 professional services teams hated the product as much as our customers. That was embarrassing. Okay, not the, you know, and this is, this is sometimes the reality of people who build products is that, you know, it just doesn't get to where you want it to be right away. So a few years ago, Barry and I were an employee three and four or something like that for AQUIA and we had met at online through the Drupal community, but we'd met in Barcelona in 2007 when we were in stealth mode with AQUIA and we were actively in the process of raising our Series A and we went ahead and we were fortunate over the next several months as we did some validation around the business that the Drupal market was going to be big enough to support a startup servicing that community and we were fortunate enough to get a Series A around of seven million dollars. And so, you know, that presented a really wonderful opportunity for us to build products for our potential future customers. And I'd like for you all to take a moment and just think for a second, what would you do if somebody were to particularly dress, were to come to you and sit down with you and give you seven million dollars and say, let's go build some great stuff and get some customers? And to answer, it is, it is really, I remember that Karen and I and Jay Batson were sitting in one room in an office in Andover looking at each other saying, who was going to pay us and for what? And we knew, okay, Drupal and services, but what exactly? We really had no idea. So, I'm going to tell you the story of what we did. But before we got there, we didn't have when we started AQUIA the insight of some of the great tools that are out there that are available for you today. In particular, a project that was released this last year called the startup genome that went out and interviewed in fairly in-depth process over 600 startups and to see what the code for cracking innovation and innovative new products was. And in that paper, and you can go and you can register for it in the startup genome and download it and I encourage you to do so, they found that startups move through six phases and if they move through those six phases in order, that they will be successful. And if they find that they skipped those phases, they are unsuccessful. And that was the conclusion of the research. And what they said was that you first you go through discovery, then validation, then efficiency, then you scale and then you hit profit. And then you hit renewal or decline. And basically, this is the core crux of the issue that we were at. We had the money, we had a baseline, but we somehow had to get to this ideal killer product, right? And how do you get there? Do you get there through brilliance? Do you get there because you're a genius? Do you get there because you have great architecture or great design? Do you get there by listening to your customers? How do you get there? Right? Not so I think what startup genome tries to talk about. There's also been a great book that every executive at Aquia and our leadership team is required to read called the Lean Startup. And this has become very popular, New Relic, a lot of companies are offering this book for free to read, but it's basically this idea of how can we develop products in a really efficient and effective manner quickly. And they start, and most of these are targeted towards technology companies, but they start with technology commoditization, right? Something, a great tool like Drupal, which is a basically commoditized content management, has commoditized social features, has commoditized the distribution of content through channels. So that's one. And you all have access to Drupal and you can all come up with product ideas and go build a great product with Drupal. But it's not enough to just have the technology. The next thing is you really need to come up with something called innovation accounting, which is really how do you measure the innovation that you're building? How do you go out and design something and create it and get some data about how it's doing and then get a feedback loop? And we'll talk a little bit more about that. And then that's not all. Not only do you have to have great engineering processes, but you've also got to do something called customer development. And you've got to go figure out how to listen to your customers. And the cool thing about the Lean Startup and the Startup Genome is they found that companies that could figure out how to listen to their customers correctly grew seven times faster. So growing really fast and maybe not needing that next round of funding, something that's incredibly a good opportunity. And then last but not least was another book in this series by a guy named Stephen Blank called The Four Steps to the Epiphany. And it was really about customer discovery, customer validation, customer creation, and then building the company. Okay? So as we're talking about these things and you're thinking about your customers and you're thinking about your products, note these down and come back and research them because you're going to find them to be really valuable for the product that you want to build. So for us, when we first started in this little office in Andover, we reached out to two members, prominent members of the Drupal community, Mike Myers, who had raised $14 million to build out his website Now Public and Ken Rickard, who had worked for a large newspaper chain, two of the people who were building some of the largest Drupal properties out there. And we brought them and flew them up to Andover. Do you want to talk about what we asked them? I don't remember in exact detail, but we were forming the basis of the first thing we were going to offer, which was the support model, that Acquia would answer questions that you had about deploying Drupal for your enterprise. And I remember we drew up, we wrote up on the board like 10 or 15 different things they might want from Acquia that they might want help with our services. And we told them you have 10 points go vote on the board, however you want to distribute your points. And so we were just trying to get from them, like what is it that you want? What we came away from in those very early meetings, this is like January, February, March, 2008, was that the first major impediment for these companies, these two companies, I mean they had Mike Myers and they had Ken Rickard, right? But a lot of other companies that wanted to use Drupal didn't have these well-known members or the very experienced members. And so their concern was I deploy the software and I have a problem with it. Where am I going to go? Am I going to post a message on the D.O. forums and the mailing list and