 Ghost is a blogging platform that is free, open-source, beautifully designed, and available to anyone who wants to use it. The concept of the blog has transformed journalism in the last ten years, like the MP3 did to music, so some of the biggest ideas and discoveries of the last decade have been broken through the blog. And yet, in the last ten years, not a lot has changed when it comes to the software. We're still stuck with stuff that's complicated, more complicated, full of cats and porn, or simply has no sign-up button whatsoever. And Ghost is what's going to change all of that. I know how to build blogs, and I know how to build blogging platforms. I've built blogs for Microsoft, for Nokia, for Virgin Atlantic, for EasyJet, and for many others. I also spent two years working as the deputy head of the WordPress user interface team. But WordPress has grown up, it's not really about blogging anymore. It's moved on to be about websites and content management and all sorts of amazing things. But what I care about is blogging. WordPress, by all rights, is no longer just a blogging platform. And that's exactly what Ghost is. It's just a blogging platform. It has a beautiful dashboard that shows you everything you need to know about your blog in one place. Managing your content is as simple as browsing through it. When you need to edit your content or write new content, you can see markdown on the left and preview on the right. It's one of the most simple and beautiful ways of writing for the web that we've ever had the pleasure of using. Uploading images directly in place where it will appear in your content just works. And adding tags or categories is quick and easy. It's mobile optimized, it'll work on any device, iOS, Android, whatever you have. And tablets as well. The split view works in both portrait and landscape modes. We've built Ghost around three really important principles. The first of that is that Ghost is built for its users. A lot of open source projects suffer from being targeted far too much at developers and that's not what we're doing. The second one is that Ghost is free. The MIT license means you can do pretty much whatever you want. No restrictions on themes, plugins, conferences or anything else really. Lastly and most importantly, Ghost is being made for love, not for profit. If successfully funded it will be set up as a not-for-profit organization. Now why should you care about that? Because it impacts on our motivations when we're creating the software. Do we want to make millions and sell to Facebook? Or do we want to make something that's genuinely good and serves its users? Not its investors and shareholders. It's not just me working on this. I have an amazing team working with me who have the technical chops to make this happen. Rob Hawkes, technical evangelist for Mozilla until very recently did the first pass on the Ghost code. And now heading up our technical team is Hannah Wolfe, senior developer at Moo.com. We've worked really hard to get to this point and we've got a working prototype. But we need your help to finish Ghost and to ship it to the world. It's not just about blogging, it's not just about making something that looks good. It's about giving writers tools to push blogging and to push journalism to the next level. We're not just making this because we want to sell it. We're making this because it needs to exist. Ghost is about the future of the freedom of speech and it needs your voice. So having me, very nice to be here. Just so I know who we're talking to, I guess most of your developers. How many of you are founders or working on your own startup projects? This guy, I'm not sure. He forgets. He thought about it. So everyone else, are you working for startups here or what do you do? Yeah, it's in the start-up. Yeah? Yeah, and I'm actually going to take it. We're actually doing it along with a bit of course. Oh cool, good to hear. It's the problems. We're using it for notes. Yeah, it sounds good. Any designers? Non-developers? Okay, so we've got all full-time employed developers. Good to know. Trying to remember all my JavaScript knowledge now. So I think we'll start with asking John what he did before starting Ghost. Because I don't know if most of you know. I think we did a bit of a background research on John. But maybe John can share what did you do before starting Ghost? Sure, yeah. How many people have heard of Ghost? Sorry, this is my last one of these. Okay, good. Okay, so there's some knowledge. Okay, so I used to be a freelance web developer years ago. That's kind of how I got into everything. And I used to build websites for all kinds of clients. But it ended up mostly being blogs. That kind of ended up being my specialization. Building blogs for all kinds of companies. And they would normally get an agency in to do that for like $100,000. And then I would come along and say, I'll do it for 20. And they thought that was a great deal. And I did it in the space of five days and then took three weeks off. So that was kind of how my freelance model works. And the more I did that, the more I thought, well, I should do more with WordPress. And this is working out quite well. So I should really know more about WordPress. So I started contributing to WordPress Core, the open source software. And at the time, they were just forming their design group. So they had developers contributing to open source, but no real designers. So they just started this design working group, which no one really showed up for. But I did. And I came along and they had a weekly meeting and I went every week. And slowly but surely, I kind of did more and more and more until the point they said, well, you should just take over this group with the deputy head of UI for WordPress and help us manage this whole process from conceptualizing designs for the back end admin to building up the front ends, shipping code for that, just the HTML CSS side. And that was my introduction to contributing to open source, which was pretty good because it was on a platform that was powering what's now about 25% of the web. So every time you successfully committed something to the open source code base, you knew it was going to be used by literally hundreds of millions of people. Which on the one side is really scary and on the other side is really, really gratifying when you see a release go out with your name on it. So that was kind of my early start, I guess. So we were wondering like, why did you choose to launch this open source project to Kickstarter? Yeah, that's a good one. So I came up with the idea for Ghost after working on WordPress for a long time and seeing it move towards this kind of content management system direction. And I put up this blog post, which kind of was the precursor to that video. Kind of explaining the idea and just saying, I've had this idea for a long time, and who wants to get another blogging platform? But here we go, this is the idea I had. And it went to, I thought I'd get a few thousand page views. It hit number one on Hacker News on, product content didn't exist back