 So while they're fixing the tech, how many people here have ever failed? Okay. And in terms of professionally in civic tech, would you say something you worked on just didn't work and you had to shut it down? Okay, great. Well, that's great. Glad we're all in. Hi, it's like a big AA meeting. Hi, my name is... It wasn't my problem. Oh, okay. So I just want to say, first of all, I'm glad to see that this is meant to be a learning space. Clearly we need to learn and can learn. In fact, perhaps the only way we learn is by making mistakes. As Thomas Edison said, you know, he didn't fail. He just found 10,000 ways that didn't work. And okay, slide. Let's see. It's not going. I'm pressing the wrong button. Let me use this. Did that work? Okay. Now I've made the screen go off. I've already failed. The entire screen has gone blank. Yes. Tell me what button you want me to press. And that should just be scrollable. Yes, it is good. Okay, scrollable. What did he do? Oh, come back, sir. You're going to have to do this for me. Which button? I'm just going off and down on this. When he presses it, it works. Okay. And as Esther Dyson likes to say, you always have to make new mistakes. So I'm literally going to scroll through these really fast. Also, I just want to say that the civic tech field is full of successes. We're not going to spend any time talking about them in this session. But I want to acknowledge that. That this is not about dragging anybody or obsessing with the failures. It's just to be honest, we've been around long enough. Let's try and learn from some patterns and some clusters of types of failure. I literally have to scroll down through everything to get this to Gs. Okay. So failure is not unusual. In fact, in the whole startup field, something like 75% of VC-funded startups in the United States fell. These are things getting more than a million dollars in early-stage funding, according to a Harvard study. And the most common reasons gleaned from lots of failure postmortems will sound pretty familiar to you. And this is just for tech startups, not just for civic tech. The biggest reason, not finding a market need, running out of money, not having the right talent, getting out competed, not getting the pricing correct, or building a product that people don't really want to use. So to get into the specific patterns of civic tech failure that we found from looking at the civic tech field guide, so we've got about close to 2,000 instances in the field guide of which about half are actual products or entities or platforms or tools. And what we zeroed in on out of those thousand were things that we knew were dead and that also had some early-stage support. So this is not a discussion about projects that people started on a weekend and then abandoned. Okay. Those don't count. We live in a space where it's so easy for people to start things and try and see if you get any traction that it would be a mistake to sort of focus just on those failures. We wanted to look at things that actually had some capital behind them, teams, a presence, users, and yet they've still failed. The first, obviously, the biggest reason over and over again is just the failure to design really with users in mind. If you build it, they don't come most of the time. And this is why the Lean Startup methodology is so useful to do constant user research and to be paying attention to what users actually want. A second reason which is often isn't talked about enough, and we saw this especially in cases of civic tech organizations like Sunlight, for example, where I was an advisor for 10 years, or My Society, where you have multiple projects and you actually are trying lots of things, but at a certain point you decide to prune and you realize you cannot support everything. You should put your resources on the things that are either proving to have the most value or where you actually also think uniquely you play a role in the ecosystem. There's a good example of a project that My Society shut down called Pledge Bank, which was actually a thriving project helping people aggregate collective action. But by the point at which they shut it down to their own explanation, it's partly because there were other places where you could do that as well. There was no special reason that My Society needed to keep this thing going. And then the last reason that we often see things shut down, and again this isn't talked about enough, is that founders often get acquired. Teams get acquired. A good example of this was when Change.org bought Elect Next, which was an opinion matching site for people trying to decide who to vote for. Change.org didn't continue that product. What they wanted was Kia Danambam, who was a great product manager, as part of their team. And this is a perfectly reasonable ending, I think, especially when you consider how hard it is to do all the things that founders do, and that sometimes it's better to go inside a larger company where someone else is worrying about problems like meeting payroll and you're just getting to do the thing that you love to do best. So we often see that happen. So now to sort of drill down a little bit further. Out of the 44 entities, and I should say you can look on the Civic Tech field guide yourself. It's bit.ly slash organized Civic Tech. We've opened a new tab that we call the Graveyard, and