hope I get an answer? That wasn't enough for large companies. So hence we landed on this support model as the first thing to do. And so what we validated was, and the success of the meeting was if we were to offer it, they would buy it or something like it. So we moved to the next second stage of the startup genome, which is learning. And as a startup you might be thinking what am I learning? Is my job to build a product? Is my job to make a profit? Is my job to hire employees? Is my job to get a great team? Is my job to disrupt the marketplace? And really as a startup, and we maybe didn't know this at the time, but we really were answering two questions. Number one, can we build a sustainable business? A business that can self-fund itself? Keep going on with customers and stay profitable and get the payroll paid. And then the second one is, should we build these products? And really what we were doing is, as a startup, this was an exercise in learning. How fast could we learn? How quickly could we learn? How could we validate our assumptions as part of our learning? And that drives a lot of what we're doing and why you spend time with customers is to really learn. And so the biggest cost time advantage is that don't build features that your customers don't want. How many people have worked with an engineer who has a great idea for a feature? How many of you have failed to sell it? The more hands went up in the second question and the first question. And so it's really critical that you build something that your customers actually want to buy. So this is where we can describe the first major pivot that occurred to Acquia, the first bit of learning that you might say we had foisted upon us. So go to the next slide. So we decided, okay, we're going to provide support and we're going to do it through a portal called the Acquia Network. But we don't just want to be a support company. We also want to be a services company where we build products which provide services to customers. So we built this thing called the Acquia Network. We hosted it in EC2, which becomes a much bigger part of our story later on. But so through the Acquia Network is how you access your support. So it was a website that you logged into your support tickets and then we thought, okay, we want to provide some services. What kind of services can we provide? And what we came up first was with, you know, customers want to know that their websites are up. So we built the first service of the Acquia Network which was the Heartbeat service. Did anyone in here, was anyone here a customer of Acquia before like 2011? Did you ever use the Heartbeat service? I see a hand going up. Was that useful to you at all? Kind of? Thank you. Very kind. So we learned two really critical things. One is we put a fair bit of time into building that service, but we didn't really validate very carefully whether that was exactly what customers wanted for a thing that measured the uptime of their site. In fact, the way it worked was only when your site ran cron did we ping you and check whether you were online. And then if your site didn't run cron like you chose to turn it off for some reason, we would start sending you alerts that you were offline. You're like, no, I'm not offline. I just stopped running. I feel like it didn't make a lot of sense. We didn't really do a great job of validating it. The other more important thing was that we spent, you know, when we brought Ken and Michael Myers up to Acquia and asked them about support, we were asking them, you know, are you interested in if we, you know, if we support Drupal and a certain core set of modules? And they said yes, answering the question, we would like support for Drupal. And there was this really subtle disconnect that was happening that they thought we were offering to support anything in Drupal and we thought we were offering to support a limited subset of core plus the top 100 modules or something. And so we launched the Acquia network in September of 2008. And nothing happened. What we discovered, the learning we had foisted upon us was customers weren't interested in a subset, support for a subset of Drupal because 100% of them had a custom module, a custom theme, something that wasn't in our list, something from the long tail. And so the box that we had drawn over the support we were offering did not fit any of them. And, you know, the sales bell didn't ring for a couple months, at which point we engaged on our first major pivot. And so if you're not familiar with a pivot, a pivot is when you basically do a course correction and you reevaluate whether you should continue to develop the product in the direction or do you change, right? And so there are a lot of good examples of companies changing direction. And I think the key point here is that that first phase where people who use Drupal really think about it, it's like, if I take Drupal and Cloud, amazing. I've combined two technologies that are really cool. Now I have a product that's successful, right? You know, Drupal plus Node.js, awesome. Drupal plus Mongo, awesome. Drupal plus Redis, awesome. Pick your technology and people will think that if they combine the two. Drupal plus CVCRM, awesome. And people think it'll sell. And we learned that it's more than just, you know, the introduction to technology. And so they come up with this concept called innovation accounting as opposed to having a long project roadmap with milestones and a total number of users and things like that. The idea is to really get a flywheel spinning of getting ideas and code and data and learning and building and measuring. And the whole point here is that when you're in a startup, or you're building something for customers, you really have no idea what you're doing. You think you know what you're doing, but you really have no idea what you're doing. And so you're trying to reduce uncertainty by spinning that wheel really fast and getting around that first loop of building something, getting it out to customers and testing. And so there's an actual engineering process and some practices around angel development and minimum viable products. We were doing this back in the summer of 2008. We were using an agile software development process called Scrum. And so every three weeks we would have a cycle where we'd talk about the features we wanted to build and we'd build them in that three weeks and then we'd deliver it and we'd have an internal demo meeting and show everyone all the cool stuff we'd built. And we're still using that to this day and it's actually a really great development methodology. But I think looking at this slide here of the innovation accounting, what I'm basically trying to explain