then, but on all the equivalent things. And had just over a quarter of a million page views in about five days. And so I thought, okay, this is something worth working on and spending more time on. I've never seen any of my ideas get that much traction immediately before. But I'm not gonna work for free, like I do still need to eat. So how am I gonna fund this? I didn't wanna go and do the whole VC thing. It doesn't appeal to me, that's not my big motivator. But I did want to be able to still actually pay myself a salary and work on something. So I thought this crowdfunding thing seems to be working quite well for other projects and it was shortly after, what was the open source games console, Uya, I just done a Kickstarter. And so I thought, well, people seem very enthusiastic about it when there's a blog post. Let's see if they're gonna be enthusiastic about it with their credit cards. And use that kind of as a decision point of whether to spend any more time on this. So it was the most logical model really. We had lots and lots of community interest. So I thought if you have community interest, see if you can get the community to fund it as well. And then you'd be giving up zero shares, zero control of the company, and you'd still be able to get it off the ground. So it's kind of a natural approach that felt like it at the time. And you were previously deputy lead at the WordPress UIT, correct? Yeah. And when you started Ghost, you decided to use Node for this. So, and WordPress is built on PHP, right? So why the decision for Node and not say something that you were probably familiar with, say PHP? Yeah, it's a good one. Early on, the idea was actually to fork WordPress. And that was my kind of designer brain wanting to sort of take the easy path. And the idea would be you preserve the ecosystem of WordPress, so you could still use the plugins and the themes. But you'd kind of do all the new stuff, which would be cool. And then after talking to a bunch of smart people, someone pointed out to me that you'd get the worst of WordPress, which is all the technical baggage for the last 12 years. And none of the benefits, because the community wouldn't have A, any real reason to switch and B, still compatibility problems with actually using any of the rest of the ecosystem. So it really wouldn't be very interesting and it really would be very hard. So then what do you go to next and what's the other potential options? Didn't really know the answer. Obviously people have done Ruby blogging platforms, they haven't been very interesting. Node at the time was still very new. It was just starting to get a little bit of mainstream adoption. Not even PayPal adopted it at that point. But one of my friends Rob Hawkes who's in that video kind of said, well, let me just, I can make this editor thing work with a bit of Node and a few JavaScript packages and we'll just throw something together for fun. I said, cool, yeah, that sounds good. Node, I've heard that's up and coming. And the more we built on it and the more we sort of expanded on this prototype, the more we realized, hey, this is actually really fast. And turns out there aren't any other blogging platforms built on Node, so it will get a little bit more interest, get a little bit more traction because it's something that hasn't actually been done before with this technology, which is obviously a horrendous misuse of Node because who the hell needs real-time sockets on a blog? Like, it's not necessarily the best reason to build on technology, but it was for better or for worse an interesting one. So that definitely helped. And how do you manage to get the first three contributors on board? And how big is your current team right now, both your people on payroll versus the contributors that do this? So we're up to eight full-time now, working for the foundation with two more on the way and about 120 total contributors, of which may be 20 to 30, are really pretty active. And again, this kind of came out of the community. So once we had this blog post, I kind of quickly slapped a newsletter form on it and said, sign up if you want to hear more about what happens to this, and then the Kickstarter, so we just built up this mailing list over time, which reached, I think, before we launched 30,000 to 40,000 people. And then so after the Kickstarter, we said, okay, well, we're going to build this now. We're going to do the prototype close source before we open it just to keep things sane while there's barely anything there. But if anyone's a developer and you want to help out, send us an email, tell us what you do, how much time you've got, and we're going to invite about 10, 15 people in while everything's still closed to help build this thing out. So that had a kind of two-sided positive effects. The first was that it got people to help us, which we really needed. But what really attracted them to it was, A, they could see Ghost had lots of traction or lots of attention so far. And because we said we're going to build in secret, basically, for the first three, four months, it was almost like a kind of VIP exclusive access. Like, you are allowed to commit to Ghost, rather than please will you come and commit to Ghost. Which kind of gave it this little bit more of an edge, like, yeah, this is a cool, like, I'll get involved and then I'll be one of the first people to write code in this thing, which seems like it might take off. Which wasn't actually really planned. We didn't think of it like, oh, how can we game developers into thinking this is great. It was more like, we actually do need a small group. We can't have everyone in here. So it just sort of worked well. So I was just wondering, how do you manage to, how do you manage to stop, like, feature creep issues? Are you facing that right now? And like, how strong is your hold on your critical map? Yeah, that's a great one. Feature creepers is one of my biggest fears, I would say. But because it's one of my biggest fears, it's one of the things that we almost swing so heavily to the opposite direction on. So because we've seen platforms like WordPress, like Drupal, like almost any open source products you can name really, slowly but surely over time progress towards this place where they end up with 700 fucking setting screens, and you don't know what anything does. We're kind of so afraid of ending up there that we almost just say no to everything. So you will see now even three years down the line, ghosts, when you log into it, is almost too minimal. And that's because we are really, really, really conscious of adding new features and letting new stuff in because it's such a slippery slope. When you just add that one thing because there's loads of demand for it, then you've set a precedent, you know, like the legal precedent for unlocking an iPhone for the FBI. And once you set a precedent to do one messy thing in user experience, then it becomes the basis upon which future decisions are made. So when the next messy thing comes in, even if it's not as bad or even if it's worse, some will say, well, you put that thing in, so why