 It's good to meet you. I'm Brady Forrest. I'm visiting from San Francisco. It's my first time in Singapore, the nth time in Asia, but it's it's really great to be here. Before I get started I want to just call out Bunny in the back. I would not be here if it weren't for Bunny on multiple fronts. So I met him about a decade ago through a Riley when he was working on Chumbie and stayed in touch with him. We became friends, gone to Burning Man together, but most importantly about four years ago I was approached by the CEO of PCA just to whether or not I would want to come and start a hardware accelerator and that ended up becoming everyone. That's what I'll be talking about but Bunny was the first person I called when I was making the decision and so thank you Bunny and I just got his book last night which I can't wait to read on this trip. Alright, but now I'm gonna talk a little bit. Okay, but now I'm gonna talk about some lessons I've learned from working with 74 companies, hardware startups at Highway 1. Just a little background about myself. I have my career span about 20 years now. Started off doing supply chain management. No problem in case. Nope, okay. Started off doing supply chain management then ended up doing working in O'Reilly media where I started Ignite if you've ever heard of that global talk series then got into investment at COSLA Ventures with Vinod COSLA and then found at Highway 1 and wrote the hardware startup for O'Reilly. At Highway 1 worked with 74 different companies, all hardware startups and by hardware startup I mean a company that has to make something. Yep, I know it's still loud. As part of its business. So generally I tend to invest in companies that are basically software wrapped in plastic. So it's a company that although it is shipping an object they're really trying to make their money in general from data or software because they know that their hardware is going to be abstracted away. Either competitors will be fast followers or the end goal is to embed into someone else's system and have other people making the hardware. These are some of the companies that we work with. This is what the products look like when they come to us. They're very early stage hardware prototypes. They have not been through design yet. They're generally functional, wires sticking out of them, almost always just one generation from Arduino sometimes made of Legos. And then when they leave this is more what they look like. We're aiming for a complete product, something that you can really show, shows an investor what the end goal is. Even if you've only made just one. Something you can show to potentially consumers as well. Because let's be clear, although they may be functional, although they may fly, those drones may fly, that may make coffee. It ain't pretty. So you want to show something like this when you're actually going out to try and get investment or get someone that's by your product and to take you seriously. So Highway 1 is a four month program, had seven cohorts, two cohorts a year, take about 10 companies per cohort. We have engineers in-house. We take the companies on a Shenzhen trip, which is why I'm actually in town. I find the Shenzhen tomorrow to take the companies around and to see factories and kind of see what manufacturing at scale is like. We have engineers in-house and we help them out with fundraising. We invest in the companies. When we first started out in 2013, we put in $20,000 US. Now we've made it much simpler, but we put in a lot more money. So we put in 50K for 5% of the company and 100K for 8% of the company. Now this is just ground, you know, it's not a pitch for Highway 1. These are more about what I've learned from there. So before we get started, I just want to play some of the assumptions that I'm talking about here. You know, this is based on my work with startups. It's very US centric, but I think a lot of the lessons can be applied internationally. A lot of it defaults to work with consumers. About two-thirds of the Highway 1 companies have been consumer, but we've had many health B2B companies that go through as well. And increasingly, we're moving B2B because it's a much clearer revenue stream for companies rather than hardware or consumer, which is kind of a flash in the pan. And finally, a lot of it assumes that you're looking for venture funding. And I just want to stay upfront, although that is the bias for Highway 1 companies, that is not the only way to start a company. It's going to require a different level of engagement and involvement if that is the type of company that you want to start. But there are many, many companies out there that ship hardware that do not take outside investment or find investment in other ways from friends and families. But if you're taking on venture, then you have to realize that those venture capitalists are investing other people's money and they have to get a return. And so they're going to drive you to growth. And so these are the assumptions going into the talk. First one is, big surprise, hardware is not software. If you are doing a hardware startup, it is not the same as a software startup. To put it simply, software, it just kind of works. You can put up a web page, you can change a button, you can add a text box, you can remove a text box, you can write a simple web app that tests your basic assumptions and get