is that we didn't get this right in 2008. So we did the build and even though we were operating in a very fast agile process we were shipping internal releases. So we were shipping new releases internally every three weeks. We weren't shipping them to customers every three weeks. We didn't ship them to customers until September. Notice that we started in January. So we spent a way long time in that build loop and then we finally shipped it and when we hit that measure loop we told you what we measured. What we measured was zero. Very clear signal, easy to understand and hence we had that learning box. So this is the history of how we went through learning this the hard way. Anyway. In a startup, no facts exist inside the building. Only opinions. And so when we first started we had started with our CEO, Jay Batson and we had a chairman guy named Tom Erickson who's now the CEO of Aquia. And Tom said he wanted to see what we were doing and most of our business model was based on inside sales. We would do some marketing, people would come to our website they'd see the phone number, they'd call, we'd answer the phone and we'd sell them something. But then we decided well there's a lot of people using Drupal out there why don't we get somebody out in the field. So I had been running Drupal cons for a long time or for a while and knew a lot of people in the community and so I decided to call my friends at Sony Music. And they said you're gonna go see Sony Music? Sounds like fun, I'm coming. And so off we went to Sony Music in New York City and I went inside and hung out with all my friends at Sony Music from Drupal Cons and staying up all night drinking beers and chatting and stuff like that. And we got in there and I presented our support plan and they stood up and said awesome we want support we use Drupal we've got a couple hundred sites we want that enterprise support we'll buy it. And my CEO said congratulations you have a new job. And now our field sales guy at Aquia. And I was like what? I want to go back and be a community manager. But the point was that we had an opinion inside that we could sell support but it wasn't until we got to Sony Music. And then at Drupal Con DC I'd been calling the Obama campaign staff because they had launched the Office of Management and Budget through the White House had launched Recovery.gov. And Barry and I were on the security team and we were terrified that the highest profile Drupal site Recovery.gov in the world was going to get hacked. And when we pushed out a security release we had to know that they were going to update that code. And so Barry was calling the White House phone switch line and saying please could you get the Obama staff to go and update CORE. You know there's nothing to that extent. You know I mean yeah but I mean we didn't know we were doing. So we're at Drupal Con DC and I get a phone call. I'm like hi and it's like hey it's Dave Cole from the White House and we really appreciate the contact and the Obama staff campaign staff letting us know that Drupal Con is going to be in town. We really appreciate that you've offered us tickets to come to Drupal Con and give us a tour but we're not going to be able to make it. We're really busy. I'm like totally interested. Like however would you be willing to come to the White House? I'm like hmm. Yeah I'm available. And so I went and I grabbed Dries and somewhere where you were there you might have saw it on Twitter. At DC and Dries and I off we went to the White House. Now for the last five years I'd come from IBM and had been doing enterprise stuff in IBM research so I saw Drupal as this thing that was going to mature so that big companies would use it. So we were really dedicated on security. We really thought hard about architecture. We took everything that we were doing and when you look back where we were it was ridiculous but we just took ourselves really seriously and worked really hard to make Drupal really credible and we laid out a roadmap of every question we thought they would ask us and so we basically said here's what you got to do for security here's what you got to do for scalability here's what you got to do for maintenance and management and hosting and we just rolled it out for them and then told them how to work with an open source community told them how they could come to Drupal Con how they could hire developers how they could build their team and shortly after that they agreed and I can't say a lot but then they did manage to deploy on Drupal and so yeah that's we transitioned from an opinion that we could do something to fact. And right around that same time is when we made what I'll call pivot number zero for Acquia which is we particularly talking to these two customers and some others when we finally figured out oh you know this box that we've drawn around the support that we're going to offer it's not going to work right customers were hearing something different so we made our first pivot in a big tent where we said okay now instead of defining that we're going to support Acquia Drupal in the top 100 modules we're going to support everything anything Drupal 6, anything Drupal 7 custom code, contrib, your theme, whatever so that was a really major shift we'd spent nine months with this model that we were going to it was like the equivalent of Red Hat Linux where they ship a certain set of software and support it we thought we'd do the same thing except it didn't work so enormous shift the entire we were involved to decide okay we're going to really expand the definition and then we started getting some yeses like Karen was just explaining so we started signing up more and more customers and we started supporting them and pretty soon we started getting phone calls hey, we're supporting us but our servers are offline hey, Apache is down well, we're here to support Drupal we're not here to support Apache we're not here to support MySQL we allocated 32 megs of RAM to your 200 module Drupal site but MySQL is not going to run, please go get a system administrator who will tune your MySQL and more and more of those customers basically either thought they could host it themselves and clearly couldn't or were like we don't want to host it ourselves and so we've discovered Pivot number two or Pivot number one Pivot number one Pivot number one if that was the other one was zero was a lot of our customers really needed hosting or wanted it and so we found ourselves having to go there and so this is a really interesting thing because when you look at the start-up genome project one of the things they say is there's a right amount of Pivots that you should have in your company if you have zero Pivots you probably weren't