not do this? So we swing really heavily towards going no, not going to do that yet. We're going to keep building what we've got, we're going to keep cleaning up what we've got, and we're going to wait and see and test for what we're going to add next, which is very, very difficult and quite uncommon in the open source world, which doesn't usually have a heavy focus on user experience. It's usually more about configurability and customizability. But I would say so far it's proven to be a strength. One of the things that you did add to Ghost is that you chose Markdown instead of a traditional GUI sort of text editor. Yeah. And did you feel that, you know, your customers when they come in, they'd have to learn this Markdown thing. Maybe all the developers here, so most of us would know Markdown. So were you targeting just developers or were you afraid that, you know, customers wouldn't know this Markdown thing, then they have to learn it? You know, it really wasn't complicated, and this is, I think, one of the most important lessons you can take from looking from an outside perspective at any startup, any company, any open source project is it's very easy to look at what everyone else is doing and think they must have a really important plan for why they did that. But the simple answer to why was I just had a Markdown editor on my Mac and I liked it and I thought Markdown was cool. But it's, and you'd be surprised how common that is with reasons for why stuff exists and reasons, certain buttons or certain colors or certain people do certain things or certain companies. Like there's a story that came out just after our Kickstarter. I've got some friends on the Microsoft developer evangelist team in the UK and I've known these guys for years. They used to work with the WordPress projects and they'd host open source days and because I was one of the few people on the core team for WordPress in the UK, they would invite me to London and come talk to them about what Microsoft did with open source. So I got to know two or three developer evangelists at Microsoft and they're just really cool guys. Like it's, you say to them, why are you working at Microsoft? And they're like, because we think we can fix it. But they do, now honestly they're great guys, I've taken the piss, but they're great guys, they do great work and they just became kind of vaguely decent friends. And so when the whole ghost thing happened on Kickstarter, I called one of them up and I said, you know, this whole thing's happening, open source Microsoft, that's the thing that you like, you told me, you should, you know. And this was my friend Martin Beebe and he was like, yeah, yeah, let me see what we can do, but you know, it's Microsoft, it will take a while. And so about a week after the Kickstarter ended, he called me and he was like, yeah, we want to do something. And I was like, dude, do you know how Kickstarter works? And he said, yeah, but we still really want to support you. And I was like, okay, well, you know, the Kickstarter prices are all gone now, so it's going to be more than that. And he was like, that's fine, name a figure. So I said something reasonable. And he was like, yeah, cool. So Microsoft gave us this follow up sponsorship after Kickstarter. And in return, we made sure Ghost ran on Windows. And I think the only other thing we had to do was add a tablet at Microsoft Surface as one of our marketing images to the site. And that was it, simple as that. But the news stories that came out, oh, Microsoft throws its weight behind Ghost to compete with Tumblr, who Yahoo is just acquired. Fucking kidding me. I just called up my mate Martin and said, mate, you've got a marketing budget. Could you help brother out with his open source thing? And he said, sure. And we figured it all out and it was an awesome thing. But there wasn't a grand plan behind it. And there are a lot of stories like this. So be on the lookout occasionally when you think you spotted strategy in competitors and startups that you're working for and things that other people are doing. Because sometimes it really is just simple. There's no conspiracy. So another thing was I was so related to this, I was looking at GitHub. And I saw one of the repos that you're working on was like a desktop client. Yeah. Yeah, this is part of the product plan as well. Yeah, it's great. So Slack came out with their desktop client shortly after they launched. And it looked quite cool. We were like, this is nice. It works well. Didn't really think much of it. And found out it runs, it ran on MacGap, I think at the time, which is just this little OS X wrapper for, sorry, OS X wrapper. Apple people get very angry when you say that. OS X wrapper for JavaScript version of the site, basically. And then they moved it over to Electron, which was Atom's thing. And then a bunch of other apps popped up using Electron. And then almost exactly the same story. One of our developer evangelists friends at Microsoft called Felix Riesberg became a core team member on the not Atom, Electron team. And also was contributing to Ghost. And so we thought, well, it would be nice to have a desktop client. Why don't we start seeing if there's anyone who wants to build it? So we worked with one guy and he said, yeah, we'll have to get there prototype. It looked good. And then I said to Felix, could you help us build this? You know Electron. And he went, yeah, sure. So he's actually flying in tomorrow here. And we're going to work on that this week, which would be pretty cool. If anyone wants a sneak peek, I'm working at the hub the rest of the week. So you can come and see it in action. So there's a quote online. No, no. Not true. Oh, you don't want to reject this. So I'm going to read it. Apparently, you said, if Facebook came to me tomorrow and said, here's $1 billion, we want to integrate Ghost into Facebook. No matter what the human side of me wants to do, I cannot. So does this still stand today? It does. It does. And this is, this is one of my favorite subjects to talk about actually. This is, this is where we are super weird compared to your average startup, really compared to 99.9% of other startups. So Ghost is built as a company and an organization on the basis of what would a company look like if you tried not to get rich from it, right? Which is the opposite perspective from basically everyone else starting a company. And there are two main reasons for that. The first one is I spent a long time trying to kind of figure out what I wanted to do with my career, with my life, with everything. And asking myself that same question that you always ask yourself when you're trying to figure out what you want to do with your life. And you go, what would I do if I won the lottery? And I had $100 million. And the beginning of that game is really easy. You know, you go, I'd give my boss the finger. I'd buy that orange Lamborghini downstairs. I'd fly around the world. I'd start a company. I'd do all this cool shit. And then it gets a little bit harder. And then you're like, okay, well, once you bought everything, then you