it out to 10,000 people with a minimum Google ad spent. And you can do that week after week as you're starting to iterate. You can use various frameworks to quickly whip up an Android app. It is just not the same as hardware. Hardware is more like this. It is a lot of development and then maybe it works. Now, hardware has gotten a lot easier to do. Arduino, MakerBots, tech shops, which are tech labs where you can get access to machines have made it more like software. You can download code and get tech specs from, say, Instructables or Thingiverse, but there's still fitness assembly for or kitten assembly for all of the units that you decide to make. And so testing it on 10,000 users is basically impossible at an early stage. And that's caused there to be a very different investment landscape for hardware startups. You know, although there are some big returns, most of these are still about the software. You know, Nest and DropCam were both bought because of the services associated with it, not because of the objects. And a lot of these VC firms will invest in hardware, but they need to see serious software and the majority of their portfolios are software. There's only a handful of company or of investment firms that really focus on hardware. And that's led there to be kind of a different ecosystem around hardware startups. So this is really the way I see hardware startups from the beginning. You know, you have your prototype where you're first looking at the unit, then you're validating it with the customer. You figured out what you're trying to make. You've got your product design, then you go into manufacturing. And then once you've actually started to sell a bit, then you try to get to scale and really distribute it. So at that prototyping stage, you know, you're probably in university or on a crowdfunding website or an equity funding website. And you're building out the proof of concept and you're getting the founding team. You need at least one engineer. You need at least one software skills person that can also be an engineer. But you still need someone who's going to take the lead on leading the company and making business decisions. While the engineer is focused on the tech. I find the best teams are three to five, though certainly worked with teams that are one to two. I generally do not work with solo teams anymore. Because even if a solo founder is amazing, if they haven't been able to find other people to lead, then putting money into them doesn't really make sense because they need to scale as a person as a manager. At the validation stage, that's when you're taking those units, those Arduino units, and you're starting to get them to your customers. And that's when we see a lot of companies go into accelerators. And that can be highway one, that can be hacks and Shenzhen, white combinator, tech stars, those are and 500 startups are all starting to look at hardware, but they don't focus on it. But that is a great place for customer validation. And to start to feel out the investment cycle, where investment teams really start to dive into the companies is at the product design page. And this is where I think is very, very important for a hardware startup. And there's two key points right here. One is the design for manufacturing, the DfM. So you've made this Arduino product, you've worked with the designer, you made it look very, very pretty and sexy. But do you know how it's going to be made? Because that will inform your margin stack. And the margin stack is all the costs that come on making your fully loaded product. So there's the bomb cost, you know, the cost of all the parts, then there's the manufacturing part cost. And then there's shipping, there's distribution, retail. And so if you have been working with customers who only want to spend $100, and it's going to cost you $50 to get in their hands, you don't have a business, or at least you don't have a venture fundable business. And so you need to start thinking about these things very, very early. And sometimes that means cutting features. Sometimes that means going for a different market. Very often, we have our company start off doing consumer, and they end up going B2B or government, because the costs don't work for consumers. And then as they scale, and maybe things become cheaper, they find a cheaper way to ship the product, then they can go back to consumer. Then after you actually know what you're making, you know that you have a business or at least the best of your ability. That's when you're going to go to manufacturing, start to approach companies like Flex, PCH, perhaps Foxconn, and maybe start to consider crowdfunding, if that's something that you want to do. Now I'll talk about it a bit later. But I think, you know, too often, we see companies go to crowdfunding at this stage. And if you are going to ship, if you think you're going to ship a lot of product, then that's the wrong stage. If you're just trying to ship 10, 50, 30, that's something you can do with pizza and beer and friends. But if you're going to be a real business and you're trying to do a blockbuster Kickstarter, I think you should focus on it more at that stage. And then after you've actually made your product, then you need to start thinking about scaling. And scaling is not just manufacturing, but it's actually its distribution. So selling from your website, we