listening and you're going to fail if you have four Pivots you've probably changed your company's business so many times you're out of cash and you're going to fail so the optimal number of Pivots is to be around to two major Pivots in your start-up early on and so you should be anticipating that you're going to fundamentally have to change the products that you're developing as you're listening to your customers go there so change significantly one to two times and so the question is should we build this product and Mark Andreessen for those of you you know might not familiar with him he's the original developer founder of Netscape he went on to build a great company called Opsource which had a massive series of layoffs in the middle of the first recession and like miraculously 18 months later he turned around and sold it to Hewlett Packard for hundreds of millions of dollars I don't know what he did in those 18 months but he certainly affirmed himself as a true visionary and he's gone on to be founders in Facebook and LinkedIn and a whole series of other ones and so he says in a great market a market with lots of real potential customers the market pulls the product out of the start-up and that's what happened to us we were trying to support people their servers were failing and we just got sucked into that business to go there and so as you're working with your customers on your products you're going to find yourself getting sucked into businesses that you had no idea you were going to get into probably even thought you didn't want to get into so what are we trying to do from a customer validation we're basically trying to minimize the total risk of failure by checking your theories against reality okay and so we got to this point where it was like okay we're forced into the hosting business what are we going to do is Jason from Red Hat here no okay good I can tell you the story so Jason was from Red Hat and Jason's job was to basically print magazines glossies brochures sort of a graphic design background and they were looking for a CMS and said we need to go get into this web space a little bit more we want to put up different kinds of websites and I said well what's your team look like and he says we're effectively content people and graphic designers and they basically had no tech skills let's give it a shot would you like us to host your website for you and they said oh boy would we that would be great if you guys could just take care of that and make sure it all stays up that's great and at the time we were building our stack on Ubuntu and we said would it be alright if we ran the Red Hat website on Ubuntu no so we went out and you know we had no choice so we stood up a couple of Red Hat servers and Amazon Web Services and we were in the hosting business we were hosting part of Red Hat's sites and we were building a product and he said I'm building a product and it'll be ready in six months or something like that I think six months is what he wanted and I said great I'm going to start selling it today and so I was in San Francisco and I went and I found a great customer called Mother Jones that was having a big website and I actually have been working with them to get a case study printed on the front page of Drupal.org talking about how great their website was except they weren't really comfortable with me posting it on the front page of Drupal.org because it was kind of crashing a bunch so I really like to get Drupal working first before we start bragging about how great it is and I said okay well how about we help you and so they said great love to get some application support from you love it if you could host it want to talk a little bit about what happened then well so the first thing I want to say is that if you read the Lean Startup book they talk a lot about the minimum viable product and you know what can you do that's the fastest thing that you can ship to get any data any validation at all anything from the measurement phase it really kind of annoyed me back in what was it May or so of 2009 when I'm we're trying to build this hosting product and Karen went and sold Red Hat on custom configured servers but then we read this Lean Startup book a few months ago and they described I don't remember what the example they gave in the book but it was some company that was providing some kind of like shopping service and their minimum viable product was that like you would submit on the web page the products that you wanted bought for you and it was like you type them into a text field and the CEO would print the list drive to the store, buy the products and bring them to your house unbelievably inefficient right but it gave them the data that there were people who were willing to shop online which they didn't know before they had done that validation well as annoyed as I was when Karen sold this thing to Red Hat because it was completely not what we were building it did give immediate validation that if we provided a hosting service that there were customers for it so I apologize then we brought on Mother Jones so AQUI at the time had different levels of support that we would sell we had our enterprise level which was the all you can eat 24 by 7 no matter what happens what you break doesn't matter we'll wake everyone up to solve the problem and then we had lower level tiers than that that were obviously a lot less expensive we had a professional level tier that had like 6 tickets you know limited number of Drupal support tickets you could submit and so we thought ok well we have this hosting platform we'd like to sell it to as many people as we can so we'll sell it to anyone who has any of our subscriptions you can have an enterprise with support or you can have a professional with support or maybe you can even just have the I'm sorry with hosting not with support you can have enterprise with hosting professional with hosting or maybe just hosting just by itself you know and you don't call us for Drupal support maybe you have your own developer team you don't need our tech support and in fact mother Jones was our first hosting customer and they started with just hosting and then a few weeks after they launched their site crashed now the product our hosting product was brand new support team didn't really know it that well everyone including us including me we all sort of figured you know it's probably our fault I mean certainly we had found lots of things where we put bugs in the system and it wasn't very reliable but the site was down and so all hands on and we went and debugged and debugged and debugged and eventually we discovered that they had made a small change in Drupal core which it turned out was a bit of a