�ݺ�ߣ

Editor's Notes

• #2: Thanks, it’s great to be here.\n
• #3: When Andy Baio was first organizing this conference, he asked me to speak in the hopes of bringing some sheen of wisdom from an oldster. I tried to talk him out of it, since the conference is focused so much on cutting-edge things happening today, but he argued that someone like me who has had a long term project could offer some insights to anyone just starting out.\n\nSo that’s why I’m here.\n
• #4: In 1999, I started a community blog called MetaFilter. I went with a community model because I didn't think I could do it all myself and much of the early success of the site was not only being in the right place at the right time, but having the right tools available to visitors, and me taking it seriously soon after it launched. It currently has about 60,000 paid members.\n
• #5: The site grew and grew, and at the tail end of 2003 I added a subsite to it called Ask MetaFilter which was a Question and Answer site. By 2005, I finally got to quit my dayjob and work on the site full time. Shortly after that things got bigger and bigger and I hired my first employee the following year and now we're six people helping run the site.\n\nThis was a long process that resulted in me becoming a reluctant businessman. I had to learn what it is like to get a lawyer, investment advisor, accountant, etc. It helps to talk to old people who know this stuff.\n\nI learned a lot along the way and I hope to share some things today\n\n\n
• #6: I thought a lot about my history on MetaFilter and other sites I’ve run, and I came up with three big lessons from running long term projects that I think could be useful to anyone starting out today.\n
• #7: 1. The first lesson is that Failure is the Greatest Teacher.\n\nThe word failure has a lot of negative connotations, but I've always embraced it. \n\nFailure is great because on the most basic level, it can teach you what doesn't work and hopefully help you get closer to something that will. Think of trying to make your way through a maze, hitting dead-ends means you can go back and try a different turn. It’s not the end of the game.\n\nFailure is only a bad thing when you ignore the potential lessons it can give you. If you treat it as a dead-end and stop right then and there, only then have you really failed yourself.\n
• #8: Thankfully, we have this really awful word I hate hearing but I like the sentiment: Pivot. \n
• #9: People "pivot" their products and projects when something isn't working in an effort to figure out what would work. \n\nYou will never hear me use this word, but I like how it puts a nicer face on reacting to failure.\n\n
• #10: When you start to be ok with failure, you can do things that can best be summed up by the phrase "Throwing Shit At The Wall To See What Sticks". Pretty much every friend I have in this industry that doesn't work in a cubicle farm could have the job title "Throwing Shit at the Wall to see what sticks". \n\nIt's what creative people do, and should be doing as often as possible. Don't fear failure. Always be trying new things. \n\nCreate an environment for yourself that allows for this, whether that's setting aside time to noodle on projects (like Google’s 20% time) or build yourself staging servers to test home-grown code on, if you give yourself the conditions to make this work, you might be surprised by the things you come up with.\n
• #11: It's important to not be afraid of failure. \n\nWhenever I fail spectacularly, I always take notes. I love looking at old text files and drafts of blog posts and things I wrote in the aftermath of shuttering an idea or suffering some public embarrassment over a project. I try to take notes on my failures as soon as I can, when the lessons are fresh in my mind. I've currently got two big projects that are working out pretty well but I've probably got a dozen things in my past that didn't work out so well.\n
• #12: Let's go over a few:\n\nTicketstubs - The idea for this site was that one day I was moving out of an apartment and I noticed I had a small pile of ticketstubs I’d carried around for the past 10 years or so. Each one had a great story, so I thought I’d build a site around that.\n\nSo for this site you scanned in a photo of a ticket stub and told a story about it. Most of the failing of the site was people thought it was a gallery or a hall-of-fame and didn't understand why I didn't let them post all 58 Bruce Springsteen shows they'd seen, or they only wanted to tell stories of how great the performer was. I only wanted the story of the one show they took their wife to on their first date and why they saved the ticket. I learned from this project that simple ideas work a lot better than complicated concepts and if most of your first time visitors don't get the reason for the site immediately, you need to try harder or simplify the idea more.\n
• #13: PVRblog - This site had its heyday that is now clearly past. In 2003, I started a blog featuring all the tips I was finding for TiVos and I was blogging about news coming out about personal video and it was a total blast and I was obsessed a bit with the subject matter and by coincidence Google’s Adsense system launched a couple days after I launched the site, and suddenly it became a really lucrative thing. At its height, I was making only one post a day and the site was bringing in $4-5k dollars a month. It paid my mortgage and bills and made way more than MetaFilter in the early days.\n\nNow, I come from a world of academics. I was a grad student in the sciences and grew up on the college internet and my first web job was at UCLA. In the academic world, everyone shares anything they learn so we can all get collectively better. I did the same and wrote up an article extolling the virtues of Adsense and pointed out that I might have the first very profitable blog and that anyone could do it.\n\nThere was a huge response to my essay and Google was stoked on it, making my site one of their case studies, but a few months later I realized there were seven blogs all about tivo-related things, including one by Weblogs Inc., all covered in ads. \n\nAnyway, long story short, after a few years I wasn’t so into this subject any longer and let it languish and die a slow death. Part of me wonders if the lesson here is I should have kept my big mouth shut and just enjoyed the runaway success when I had it, but I’m conflicted about it. I’d rather live in a world where people share the things they learn.\n