have the best margins. But then you're in a Google war. So you want to start thinking about Amazon, you want to start thinking about going to Best Buy and other retail companies and thinking about distribution. So this is just an early look at the hardware ecosystem. Another thing, another way that hardware is not like software is the funding gap. So as this base level idea is we're just going to say for argument's sake that this product costs 2 million to ship. And if I was to look at the cost breakdown, that popped up a little faster than expected. You know, it's 200,000 for prototyping, half a million for DFM, and then another over a million dollars for manufacturing and working capital to fund your inventory. But the reality is, friends and families going to be 25 to 50k, maybe you'll get 100,000 from an accelerator. And then the average crowdfunding campaign is 125,000 US dollars. That is not enough to really ship this product. And I've seen too many companies come to me at Highway One, where I'm only giving out $100,000. And they say, we don't have the money to ship this crowdfunding campaign. And we've accepted the funds. And we don't have any other choice. What can you do for us? To which I say, nothing. Like, good luck. You should have come to us before you decided to take people's money. There are many ways to go about this. But taking other people's money when you don't know how you're going to live, I don't think it's a good one. We actually had one of our companies at Highway One, where they set their goal for 50k for a crowdfunding campaign to be clear. But the internal goal was 250k. And they knew they could not ship it in a realistic amount of time at the quality level they wanted, unless they hit that goal. They ended up reaching 100k. And it was the hardest thing that that CEO did. She canceled the campaign the last day. Didn't take anybody's money and charted a new path for the company. Because it was clear that the product that she had developed would not fit the customer's needs and would not get the company to where she wanted it to go. All right. Second lesson, you need customers as much as possible. And for this, I want to look at a company I did not work with, but I'm friends with the CEO, Pebble. Now Pebble was, you know, one of the first big Kickstarter campaigns. Eric broke every rule I just talked about. He, you know, he went on Kickstarter as a Hail Mary, and it actually paid off for him. But he's the rare one. The thing is, most people think that this was the first Pebble watch. That's not true. This was the first Pebble watch. It was the impulse smartwatch. It only worked with blackberries, and you could only do text and emails. He still has that inventory sitting in his closet to remind him about customer development. And when I was working on the hardware startup, you know, this is what he said there to me, you know, don't worry about making something big from the start. Instead, get it out there, get feedback. Don't be so precious with your prototype. Instead, follow a process more like this. So when you're prototyping, just keep iterating on the prototype, then test and then realize you have to go back to the prototype and keep doing that until maybe it's ready for customers. So demo, and then realize you're going to go right back to prototyping. And then maybe you'll get a little further. And all these things, these take more time. It is very easy for you to just prototype and do something new. But once you engage with a customer, that takes up a lot more time. You know, so they ended up doing okay for a while, and kept going back to Kickstarter, but weren't able to raise the funds to keep going. And so as probably most of you know, they ended up being bought by Fitbit mostly for their app store. And Eric is now in Shenzhen for the next six weeks, and I'm having dinner with them tomorrow. So lessons to be learned. The next lesson is prototype, test and iterate. So this is a highway one company, compression kinetics, you won't have heard of them because they haven't shipped yet. But what they are making are stockings that go on your feet and do active massage. So probably heard of compression stockings, they only work if you're moving around. But if you're just sitting there, they don't. And if you're a professional athlete in say the NBA or NFL, then after every practice, you go in a room, and you hook up a machine like this, and you put these huge stockings on. And this really loud air compressor runs for a half hour and pumps blood up from your feet, because it improves recovery and keeps your muscles limber. They love the effects, they hate the experience. Also $3,000. This is what compression kinetics came to us with. So very, very simple, battery pack, Nylitol for shrinking these, this plastic webbing together and Velcro straps. They ended up partnering with IDO to experiment with what does it look like? Who, you know, and to take that into account, they said, Okay, who are they going to work? Like, who is this going to be sold to? Is it hospitals? Is it travelers? No, they decided on runners. Runners already wear compression stockings while they're out running. So how could they work with this customer's flow? That'll probably be their first customer segment. And so they started prototyping and prototyping and experimenting and trying out different colors. And looking