mistake and that was taking their site down so what we learned from that this was sort of the second major pivot we had this hosting product we were selling it we thought we could sell it just to any of our customers but when we realized okay if we're selling guaranteed your site's gonna stay up and your site goes down by the time we're done debugging the problem maybe it's our infrastructure maybe it's your code maybe it's your custom module we have provided enterprise support for that customer we have you know dug all the way through the details of their custom code we realized we can't sell this without enterprise support because if we do we're gonna provide it anyway and we can't afford that so second major pivot I mean within two months of launching the product was we just stopped selling it for anyone who's a customer you could no longer buy the professional level of service with a limited number of tickets and be on our hosting platform and this will play substantially into our conversation later so again we're talking about the four steps of the epiphany we're working through our phases of customer development and we're still on customer validation and so our goal is to you know find the minimum set of features that we can require to get the early customers the second thing is we want to test the business model risk and certainly you know when you build a product and you launch it and you put service levels around it some of your customers are profitable some of your customers you break even in and some of your customers you lose money on right but the idea is you don't want to lose money on all your customers and with the scenarios that we were running to where we were putting 10 hours of debugging 18 hours of debugging of top tier engineers we were losing money not on everybody but on some of these things that we were putting ourselves business model risk and so as we tiptoed into this hosting business and it signed all these contracts to deliver hosting for a year we started having some real tough questions internally and you get to this point when you've got a product called Persevere or Pivot right and people in our company were like we need out of this business it is so scary to have everybody in the company freaking out at 3 a.m. in the morning trying to keep these sites up it's so painful that we can't do all the other things that we want to be able to do maybe we should just get out of here and so you have to make a decision and this is the great failure of so many startups is unbounded optimism is that we'll just try a little bit harder we'll just put a little bit more of the staff payroll on the credit card we'll just stick it out for a couple more months and keep burning and maybe we'll succeed and you can't do that you'll die you'll just plummet right off the cliff into the ground you've got to be prepared to pivot you've got to be prepared to make a tough decision and in this particular case I believe passionately in this I think Barry believed passionately in it and at the end of the day we chose to Persevere but don't take that decision lightly okay well and we persevered with the Pivot of we're only selling this with enterprise support basically we were like we have to be paid in order to do all this manual work that we know we're going to do at that point so the next stage the third stage of customer development is customer creation and what we quickly discovered was that trying to guess how much hardware you need to run a Drupal site that's completely custom that may have 200 contributed modules that may have 100% authenticated traffic may have 100% anonymous traffic may have 30% authenticated traffic had a huge impact on whether or not that site was going to stay up on that particular amount of hardware so we added an engineer to our team a guy named Chris Yates to our pre-sales team and he took over and as part of that product he had to go through and size every single site before it went live to make sure that there was enough hardware and the customer was willing to pay to keep that site up and that was pretty tough so we see here we've got a baby customer getting sized for a suit to make sure it fits now what we found early on was there were definitely companies that were very interested in our product and on one end there were startups and the startups were new they were trying to move fast and they were like if you guys can take care of this we'll go focus on all that other cool startup we're doing just run our Drupal site for us the problem we found was that startups were really conservative with their cash and they really didn't want to pay for service levels and they also kind of sometimes were a little bit sneaky about telling us how big their sites were because some of these were like social media companies and so they'd be like oh yeah Super Bowl were like the most linked to site on the internet oh thanks for letting us know 18 hours in advance that you're going to need 10 times the amount of hardware and so they would crash or when we said our site was Drupal what we really meant was it's mostly Moodle and like a million lines of custom PHP code and I think there's like some pages from it from Drupal oh awesome thanks for telling us it's not running well and so these are the kinds of problems we started with startups and then on the other hand we found some good places where enterprises they were like you mean it costs less than a million dollars because that's what the IBM guys charge us and we're like did we say 10,000 to host your site we meant 100,000 and they were like this is awesome we're going to move everything over and so sometimes in sizing we also had a great experience I got a call from Disney they were using it for Drupal in a big way they were using it for ABC Family they were using it for a bunch of their properties they were using it for the TV show The View and we were really thinking awesome Disney's going to take off on Drupal and we had a champion over at Disney and there were a lot of digital agencies that were delivering Drupal sites for Disney and they basically called us and said we want to see if you can size Disney for Drupal and I was like amazing let's do it so it was me I get on a phone and they were like we do 30 billion pages a month what can you do I'm like well it's pretty scalable yeah and so we're trying to come up with use cases we did a 20 million page view a month site once and stuff like that and I'm like does Drupal have a CDN built in and I'm like no but there's a module but it doesn't have a CDN built in but it's not built in therefore it's not scalable thank you it's not scalable hmm and so that's the way the conversation went and I went to my CEO and I