have to go and learn stuff or give money away or help charities or something. But okay, let's say you've done all of that stuff. It takes maybe two, three years until you've burned through enough money to kind of satisfy all that bullshit. And then, okay, what do you want to do next? Like, what do you actually want to spend time on? Assuming you're going to live, like, let's say optimistically till you're 80. What do you want to fill that time with? And for me, that answer came sitting on an island in the Philippines, which is actually where I grew up, but I'd gone back there. And I was sitting there with a couple of friends. We were watching the kite surfers on a little island called Boracay. And I realized I'd be doing exactly this. I'd just be hacking on some open source stuff. And I'd be traveling around kite surfing when I felt like it and having enough time to hang out with cool people. And it struck me at that moment that I was already living that lifestyle on what, maybe 40, $50,000 a year, maybe less. So then why go to all the hassle of trying to become a millionaire? What a waste of time that would be. So from that moment, my motivation to make a shit ton of money would just disappeared because I thought there's no point. And then the second perspective was seeing all of the other open source projects that are around that start out with really noble ambitions and then over time morph depending on who buys them or who acquires them. Express was open source and great until it got given to Strongloop and then Strongloop got acquired by what, AOL or someone? I don't even remember. Even worse. So you see these things start out as good and then slowly over time just disappear and get manipulated and transformed by whichever corporate shareholder happens to have them most recently. So I thought, what if you started a project and a company with a sole ambition of not trying to get rich off it? What would actually happen? How would that influence the product? How would that influence the customer base and the developers and the people in the community around it? What if you came at it from a totally different angle? Well, there was absolutely no goal of selling out at the end. And so that's the reason we set up as a non-profit. So I own Zero Shares and Ghost. My co-founder Hannah owns Zero Shares and Ghost. It's a company limited by guarantee, which means we're trustees, which means we can steer the ship. But if at any point we quit, there's no big severance package, there's no bonus, there's nothing to sell, we can simply hand over the trustee ship to someone else. So I will never get rich off it. And Zuck can never take it over. Which has its drawbacks. It makes things very hard. There's so much we could have done with a bit more funding than we had. A lot of extra opportunities we could have had. But those are all in the short term. Short term, it's hard. It's hard to build something when you're purely bootstraps and you have no other option but to be bootstrapped. But I think long term, I think in the next five to 10 years when we look back at it and once our competitors have burned through that money and tried to use advertising and failed and fallen over and we're still around, even if Hannah and myself aren't around, someone else is still around. I think for the longevity of the projects, I hope it will have proven to be something that's a powerful idea and maybe something that will translate into other projects as well. Have you seen the new spin-offs from Ghost? Spin-offs like Fox? Yeah. Not yet. There's been quite a lot of people who've tried to recreate the Ghost admin on top of WordPress, which is quite weird. And they call it things like, gust. But no, no direct walks. Shot for disgust. But weirdly, that's actually going to get easier in the future because WordPress are adding an API now. So you could conceivably actually, someone already hacked this, you could use Ghost admin as a client via API to WordPress. The only kind of disadvantage or the kind of sticking point where it trips up there is WordPress have built WordPress, right? And then they're tacking on a JSON API. So the API basically just goes to PHP and does an equivalent function and then sends something back, like the API is just over here. Whereas Ghost is API at the core and then we build stuff around that. So you have the server side, which is an API and then a client, which interacts with it. But WordPress has, the API is going to take a long time before it really gets full coverage of everything. So it'll be interesting to see how it develops. They're very excited about it. I'm quite interested to see where it goes. So we're going to stop asking the questions and open the questions to the floor. If there's any questions you guys might have. Yeah. How does it, how does it tool change? Like what tools are you going to use for the management project? Is it all Github? Pretty much everything Github tests through Travis. That's, it's pretty much the same stack for all open source projects, which is Node on the server, Ember on the client, Travis for tests, Github for managing everything. We still use Grunt because we haven't yet been hipster enough to switch to anything else. And then on the, our closed side so far, which is basically the website ghost.org, we use currently Ruby on Rails, which causes us nothing but hell. Like legit never make me install a Ruby environment on a Mac ever again, because holy fucking hell, I don't even, but we're working on switching that out to this exact same stack as well. So it's going to be Node and Ember. Should be good. I'm preaching, you have a question? Yeah, sorry. Yeah. I actually found out about ghost through when you all took over Rune IO by Samsung. Oh yeah. Like, so how does that, how did that work out? Like how did Rune's code base integrate into the ghost every day at all? It didn't. So what just happened? So Sam, Samsofts, who's apparently doesn't know engineer based in San Francisco, likes working on lots of little projects with Drew Wilson, his friend set up a blogging platform called Rune.io. And it's a simple thing, quite similar to ghost, except it was just this close source single thing, and that was it, you log into it, like a Markdown version of Medium kind of. And it got a little bit popular, had about 80,000 users, but it never went beyond that. And at a certain point, after two or three years, Sam got really fucking bored of it and stuff kept breaking and he didn't want to fix it. And it cost him about $2,000 a month to run, but it made him about $300 a month in revenue. So he was having a bit of an existential crisis towards the end of 2014 of like, what do I do with this? I have all these users and all these bills. So he was having a little tweet storm about how depressing it was, about it not working. So I just sent him an email and I said, well, like if you don't want to do it anymore, we can help your users out and just make it an easy transition. And he said, sure. So we just organized it like that. It was very, it wasn't, we kind of framed it as an acquisition to try