at orange gager movies, and other colors. And this one's blue, and then green. Oh, now we're into pink. Yes, and Velcro added to the pink. Ooh, I don't like the yellow. I think the gray looks kind of nice, though. Oh, no green. And kind of spider like. And in the end, this is more what it looks like. So the idea is, a runner will go running with their compression stockings. And there's a small magnet here. And then this is their part of the This is like the true product. You can strap it on. This finds the magnet. So you've fit it in the exact same spot every time has the battery. And it uses buckles and Velcro so that you can set the exact strap, but then you can keep it right for your body every time you don't have to re tighten it every time. But in the end, they went through about 30 or 40 different prototypes. John would be up there selling every night. He got really good. It requires that kind of dedication. So remember, prototype tests, then customers, because all of these things take more time. I'm really glad they didn't stick with the yellow. I think these look pretty nice. Now, when you're prototyping, it's important to realize that you can prototype in 2D, 3D, and 40. You can draw very, very quickly, and then mold something out in clay. You don't have to go right to 3D printing or right to CNC. Think about the amount of time that each of those steps take. And so when we were working on the hardware startup, we kind of think of it this way. 2D is the concept. So what is it going to generally look like? 3D is the form. And then 40 is the experience. And each of these comes with their own set of tools. So in the 2D phase, you're looking at laser cutting, you're drawing, and then sometimes you can start to make, say 2.5D structures just from a laser cutter. Then after you felt it out a little more, you think you know what you're really trying to make, then you can make a 3D printed object. And then later, go to CNC before you're ready for 4D. And 4D is probably more video. It's working with the customers. It's seeing what is that experience like with your prototype. And so again, it's really easy for you to work on your 2D and then go to 3D and wait on the 4D because time increases as you go to the right. And so do the dollars. It costs a lot more for you to film a video or to work with customers than to just draw something. So we talked about this before, but crowdfunding will not ship your product, at least not at scale. This is the story of InstaCube. InstaCube back, I think it was in 2012, put up their Instagram photo frame for $99. And I was like, wow, how can they do that for $99? I've never seen a photo frame that cheap before. Amazing. What a great idea. Myself and many other people thought so. I checked out who the creator was, and they were D2M. They were consulting from that help companies manufacture in China. I was like, great. This is awesome. I am super going to be excited to get my Instagram photo frame. That's great. Well, it was expected in March of 2013. I got it the fall of 2014. I knew they were having problems. Well, first, you know, when they missed the ship date of March 2013, and probably a little, but really it came through. But it wasn't too surprised when they said they were switching manufacturing partners. And they cannot ship in July. And that update came after the ship date. And too often what I see is, this is what people think a hardware startup is. I have a 3D printer, I'm going to crowdfund, profit. It doesn't work like that. Cold reality is a lot more IDA, do a whole bunch of things, set up vendors, set up Slack, find your team, test engineer, then maybe ship. So what I think Instacube did was they went to crowdfunding in September of 2012. They then started working on DFM. They went to manufacturing in 2014. And they did actual ship in fall of 2014. So their stages looked like this. In reality, I think you need to do this because you don't know how you're going to make your product until you've done DFM and lined up your manufacturing partners. You haven't done your margin stack. They couldn't ship the product for $99. And it ended up killing the company. Because remember, going from left to right costs a lot more money. And in their end, they ended up getting written up. They got death threats. People quit the company. It was pretty bad for them. The only reason I keep my Instacube is just so I can hold it up in talks like this. And I'm like, the manufacturing is so shoddy. It doesn't work. But it was too big for my suitcase. So in the end, just remember, a hardware startup is not a software startup. You need customers. Don't be precious with your prototype. Get it out there. Learn as much as possible. Prototype, test, iterate, go back. Keep playing around. Don't waste your time. Prototyping can be very simple. It should be very simple before it gets complicated. And do not rely on crowdfunding to ship your products. Applications for the next class of Highway One will probably think it's like Mayesh. Oh, and I should say, I have stepped away from day to day operations of Highway One. And past the torch off, I left PCH in December. But I still help out quite a bit with the teams since my being here. And can answer questions about Highway One, hardware, all that type of stuff, either now or, yeah, I think we're about wrapped up, or after the session. Thank you very much.