said give me engineers we're going to build a billion page view a month's infrastructure and prove to Disney that we can do this and he said no and I got this really career level embarrassing email back from Disney about what a massive failure we had been in convincing them that we could help them deliver like a cloud of shame for about two and a half years and then in October I got a phone call hi it's Disney we've been failing in our CMS deployments for the last two and a half years we've made a corporate decision to move it all to Disney to Drupal woohoo and so we're back in the game and so Disney is now hiring and maybe some of you are in the audience some of the Disney folks are here in the audience but they're hiring a lot of Drupal developers and they're trying to create a platform and deploy a lot of Drupal sites so trying to fit for customers and trying to create those customers and get that pipeline is a really important part of the experience the fourth stage of the Marmar startup scale is to focus on scaling and really what you're now doing is you're now at the fourth stage of customer development you're in the company building stage and really what your goal here is to validate your founders beliefs but we could create a company that would help enterprises and large organizations succeed with Drupal no matter what and so we just put our heads down and tried to figure out how to do that and that meant tweaking the products and adjusting the products and making it work but also meant we had to scale the business now I was one guy and I went out and I met Sony and I went out and I sold the White House but that wasn't going to work I mean I can only sell so much so then we started building teams we started hiring more and more sales people and we had to scale them up and then we had to do more and more marketing activities and effectively what you want to do is you want to build a sales funnel and how many people have been the lead technologist and had to go into the field and sell products how many people have failed to convince their sales people to sell as well as they do you see the problem so you got to figure out a way to scale and get that funnel of sales coming down through the pipeline now in that startup genome report they came up with some analysis that talked about the theory of market types that as a startup there were a series and the startup that we were was called the challenger market type we're the short guy Vignette's the big guy we're trying to go up and take over the enterprise CMS market and we're going to get our butt kicked but that's the business model it's an enterprise sales model enterprise sales field force and that's the way you go to market it's pretty well for a while but then we found ourselves in a situation like Wally where we were persevering and persevering and stacking that garbage higher and putting in nice little neat boxes for what felt like a thousand years and then we kind of found out that we needed to get into another business so Barry's going to talk a little bit about where we needed to target so when we last left our pivot story it was October of 2009 and we decided we were going to sell hosting only with enterprise support subscriptions because after all if we debug your site we're providing that level of support and that let us down some very dangerous roads we made at the same time we had a lot of people in our company who were doing another project called Drupal Gardens including most of our visual designers and so whenever we came upon a question of okay we're adding this new feature it's a really cool feature what are we going to do for the UI how are customers going to use it we have design resources at the moment and it doesn't really matter that much because every one of our customers has an enterprise support contract they're allowed to submit an unlimited number of tickets and so we'll put in the best user interface that we the engineers and I suck at designing user interfaces the best interface that we can figure out and if anyone doesn't understand they'll submit a support ticket it'll be explained to them and then they'll get it great big mistake we were completely failing to pay attention to our customers one our support team hated the product because guess what it was confusing for them too and they were always having to come to us asking questions it made them feel like it just wasn't a very good product our customers as we predicted were having to submit support tickets in order to get basic thing done now how many people in the room like submitting a support ticket in order to do something that's necessary for their job yeah no hands no one likes asking for support we wanted to be easy to use so we led ourselves by making this necessary first pivot led us to this very dangerous position where the user experience of our product just wasn't very good and the other thing we realized around this time was we're selling to enterprises but the people who are building the websites at these enterprises they're not the CIO they're not the VP of marketing they're developers and the developers who work are just like the developers that work anywhere else they're programmers they like a good developer experience just like anybody else and we had simply completely failed to build the product for them so this was now towards the end of 2010 and we made another major pivot so we had this product we were selling it only with enterprise subscriptions with I don't remember what the prices were at the time but five digit or higher prices frankly was a good thing because they would have hated it as you saw that slide at the very beginning of the presentation where our own engineering and PS teams didn't like it because it was too hard to use so we started in January when we finally got the religion we got this message that the experience was no good and we crushed ourselves for about two and a half months and completely redesigned the user experience added a whole bunch of automation for things that previously our support team had done for the customers and then we launched the new version of the product called DevCloud a year ago actually in Chicago which was a huge improvement and we got a lot of positive feedback and we lowered the price a lot so now instead of a five digit price it was a three digit price and then over the summer we're selling the product and things are going pretty well but we start getting more feedback so we had started building this product in 2009 when SVN the version control system that people were using but in the meantime the whole Drupal community had shifted to Git and we thought Git, SVN they're not really that different people can use SVN didn't