and get a bit of marketing buzz out of it. But it was more just, we made an easy path for all ruin users to move to Ghost. We gave all of his customers free credit for two years on Ghost. And then we added Sam to our advisory board, which we just talked to him more often about what would be good to do in the product. Right. So assuming that we have a lot more funding, but that we're keeping for Ghost, well, what else do we do? If we had a lot more funding? Yeah. Many things. But I think the biggest one is hiring. We move slowly. So the team is eight full-time now of which three and a half, the half being an intern are engineers, which is really an absolute minuscule full-time team. And with that team, we support 600,000 users, 30,000 individual, no JS applications, hosted on an infrastructure of about 16 servers, serving north of 100 million requests a month. So we are spread thin. And keeping up with running that is hard enough. And then actually rolling out new features or fixing bugs or doing the things that we have from this community that's demanding new stuff all the time is slow going. And there's so many things people want us to fix and do. And we're like, yeah, that's cool. We can fix that and do a small thing. But what we really want to do is work on this big project that we've had brewing in the backs of our minds for the last three years. And we don't have enough hands physically to do it. So hiring people is the number one thing. I don't think there's anything else to really spend money on to be honest. Like we're a distributed team. We don't have an office. We don't have overheads. This is the only known ghost t-shirt in existence. So yeah, maybe I'd buy some t-shirts for the team, I guess. So full disclosure, this is more of a question for Hannah. My role has always been the business side as well as the front end. So HTML, CSS, SAS and design. So my knowledge of JavaScript is enough to get the full big picture of how our stack works. But sadly not quite there enough to tell you what key problems we had with Node.js and how we overcame them. I will say that one interesting decision we made in our architecture when we started out was to effectively bundle everything into one repository where the Node way might have been to modularize it far more early on. So our client and our server sits in one repository at the moment. So most Node people who come to the project go this is a fucking mess and leave. But what that allowed us to do in the early days of ghost was to hack things very, very quickly without needing to create a package and a module for every little thing we wanted to do and then tying it together properly we hacked everything very, very quickly in a single repository. And then as we start to figure out what modules make sense and what structures make sense and the product matures then we start to break things out of that kind of mess of a repository into individual microservices which then get pulled back in as modules. So we're probably just now about to divide out our client and server apps into separate places now but working in that way early on got criticism from more senior developers who are used to doing things the proper way but it allowed us with a very small team to move very quickly because we didn't have to conform to doing slightly more proper architecture early on. So I'd say that was at least one architectural challenge that we overcame. You have a question here? In terms of open source governance how do you balance between the node foundation focus and as well as external projectors? Yeah, that's a great one. I would say so the foundation pretty much steers the ship and because we're a not-for-profit organization and a foundation we get very, very little pushback on that so where you might have an open source project let's say run by Airbnb they have to first consider the motivations and the ambitions of Airbnb and then also of the outside stakeholders and then their ideas of what is the most important thing might not completely align because Airbnb might be optimizing for customers revenue whatever outside shareholders might be optimizing for other stuff we have pretty strong alignment so whatever the foundation is doing it's only in the best interest of the product of Ghost itself so external contributors for us are very understanding very on board basically with what we do but yeah, essentially how it works is the foundation pretty much leads the direction of products and of roadmap and then we do our absolute best to involve the community, involve external contributors in the process as much as possible because if you don't get buy-in then it's not going to be a problem but you're also not going to get any help so involving external contributors in the process as much as possible is what helps then get people writing code if you just shoot down ideas and say we're going to do this people aren't really going to help you so it's a balancing act but we're in a fortunate position of having it a little bit easier than other people Any other questions? Open source, business models open source technology but it's for profit or profit I mean very yet in terms of sustainability No No I'd say we've hit a unicorn open source mix and by unicorn I don't mean billion dollar valuation, I mean rare I mean diamond in a sea of shit Funding open source is hard it's really really hard and you know what if I was going to do this again with anything other than ghost I probably wouldn't do it the same way for ghost it makes sense we've got a product that needs a server and you can sell a server and you can sell managed service around a server but for anything else if you want to build an open source app which doesn't need a server it's really tough to manage and the selling support shit doesn't work no matter how much open source hippies and you really need to understand there are two camps there are open source pragmatists who use open source to their advantage because it makes sense and because it's a good thing to do and there are open source hippies who just think the code in the world is sacred and should be given around like a joint at a party like separate the two open source business models are really hard if I was going to do anything like this again that was not this product I would have open source components so I might have an API which was open source a client which was open source a set of libraries which were open source and open them to contribution but make the final product something which we bundled together which is pretty much what github do actually 80% of all github's code what makes github.com is open source components so they have open source bits and pieces that make github work but the final bit of work that ties it all together and makes it be github.com that's their core value that's the bit they keep closed and I think that's a more viable a more commonly viable path for monetizing open source but yeah it's not easy allusions it's tough it's really tough so there's just one thing about closed source bits a couple of things one thing right now speed that's what I