that's a provisional name. If people have a better one, let me know. And again, this is a first pass. This is mostly American focused data, and one of the things that Matt and I are working on, and in fact inviting people to a meeting Friday morning, is to help us figure out how to get more contributions into the guide for more international settings than primarily just the Anglo-speaking internet. But of the 44 that where we were able to do some deeper research into what these sites were and why as best we can tell they failed, there were some clear clusters. It's great. I have to really scroll down. So the first one is efforts to solve collective action problems, and there's a recurring pattern here. Sites that try and get people, and you think in the abstract, this is a really good idea. Let's make it easy for people to get together to take some common action. The biggest failure in this sector is a site called The Point. How many people here know what The Point was? One, good. How many of you know what Groupon is? So Groupon is what grew out of The Point. When Andrew Mason started The Point, he actually raised $7.5 million in early venture funding for this. And it called itself a tipping point-based collective action website dedicated to getting people together to accomplish a goal. The problem they had is that the barrier to action was really high. So for example, they were trying to get people to pledge to be part of a campaign to get Aquafina to switch to biodegradable bottles. And they needed 50,000 people to sign up for that before that action would be triggered. They wanted to get universal music to sell DRM-free songs. They needed a million participants for that action to be... So clearly they set the trigger level way too high. And they were also asking people to do something that didn't sort of flow easily out of the normal course of what humans do. And actually Andrew Mason has said that the idea for Groupon came to him when he noticed that users on The Point, there was a group who were actually building a campaign together to save money by buying things together. And it turns out that personally saving money is a pretty strong motivation. Self-interest works really well there. And that's when he pivoted to Groupon and the rest is history. But we've seen the pattern again and again. And I think there's a lesson here for why these kinds of efforts fail. The first one is that the collective action dilemma is real. It really is not a common sense to think that your small action will make a difference. And even when you try and show people that if we all do it together, it will make a difference, by then they've already gone to another website. So asking people to do something really out of their ordinary is a pretty high barrier to action. There are other reasons we think these kinds of sites fail. One is that we think that there are easier entry points for collective action. There's a reason why signing a petition gathers lots of people. It is not a high barrier to entry. And frequently the most interesting thing for me is to watch people enter at a low level and then move to a higher level. So whether it's on petition sites like change.org, which is a very successful example, or just people doing it on their own through hashtag campaigns followed by, you know, put money in my GoFundMe account. You see people solving the collective action problem just using simple consumer available level tools rather than saying, oh, I got to go to that website that somebody built special for this. In the interest of time I'm going to move on. Second set of clusters, and this is the one that I see most often people come and pitch us and I'm like, please don't do this. But yet again, it comes up again and again. People trying to build the one-stop shopping solution for all your political needs. Right? And there's a long history of these. And, you know, some of them with quite, you know, notable levels of usage. For example, voter.com, what a great URL, right? In 2000, they got 17 million unique visits around election day, and yet within a year they closed down. VoteIQ, which launched about 10 years later, I remember when they came to PDF to launch, they were so excited and I felt so bad for them. Because even though they had $2 million in venture funding, all my instincts said that this too was going to fail and sure enough it did. And I think there's a subtle lesson inside of that story, which is just because you can raise $2 million from a VC doesn't mean that either you or the VC is smart. There's a lot of very gullible money sloshing around out there and lots of people who actually don't do this research and get excited. I just noticed the other day that Ballot Ready, they just raised $1.5 million. Again, another site to give you all the information you could ever want about everybody who's going to be on Ballot It. How many people here think that's going to fail too? Raise your hand. Come on, you've got to show some more courage. I mean, clearly that, you know, if you can get that same information now from Facebook, why would you go to another website to do it? So lessons in the failures of these kinds of products. The first one is that politics is actually complicated and most people don't really want to know the details. They just give me the headlines, right? So this assumption that there's a