matter if we were right the customers had decided and the customers were speaking and they kept telling us why can't I use Git I want to integrate with GitHub I want to integrate with D.o so after vainly trying to argue that we were making the right choice the product was pulled out of us the customers forced us to add this particular feature there were a few other things that we did where when we started when we launched DevCloud we were still kind of an enterprise company so we thought we're just going to sell for a one year contract because with our enterprise customers we'd always sold one year contracts well it turns out developers don't buy hosting in one year increments they buy it in monthly increments and we had a lot more of 2011 this was only last summer we had there were these major objections that we finally started implementing and around the same time we got design resources and usability resources coming off the gardens project and we began very frequent usability tests where we put the product in front of the customer have them try it have them tell us all the things that were broken and then the usability people come back to us with the top six critical issues and the top six major issues and we started knocking those down one at a time so that by the end of the year by the end of 2011 or so we started seeing our usability numbers here's an example our usability people would hand back things to us with little arrows saying this is terrible no one understands and so by listening to customers and iterating and doing these usability tests every single time we released the product we the story really started changing and it was interesting because we had this experience that was not really that great for a long time and everyone hated it and we got into a position where we thought it might not even matter if we make the experience better maybe people are just going to hate us no matter what I don't know we're the big guys we're awkward but it turns out we listened to the customers we started doing the things that we said the usability score started going up and people started liking it more so the rapid iteration of build a feature measure the results with usability testing feed that back in we've learned something and iterate really started paying off and so this brings us to the fifth stage of the Marmer stages of profit maximization we've built the fundamental product now let's just crank the wheel and sell a lot more of it and by making it easier to use the product goes out and profit maximization starts to happen so we get to the last stage of the six stages called renewal and what we found was that some customers enjoyed their experience but often they would use a ton of support when they first launched their site and went live and spend the entire budget that we may have had for that customer and then as they got towards the end of the year they would say, you know, things seem to be working and whenever there's a problem you guys kind of fix it and I think we're ready to go on our own and so they stopped subscribing not initially huge numbers but there were certainly some who were doing that and so we said, you know, we got that feedback earlier about how good our heartbeat module was and it was kind of like, it was alright but then we started saying well let's add a bunch more things and so we started building out a knowledge base because we had hired all these people to do support and they had all these great things so they had 500 articles that were in there and then we started adding these tools configuration and mobify to do mobile and the Lullabots Drupalize Me videos for training and SEO grader and automated testing and people started saying, you know what, these tools are really great I want to keep running them on my website I'm going to renew because they're awesome they're really helping a lot and then last but not least was that classic support that we started off with in the first phase and so our renewal rates went up and investors gained more confidence and they were able to keep the business going and hire more people and so that worked really well so this brings us back to those learning phase and brings us towards what we started off in the beginning which is can we build a sustainable business right and so many startups start and they build absolutely amazing products and they take them to market and they do a ton of marketing and they hire a bunch of expensive sales people and they do it and it's very successful and then they run out of cash because the cost to acquire customers for the marketing campaign and the sales people and the initial product was so expensive that the lifetime value of that customer whether that customer is going to be there for a year or two years or three years or five years whatever it is just doesn't balance out and so startups run out of money even though they've got a killer team even though they have a great market that they can be selling into and so getting that right balance of the cost to acquire a customer and the lifetime value is something that you have to be thinking about as you get ready to listen to your customers and build products for them what I want to leave you with and what I want you to think about is that as you take Drupal and you use it to help your customers to build products to solve problems that they have it's not just about Drupal in the cloud it's not just about Drupal in some other technology stack you've got to get an engineering process in place that's very fast very iterative and you really start listening to your customers and you listen to them and you measure that listening and you get that feedback back and at the same time you're also thinking about how am I building my customer base how am I doing discovery how am I doing validation how am I creating my next generation of customers and how am I building my company we've got a lot of amazing and incredible customers and we really appreciate how much effort and time and patience they have given us not only the funding that they've given us but they've stuck it out with us and said we like you guys we really want to see you succeed and they have taught us a ton despite ourselves we finally started listening and I mean I think we were trying to listen but maybe not as listening as effectively as we should have been all along and I hope you take away from this that you can create the processes that allow you to learn from your customers and launch a million great products it's been a really exciting time I know Aquia has an interesting position in the Drupal community with Drees being the founder but I am really happy to tell you that this year we saw commerce guys raise $5 million