminishing returns. So, focus on that. Do the basic one first. Then let's do the first round of design. Put it out there, get some real feedback. Maybe it's good enough. Maybe it's not good enough. Who knows? Until it gets into the hands of the customers, they tell us what it is. Right. So, then we just move on here with the design, with the most basic features, so one and two, like, maybe it's just a couple. You don't have to do everything. Let's just do a couple of those important ones. Let's launch it and get the feedback, right? So, that's kind of like the approach that we did. So, if this approach in mind, oh, crap, sorry, it's indifferent. Let me jump the order. So, actually, it came down to really this core features based on that mindset that we had. So, on the higher side, it's actually really simple. All they really need to do, is to see who is the one that applied for my job, which is the job seekers profile, and I need a way to contact them. That's really, really super basic job portal. Then, for a platform, as an operator of the platform, what I really need, it's just a way to edit maybe a job post, edit the job seeker profile, you know, have a way to capture interest in a job, the way to capture the interest in a job seeker. So, be like, if I'm interested in a job, I can capture the interest the other side. So, how I capture that. I need a way to match the job by parameters. So, maybe they have like, looking for someone with an education level or something-something. Then, the job profile is someone of a certain education level or something-something. So, we're going to match them by certain parameters. We need a way to match them by distance. So, for example, you don't want to find a job that is like, I don't know, five hours away. You want to find something that's close to you, right? So, it's really simple. This part here is just a mirror of that. So, I just need a way to register my profile as a job seeker. I need to see what the job is about and I need a way to contact the job hireer. Right? This part's clear. So, it's really super simple. We realize we didn't need all bunch of all the features. So, then we're thinking, okay, so what's the easiest way to do all this? Because we don't have a bunch to figuring this all out. We only have five weeks left. Of course, we already spend the first five weeks figuring out the commercials, figure out the interviews, figure out what are the initial designs and all that already. So, then we only have the last five weeks. But we cannot spend five weeks to do this, because as a company that is paid to give solutions, we also need to present them. So, the last week we need to spend on presentation. So, actually we technically only have less than four weeks. And then four weeks is like super two or three weeks. So, then we're thinking, okay, so what's the really, simplest way to do this? So, it turns out, that actually you can kind of get away with this in many, many ways. So, the higher feature, okay, I need a way to post my job. I mean, frankly, you're just really collecting information, right? So, actually you just need a form. Okay, simple form. Okay, then I need to see the job seeker profile. Okay, so then, okay, so we just give first. So, platform features, right? You see, I need a way to edit job post. I need a way to edit job seeker. I need to like capture interest. I need a spreadsheet maybe with an API. So, I can just update. So, it's like really simple. Maybe you guys have something like this as well. I need a way to match jobs. It's maybe a script. I need a way to match job by distance, some geo coding service. Okay, then I need a way to notify matches. One way. So, we can get that working. Same thing here, some form and some event. So, actually, before I go into the actual solution, just curious, right? When we were talking about forms, right? Do you guys have any solutions that pop into your head? We just need a form. What you guys were thinking of? Okay, go of forms. Okay, make sense. Okay, go of forms. Something else apart from go of forms? Type form. Okay, type form. So, something like a spreadsheet with API. Very obvious. Addable, okay. Good, okay. Then, script. Okay, this one, everyone knows. You all can write script, no problem. Dynamic HTML page. Okay, Jamstack, okay, notion. Anything else? WordPress. Okay, yeah, GitHub pages. Okay, cool. Okay, cool. So, let's see my next page. No, sorry. Okay, jump back. So, actually, this was kind of like the grand plan. So, you all kind of lend most of it already. So, let's ignore the first side here. We don't really need to see the first side. But actually, the form, eventually, we're in a type form. So, it's really simple. We just need a type form. There's API, and we just need this to go into a database. Actually, Google Sheets kind of have got everything you wanted. They have geocoding services, which is already there for free. You will also have like the like image host thing. You know, it's not real. But you can actually use Google Drive to host. You need to hack it a little bit. So, that works as well. Then, you need to run scripts. So, you can actually run scripts in the Google app scripts. So, it's like drop posting page. And, what's the other one? Drop sika profile page. There are really two pages, right? So, actually, they are just like two super simple like re-act pages. Only because I was like familiar with re-act. You guys probably use something else. They are super easy. Then, matching them together, right, was a combination of scripts. And also, matching them together was a combination of somewhere here. Some sort of backend we created, yeah, TypeScript. There was a very simple TypeScript engine that actually still has tests on it, which I was quite proud of that. So, we actually managed to have a test, right, of like 40 plus test cases to make sure that the test, the matching logic was correct. Because the matching logic was matching across like seven or eight different parameters and there was a bit of logic there. So, then we just wanted to make sure they was matching correctly. So, what's that was really here. So, we actually did once the matching engine came up with a list of matches, actually it just populated back into the Google Sheets and in turn we really just look at the short list and then just like copy paste and send the what's that messages out. Actually, we made it easy for him. There was a script that gave him the automated messages. So, all he had to do was copy paste, copy paste. So, this was what we did. Okay, so, forward again the final product that we did actually was like that and we did all this in less than two weeks. So, the page itself there were two pages