big demand for a lot of detailed information is a false assumption, certainly in the American context. The second thing is the idea that a generic nonpartisan hub for this information is the way to provide it cuts up against the grain of the fact that most of us who are interested in politics do so for partisan reasons. We've got a team that we root for or a tribe that we believe in and so the generic nonpartisan hub, however good and civic that may sound, just doesn't seem to click with your intense political users online. Then there's the idea that, well, maybe we can make this information valuable to elected representatives. Actually really interesting to hear about Facebook's experiment in this genre with the sort of constituent badging. But so far we have not seen certainly monetizable interest from representatives who want to pay for this information. They will pay to advertise to those people, I guess. So think about that. The incumbent hubs that provide this information are already doing a good enough job. That's yet another reason why you should have second thoughts. The fact that, for example, Open Congress, which we supported at Sunlight for a long time, which was better at SEO and had more group forming tools and had a bunch of other nifty features and yet it never displaced GovTrack.us. And in fact now if you go to OpenCongress.org it redirects you to GovTrack. And GovTrack is built and run essentially by one person who's been doing it for about 13 years. God bless them, you know, and let's face it. Sometimes it's like the incumbent wins because they're good enough. Last example I think I have time for, maybe we'll get to the fourth one. Why don't opinion matching sites click with people again? These are sites, by the way, that do things like help me decide who to vote for, right? So again, to the previous point, politics is complicated. Not that many people are that in need of the answer to the question, who should I vote for? They get the answer through other means as opposed to the rational ones of, let me do the opinion matching thing. And to paraphrase the way people inside Google are said to think about this, it's not a toothbrush problem. It's not something you need to do twice a day. And so the fact that you do want to know the answer to the question, who should I vote for, but it only happens for about two weeks out of every four years, means that that's not enough of a consistent user base for you to sustain your project. The last failure in this genre was voters in, which had a really interesting idea that we should make our voting preferences social. So, and for some of us this is actually a useful thing. I used to joke that I controlled at least 10 votes in New York based on the number of people who would ask me before an election who they should vote for. So why don't I make that sort of information more socially shareable? They started in San Francisco a place where people love to share everything and yet they could not get enough traction to keep voters in going. And they had certainly plenty of money as well. Okay, I do have time for this last example. Hyperlocal, why don't hyperlocal sites that are trying to, you know, basically take everything from local blogs and local open data? About 10 years ago we had two very prominent well-funded startups in the space outside in, started by Stephen Johnson, who raised I think about $14 million in all for this project before he sold it to AOL. Every block funded with a night news challenge grant of about $1.1 million, ultimately sold to MSNBC. Why does hyperlocal keep failing in the civic space? And I think maybe we should be humble and say that no one can figure it out. AOL with patch hasn't figured it out. So, you know, places with lots of capital that local news and information of some quality just doesn't have a market answer, that's one possible answer. There are others. It may be that every block and outside in were too soon. Both of them launched pre, you know, in the era where people in the United States were just really starting to use smartphones. Not everybody had a cheap data plan. So it might be that the conditions for this market are better. I would say the fact that both Google and Facebook have now recently begun to prioritize local news more in newsfeed and in search results may actually improve the chances for people doing hyperlocal news. We don't know. But we should also recognize the external conditions at work that both Twitter and Facebook have provided so much local information to people already that they may in effect be soaking up all the oxygen available to support these kinds of efforts. So external conditions matter too. So this type of inquiry is one of the reasons why Matt and I are working on upgrading the Civic Tech Field Guide. We know that the information that, you know, we've just given you here is just preliminary. I'm sure you probably could add a lot more examples of failed efforts. Please put them into the guide for us. It's open source. And, you know, we benefit from the contributions of everyone. The URL again is bitly.com slash organized civic tech. And I should just say for both of the talks that we've done this one and the previous one, we're going to be turning these into articles on Civicist and posting our slides. So looking forward to engaging with you further there. Thank you.