we saw Pantheon raise $5 million we saw Datysphere raise $8 million we've seen Webform launched a couple of weeks ago we're seeing more and more Drupal companies succeed get this idea get that quick feedback they're taking Drupal they're building products around Drupal and they're taking it out to the marketplace they're proving that investors are seeing what they're doing and they're saying we love what you're doing your customers love what you're doing we want you guys to do it forever here's a check so there's promise I hope when we come back next year it won't be four customers four companies that got funding I hope it'll be 40 and I think that you know if you read these books I think that can really happen so thank you very much for your time be happy to take some questions for you you got a question if you could come up to the mic because it's being recorded is that yawning or quiet okay great so when you were explaining things it was funny that I was comparing this to design in our interface design process that we go through where people don't know what the heck they want a lot of times and so you just throw something out there you start morphing it into something that they actually can use that they actually wanted but they didn't realize what they wanted until you actually put something out there but I did want to ask like with your pricing like when someone's like oh I want to do hosting or something like that what was your experience with the pricing did you end up always underselling or were you overpriced or how did that go and how did you learn how to key in on that it's been a really interesting experience the first time we tried to sell any hosting I went to my head of professional service that I said what's it going to cost to set up a website in the cloud and he said the set up fee will be $10,000 and I was like oh god you've got to be kidding me $10,000 I can go to GoDaddy for $799 why $10,000 but then I started to really look at the market and see what people were doing to set up fairly sophisticated clusters of servers get all the tooling around there and I realized oh wait what does a Drupal developer make these days does a Drupal developer make $35 an hour does a Drupal developer make $50 an hour does a Drupal developer make $125 an hour hey there are Drupal developers making $300 an hour oh yeah I guess if it did take 4 days or it took 3 weeks to set it up maybe that's not so bad and then I'd go out to customers and when I go out to customers today and I say hey who's doing your hosting and I'm like what are you spending and they're like for 4 boxes we're spending about a million and it takes 3 weeks for them to get back to us and I'm like you're comfortable with that and they're like well they've got a couple of security enterprise features and that's the only choice and we save money because we buy an aggregate and I'm like really so those are real experiences between that and I feel more comfortable today where $150 for 3 sites at DevCloud and then we go up to customers where it's multimillion dollars to host all their stuff but it's been painful answering that a slightly different way figuring out the pricing is really a build measure learn loop all of its own when we started and figuring out the pricing model too so when we started with hosting something I didn't talk about was when we first brought on our first customers we were on a shared to build a mesh, a grid where we would have all of our customers on one system and they could buy slices and so you could buy one slice or you could buy 20 slices based on how big your site was and we were pricing it on that and so as Kieran said, you know, a Drupal site you can have a million page views and it can take almost no hardware because it's all anonymous static cached or you can have 100,000 page views and you can require a ton of hardware because your views, queries are really complicated so we tried to figure out, you know, how do we sell this to customers and we decided to price it in terms of page views because customers understood page views and clearly more page views is more value so we just said, okay, a slice I think it was like a slice was half a million page views or something like that and then a slice at a price and you could buy however many slices you wanted and that just turned out to utterly not work because, you know, the page views didn't really correspond to what your requirements were and the customers didn't really understand it and then we had the support thing so we iterated on the pricing model lots of times. In fact, right now we're in the middle of a major iteration on pricing with DevCloud. When you buy a DevCloud subscription you can have three sites on that account or ten sites on that account and then you can buy another subscription for three more or for ten more at different levels and we've been doing that for a year and starting around last fall we were getting very clear messages from the customers who actually loved us the most and they said this is terrific but how come I have to break it up into sections of ten? What that means is when I need eleven I'm kind of losing money for a few months until I need sixteen and then I'm breaking even again and then the last four I'm profitable and then I need to buy another ten like what's this about? So we're shifting over the pricing model again to be per site which is just our customers beating us on the head saying this is not how we want to buy the product we want to buy it this other way it's our money so we do what they tell us If you found this our talk about not listening to customers and user experience Lisa and Darmesh are two people who are very active in the user experience around Drupal itself and so they do a lot of usability studies on Drupal and they also do some stuff on our products they're doing about a hundred different usability studies here this week so if you're interested in participating or you want to see what that is like because you've got a product and you want to see what it's like to go through that study I encourage you to go ahead and sign up for that I also had a colleague of mine who wanted to plug and say that we're doing another session I'm doing another session immediately after this one called 200 migrating and scaling 200 artists in the cloud and so it's in a different room but if you're interested in seeing that it's a really great rock and roll story about migrating clouds and then we've also got another one about a panel that's going to be out on the day stage about enterprise gardens Does anybody else have any questions? Well thanks for listening I hope you took away the lessons of how to create a great startup and great products Thank you