always have to say just because we don't even have to do anything install something on node and it's like a fricking Ferrari right out of the gate so if you load tests ghost and wide press side by side on the same environment ghost will be about 20,000% faster so we'll respond to 20 requests in the time that WordPress responds to one after that I'd say design how about the static site generators static site generators usability and design in a room filled with developers this may not resonate but static site generators aren't actually easy to use pushing something to get and then running a deploy web hook isn't what your parents, uncles, sisters do on their weekends having a user interface which is well designed and works and does what you want it to do is valuable for a mainstream market and we still haven't hit the mainstream market because we're still not friendly enough but I would compare that to something like Linux versus Mac you know developers like Linux is easy to use what are you talking about this is amazing this is the best thing ever and the rest of the world is like static what generator so yeah different markets any other questions? what's your vacant list of what's your vacant list of your parents your laugh sqlite no? okay usually in a room like this I get laughs from that so yeah when you build a blog you either go to the hack and use route and you first say Mongo, get shot down say Redis, get shot down invent a new database go off and write that database for three years come back and use that database to write something by which time you've forgotten what it was you needed a database for or you go to the pragmatic route and say what does this application actually need to run okay so a blog has people who visit it they don't use the database they're just gonna cache all that content anyway they're gonna hit a cache a person who's logged into the admin who's writing a post okay so what's the database load gonna be well we've got average size of one person per blog logged in maybe going up to 10 people per blog and the most simple database in the world sqlite which is a file based database can handle 100,000 concurrent connections and by the way powers cruise missiles did you know that? like if you don't think sqlite is production ready it powers cruise missiles so we thought make it simple grab sqlite you don't even need to install database on your computer to make it work when you npm install you get a database and you can then edit that file database in your finder or whatever else you use so it's the simplest possible thing now the only exception we took from that was thinking a little bit ahead to well maybe one day we want to power mashable and mashable does have hundreds of authors and it does have a commenting system which does tie into the database which may have super heavy load so we put an ORM in between so we ship with sqlite but it has bookshelf as an ORM so you can install mysql which is what we use on our own pro-service it supports pro-scress and it will support anything else that anyone right to support for so far I think it's just those three yeah? yeah so we started out we thought okay so we're going to make money off hosting it's actually an interesting story so we need to like have full control over our servers we can't be messing with vps so we're going to go full hardware that means we have maximum control for all our resources we can do anything we want so we bought like 15 Dell servers suck them in the data center in the UK two years later figured out that was a fucking horrible idea and as soon as we wanted to scale we had to know like guys in two months we need to scale we need to order a server from Dell but so what it allowed us to do and this is actually a big secret I don't think I've talked about this publicly yet so what I was just saying about not needing the database for a blog what if you actually also didn't need the Node.js app for the blog because in terms of server resources every time you have Node.js app running you're straight away at kind of 60 megs of RAM and then maybe creeping up to 90 as soon as you're doing a couple of things so we host 30,000 individual Node.js applications that's a fucking lot of RAM to use how can we optimize this just shut the apps down just cash everything so for the first two years of ghost life we have 30,000 blogs serving hundreds of at that point tens of millions of requests per month from two cash servers sitting in front of two application servers because we could fit about 6,000 blogs on one server purely by starting them up on the fly shutting them down on the fly so we'd have a few hundred applications running at any point in time when someone tried to log into their blog there'd be a half second delay as we quickly went and said turn that one on and then you'd be logged in and you could write your blog and then once you hadn't been using your blog for a while and someone else tried to log in yours would shut down and the other one would start up so we ran this massive infrastructure a massive amount of stuff on a tiny tiny infrastructure for a long time until we hit the wall until we started wanting to build features where weird you do actually need Node.js running all the time and needing to scale and that started having a long lead time so then I called up a friend of mine at DigitalOcean who's again one of their evangelists and said you know we refer a ton of customers to you maybe we should do something more together and he said yeah actually you drive about 60% of all of our customers each month and I was like what the fuck so we said how about we just move our whole infrastructure to DigitalOcean we'll put a DigitalOcean logo on the bottom of our site and we'll write about how we use the infrastructure and you guys help us out with the hosting side of things so that's what we did middle of last year and now we're just rewriting the whole system which starts up and shuts down blogs and keeps the correct ones running instead which means we can now scale cheaply and easily we actually have someone else who's worrying about the hardware rather than freaking out every time there's a power outage in London at 2am and all of our data is now located in Amsterdam as opposed to in the UK in one of DigitalOcean's data centres which is great because Amsterdam is one of probably the best most secure places you can house any sort of data so it's important I feel like I've told you guys all of the weird things that we did that we shouldn't have done and we're now fixing but I don't know this is kind of the story of how a ghost god hears you don't realise how much broken shit is behind everything that looks good if you ever look at anyone's business or app or products and you think man that must have been easy and it seems to work great it wasn't easy and it does not work great there are so many janky patches everywhere in our infrastructure and everything we do and I know the same is true from other friends running similar sized companies just holding stuff together until you can afford to invest more and fix it and build