but I'm just showing you one. This is what we sent to our client. So, this is kind of what it was. So, it was a simple page. There was an image of the person who, they did the matching and then there was all the other parameters. Once you click on this where you can as a job hireer I can click on this and then I can see the person what's that directly. This was a later iteration. We could click on this and directly talk to this person. So, actually it's possible to do all this. This is the if you know how type form looks like, this is what the type form is. So, we presented this to job hireers. We ran a campaign and they were actually going door to door and showing them the brochure and telling them hey, you know what, we have this new product that is helping you to find hireers like blue color workers would you like to register for it? So, we showed them this and they were like okay, this is interesting. Can I register? And then immediately give them an iPad and they register for it. So, yeah, it was pretty cool. We did all this and finally I think the best part is that everyone loved it. So, we made very good progress in just two weeks. So, we had more than 500 jobs posters. Wow, amazing. 2000 plus job seekers registered and pretty cool because we had almost 5,000 job matches online. These are real job matches. So, one to six. One candidate got almost as much as six vacancies. So, I'm just trying to show you what is the real impact. We had a portal and then do it all in like I don't know like a full blown super cool looking website but we knew that like, you know what, we just wanted to make sure that the client's objective was met which is is there a market? Do you think people will use it? Is this the right target market we're going for? So, the moment we we had that narrowed down, just launched the product, you know, we get the real results, we show it to them. We didn't fake this number, this are all real. Right? Okay. I didn't want to make this talk very long so I just want to quickly just wrap this up with all this isn't possible because of all this modern tooling. So, I'm sure some of you have heard this but if not, I'll just share one more time. So, you guys have heard air table. For air table, it's a very good collaborative data store. It's great. You can put attachments, you can put. You can do relationships, attachments. It actually has some referential integrity in there. So, it's pretty cool. And then all in one platforms if I was a real life saver, you can do anything in there. CI, CD, hosting, serverless functions in Dino as well. Anyone Dino? Shout out. Okay, no. Okay. Analytics like HIP is really good. If you guys haven't figured out like you know, you ask your product team say, what do you guys want to track? They say, I don't know. You know, then doesn't matter. Just throw HIP in there. So, HIP is a platform right where it's like Google analytics but better. You put the tracking code in there. It doesn't matter what you're tracking because you can go back in time later, right? To define what were those click actions, what they mean and then you can analyse from there. Right? So, actually, it's pretty cool and it's free. So, just go ahead and use it. If you guys looking to like have some sort of like NLP kind of experiences, there's something called AI. I think it's free if I'm not wrong. So, one of the other projects we saw, they use this. You don't need Google scripts. You can run it no matter how you want. Right? So, I think that really is all. Okay? Even if you forgot everything, there's just three key takeaways. So, first one is always be accelerating your learning. You know, what are the small steps you can do to accelerate your learning. You know, sometimes you might not be able to do it by yourself. You can only accomplish this by talking to someone. So, like for us, actually, our team wasn't all entirely aligned on doing this experiment together. We needed someone to come in and say, hey, should we just do it this way so that we can quickly find out what's happening in the market. You know, it was a one, two-day conversation. You know, you get some buy-in, people say, okay, then we did it. Then we eventually accelerated and you know, learn a lot. So, you might always do it this, but you might have to talk someone into it. So, as long as they're on the same page as you, maybe the next thing you have to consider about with your conversation with your product manager next Monday when you go back to work and maybe there's been a long list of things that you guys are building. Maybe you can just ask them the question, hey, what's the minimum set of things that we can get away with? Maybe we can instead of 10 stuff, maybe we can re-paratize and just build the three most important stuff, maybe. So, that's one. You can get away with that. And then finally, the most important thing is don't build life for the last five, six months. No one's seen it. If you can shorten the time, get it out there, put it outside, let people see it, give you some real feedback. You know that would be great. So, this is a turn. Take away. All right. And with that, I think I have no more slides. I'm done. Thank you. Okay, before questions, maybe I'd just like to draw your attention to the two amazing individuals behind me. One is Arial. Arial. Arial is without Arial's support, we wouldn't have been able to make this whole thing happen. She was organizing all the food, getting people here. All right. And next is Robin. Woo! Robin is my coach here and he's also the engineering director at BCG TV here. So, three of us are here. Happy to answer questions. Later, after this, you guys can talk to us. I'm happy to take questions now as well. Okay. Do you want to join us to find out more? Robin, you're going to say... How much money do you got? How much money do you got? Okay, we have I can say two major phases, right? First phases usually is like innovation and validation. And then after that, it's what we call stage gated. We go back to the client and say, hey, say, great. Yeah, let's go to the next phase. Then we have a long build phase. So, that's when engineers get a lot more involved. We start building out the entire big system. Then, you know, we launch it. But still, it's also in like phases. We won't like do a big bang. Usually it's, we have like alfa, beta and all that stages. So, we can slowly launch it out. Yeah. A kind of, yeah, exactly. Question. Failure stories. Oh, my God. Maybe before my time. You never want to remember them before my time. Don't remember, actually. I'll tell you. It does alright. Oh, yeah. Please, please, please. Ben, we're going to a shipping company and we're exactly the same...