more but yeah it hurt for a long time we've slowly learned to make peace with that process and to just be okay with everything being very close to fucked at any given time and learning to sleep knowing that that's true any other questions from the floor how much work does it take for you to set up a foundation and how is the progress on winning this company? yeah so that's a really interesting one in the UK it turns out super hard because you call up any accountant and you say I want to set up a non-profit organisation and they go why would you do that I'm not kidding I'm not kidding honestly like 20, 30 different agencies and accountants not only do they say why would you do that when you then say oh yeah no idea like it's not something they do know how to do we went through so many agencies which actually knew how to incorporate a company limits by guarantee using the word foundation all these requirements we had conversely came to Singapore met up with an agency called Rickvin which I assume some of you have heard of who do incorporations and we set them this is what we do, we wanted to move to Singapore we have a CLG with standard terms we'll get that done and we're like what first of all you know what that is second of all there's an acronym for it awesome so they had it sorted immediately in terms of their side of things I will say that the incorporation in Singapore for that structure of company took a long time I think when you set up a standard limited company here it's like a two day situation ours took about two months we had to go through Accra for extra approvals to use the word foundation in the company name which is regulated, you have to provide some sort of public benefit and show, prove all your documents as to how you're going to do that but the process was easy it just took a long time it's right now ghost foundation limited and we can apply to have the word limited removed as well but that's like an extra thing you have to go through a whole load of paperwork for so in the UK we do because a company limited by guarantee is not a charity it's just a normal company with no shareholders however I've heard a couple of people now in Singapore tell me that companies limited by guarantee in Singapore if they prove that they're for public benefit and not for profits by the owners essentially have the possibility of being exempt from income tax sorry corporation tax but I'm not sure yet we're still looking into whether or not that will apply to us it might or it might not so are they shareholders and directors we have directors but no shareholders so what was it? nobody so the assets belong to the foundation okay so they are trustees of the foundation which means you and your partner yes me and my co-founder and the trustees yes so we act as shareholders without receiving any benefit exactly great question Singapore is number one in the world for a ton of stuff now I don't know if you know this I learned a new one today so it's like number in the world for ease of doing business number two in the world for lease crops something number it's all in the blog I put like all the number ones and there's a lot of impressive number ones but I discovered a new number one today I'm so excited to tell you this I was walking here from my hotel which is in little India and I saw two different buses advertising I don't know if you've seen this the duck and hippo safari guys you have a safari with one duck and one hippo this is freaking incredible I want to go on this and on your left sir you will see our resident duck and on the right the hippo amazing so that's the reason so after being in the UK for a few years we became very frustrated with how unfriendly the UK Europe but especially the UK is towards small digital businesses all their laws now are essentially centered around preventing Apple and Google from avoiding tax loopholes so everything they do is just like how can we stop Apple and Google from doing this and that and the other not even thinking about what knock-on effects it'll have on small businesses so the latest iteration of that is VATMOS which is a complete disaster it requires every business no matter what their income is online to charge the correct tax rate in the country where your customer lives in Europe which is different across all 82 possible variations of that and to cut a long story short every small digital business no matter what you do your payment transaction must correctly account for, issue an invoice for and then subsequently pay the tax for 6,882 possible combinations depending on where your customer is and you have to collect the data on where your customer is and hold it for a minimum of 2 to 3 different kinds of data and hold it for 10 years all of it correctly now all of this was done to stop Apple and Google from funneling all of their income through Luxembourg which has the lowest VAT instead what it's done is fucked every small business out there and a lot of people are just in Europe stopping selling to the rest of Europe which is a complete disaster for us a complete disaster and we figured out that it would make a lot of the things we want to do like a two-sided marketplace where you have 6,882 combinations of possible tax bans just completely impossible so we thought well we're a distributed team I'm actually based in Egypt most of the time Hannah's in the UK we have a couple of people in the UK one in the US, one in Austria people all over really so why are we actually in the UK it's an online business, our customers are everywhere our staff are everywhere we don't have investors who need us to be in a particular jurisdiction to invest so why are we actually in the UK or should we be somewhere else so we started looking like where else could we go and there were a few possibilities that seemed interesting but the biggest hold up was Stripe support because our whole business runs on Stripe and retooling that is basically forget about it we love Stripe, we need Stripe Stripe is what everything runs on so which countries not in Europe they charge you 72,000% tax have Stripe support and Singapore had literally just gone into beta for Stripe support so we started looking into start-up scene in Singapore what's happening in Singapore, what's going on in Singapore how incorporating a business works in Singapore and the further and the further we got into it we just realised it would just be really a place to incorporate an online business without investors without external pressures to be very sensible to be based from and it seemed like a very very attractive option so then we flew over here, talked to Rick Vinn talked to everyone and look around and went from there but yeah Stripe support plus sensible, sane taxes I think taxes in Singapore just make sense particularly the corporate taxes the admin, the sensible ease of doing business all the filing, you guys have it so easy compared to the US and Europe when it comes to just filing and doing admin it's incredible so just a very attractive place to base a business from I think when you start a business first of all you don't know how much shit you're going to face that it has nothing to do with your business particularly when it comes to tax accounting admin and as I was talking about running the size of what we do with such a small team and being spread so thin the number of things we can focus on at any given time and we didn't know what the perfect solution would be but what we did know for sure was that the amount of stuff we had to do to comply with all of the crap in the UK and in Europe wasn't working for the size of team we had and it was holding us back from being able to actually work on the products and do stuff so the biggest reason to move here was actually just to make our lives easier and I think that's a lot of what we optimize for in general these days is what things can we do what ways can we operate in just to make it easier for a small team to do more stuff you mentioned that your team is very dispersed so I was wondering what kind of models have you come about in terms of working together yeah so we fully distribute a team which is not that weird these days automatic do it, buffer do it, Zapier do it it's Zapier apparently not Zapier and they do it up to two, three, four hundred employees these days but it's getting easier and easier I think back in the early days of distributed work people would use WordPress blogs to keep up for us now it's pretty much Slack like just Slack for almost everything we use Zoom for our weekly team meetings video and audio which is great a million times better than Skype or Google Hangouts because fuck Google Hangouts and Slack are just introducing calls and videos that's going to be even better I think that was their screen hero acquisition so it's mostly Slack, bit of Trello bit of Asana, GitHub and it works, it's surprising honestly it's not that weird when I worked as a freelancer all my clients, I was in the UK all my clients were in the US and I was already working remotely so I never even considered should we open an office and go and sit in a room together all day no offense I figured I quite enjoy travelling and going to nice places like Singapore, I quite like to keep doing that so why don't we just do this remotely because I can look at my computer screen from here in front of you or from the other side of the world and it makes no real difference I think until you've done it it seems a lot weirder than it actually turns out to be it's a very easy way of working so it has the occasional hard things like you don't really know when someone's in a bad mood or if someone's super hungover so Slack on doesn't change to hungover so sometimes it'll be like hey I've got this great new idea and I'll be like fuck off, talk to me in an hour I haven't had coffee and vice versa I'll do the same thing to her and then you have to remember I don't actually know what zone this person is in so you need a little bit of extra effort to communicate the weird subconscious subtle things that you would just get from sitting in a room with someone but other than that it's very natural I'm sure you guys all use Slack or something here or in your respective offices anyway right so imagine that but not in the office and there you go so we have time for one more question of this anymore no we're good do you have a question for yourself? a question for myself, oh my god there's never been given this opportunity before oh man I should have thought of this in advance I don't think I do, no I do I thought of one today this is not really a question but I just thought of something that maybe is interesting to talk about I was sitting in the Hub thing co-working space this morning and they have apparently they have all these mentorship programs for startups who are based there and I was just sitting there at one of the tables casually doing my work and I couldn't help but overhearing at the next table what I presume was one of these advisor people giving a young founder a lot of really really bad advice and she was all over it she was like you gotta get on LinkedIn and I was like oh no you don't and she was like you gotta get on LinkedIn and I've got loads of connections on LinkedIn and let me tell you that's where it's all going down and all saying these weird buzz words that you might hear at some sort of white combinator bootcamp but then turned upside down so they didn't actually make sense and I was thinking to myself because I've never been based in a startup hub of any kind I've been in the countryside in the UK I've been in the middle of the desert in Egypt so I've been sitting on my ass on a beach in the Philippines but I've never been based in any startup hub and I was thinking to myself if I was that founder and I was sitting at that table and I was receiving a lot of really bad advice I wouldn't know it was bad advice and that's kind of scary so I think the one big takeaway I've had from as a founder in this whole startup experience is there are a lot of people who are going to give you really bad advice when you try and start something and some of them do it because they believe it, they don't know any better some of them do it because they just have some sort of ego trip and they feel like they can ride your coattails to success but I would definitely recommend taking any advice you get with a grain of salt and it's usually the people who are most sure about the advice they give you and who say you should do this and this is stupid that are the ones who aren't really sure the smartest people I know and the ones whose advice I trust the most are the ones who are the least sure of the things they tell me and they are the most humble and say I'm really not sure but here's a couple of thoughts I have based on my experience so it's always worth remembering that if you tell someone a really bad idea they'll say that's a stupid idea and if you tell someone a really average idea they'll go that is a great idea and if you tell someone a really great idea they'll go that's a stupid idea so the only type of idea you will ever get positive feedback from is an average one the other two are either really dumb ideas or really great ideas Airbnb being a perfect example everyone told them that was a stupid idea but the flip side of that is they just have a really fucking stupid idea so when it comes to advisors try and find people who've done things recently that you've been doing yourself and ask them about their experiences and I think I'd say try and steer clear of people who've either been there and done that a long time ago or people who haven't really been anywhere but still want to dole out a lot of wisdom because I think they can lead you down the wrong path as I saw today I'll leave you with that thank you very much so I guess that's it for today feel free to check out Ghost try to deploy yourself with your music free oh and we are hiring I should have said that earlier ghost.org slash careers if anyone's interested in this weird remote working thing got a couple of openings for Node.js developers Ember as well any senior quite experienced JavaScript developers very very interested cool thanks guys