• #14: Travelfilter - This is something we conceptualized, built, and started testing, but never formally launched. It was supposed to be a subsite from Ask MetaFilter, only about travel, and it would have extended features like geolocation of questions and a browseable directory of questions by place, so you could easily see every question about Hawaii or NYC in a single screen. \n\nWe never launched it because we realized it was going to take away a popular category of the existing site and we didn't want to cannibalize our own work. \n\nWe could have launched it as a new travel site separate from MetaFilter, but then it would have to start from scratch, without as much of the good benefits of being MetaFilter, like trustworthy users, and it was too big of a job to take on as a new venture. \n\nSo we didn’t burn the boats. You’ve all heard this story about Spain’s invasion of the Aztecs, right? Cortes didn’t just land, look around and go back to Spain, he told his crew when you arrive, you burn the boats.\n\nIf you want to launch something, you need to just do it, a half-assed solution isn’t going to work. We burned the boats on a meetup section of the site and forced it to a new subsite and it flourished.\n\nEventually we rolled some of the proposed features into Ask MetaFilter as a compromise.\n
• #15: I also worked as an employee on projects met with failure, usually because they were ahead of their time. \n\nBlogger comes to mind as one of those. It launched a couple months after MetaFilter did, and I worked on it for about a year after it started taking off. I could tell it was revolutionary and would change the web someday, but thanks to there not being any revenue coming in and the general slowdown in the post dot com bubble economy, the project ran out of funding, ran on fumes with just one co-founder but a few years later it did start getting revenue from pro accounts and eventually sold to Google. \n\nMy 20-20 hindsight Takeaway: start taking money from customers as soon as you can\n
• #16: \nEvan Williams was a cofounder and I think embodies a lot of what I'm saying here today. This isn’t a dig against him to say he failed several companies when he was growing up in Nebraska, pushed Blogger until it go to Google, left and launched Odeo which had promise but also sort of failed, then spun Twitter off from there. Ev’s story is pretty incredible, falling down so many times and getting right back up. He seems like a guy that isn't afraid of failure and the world's a better place because of it.\n
• #17: 2. Money is the least interesting problem to solve in any project.\n\nMore accurately: If you are working on your site/service and you only have a finite number of things to fix, It’s the least interesting thing to optimize for.\n\nMoney is a necessary evil and that's how I've always viewed it. Given the last 13 years I've spent on MetaFilter, money has been a constant low-level nagging issue, but aside from paying for healthcare, happy employees, and my mortgage, it's mostly a distraction.\n
• #18: At first, the site was done as a side-project, and it was years before any money came into it, and that was just to pay for upgraded servers. Eventually, the money became enough to pay not just for servers but also employees, and that was great but I tried not to let it dominate my thinking. \n
• #19: The times I've focused exclusively on money matters are some of the most stressful times where I made the worst decisions about the site. I figured out this was because I was aiming for more money first, not thinking of the site members first and foremost.\n\n\n\n
• #20: Jeff Bezos called this "the Amazon doctrine" last week at the Kindle launch and The Amazon Doctrine is aligning the company's incentives with the needs of its customers, not advertisers or content providers, so your customers win when you win.\n\nYou’re never forced to make a decision that is at odds with your customers. It’s really hard to do this, but I think it’s fantastic.\n
• #21: And while I’m mentioning smart famous people I should also mention Tim O’Reilly who is here somewhere today. He wrote an amazing essay in 2009 about working on Stuff That Matters, and I’m paraphrasing but when discussing money he said to basically imagine a startup like a cross-country road trip, and money was simply getting gas to keep you going on this trip. The point of any roadtrip isn’t a gas station tour, but the journey. Try not to focus more energy than necessary on money.\n
• #22: \nThe times I have tweaked and monitored and watched every dollar coming in and going out are the times that coincided with me not getting much sleep, being constantly stressed out, and having to have friends talk me down from the metaphorical ledge because I was ready to throw in the towel instead of freaking out about money.\n
• #23: \nThe best days I've ever had working on my projects were when I made a coding breakthrough or thought of a new feature and saw it implemented shortly after -- things that made my projects work better and do new things. None of the best days I can remember involved money.\n
• #24: 3. Success is fleeting, so stay on your toes\n\nI've had several of my own projects get popular before fading away and watched it happen plenty of times to know that success and the feeling of being on top are temporary. It’s a fact of life in this industry that new stuff will always replace old stuff but that’s what keeps it exciting.\n
• #25: If I had to plot a graph of a typical project's success it'd look like the classic photo of Mt. Fuji, a long approach up to a point, then back down. \n
• #26: Something like this.\n\nMost of my projects have followed this kind of path. Though I'd say MetaFilter is kind of a large mound, I’ve left a dozen projects in my wake that had their moment in the sun and then went away. I also feel like MetaFilter could very well be past its highest point of its “mound” and on the way down.\n\nThis story of growth and decay is told over and over again.\n
• #27: Look at web publishing as a whole. We started with pages and you had to have your own server, eventually moving to free hosts like Geocities, then Blogger and other apps came along and said forget writing whole pages, why not just write a few paragraphs and we had posts on blogs. \n
• #28: MySpace came along and said if blog posts are too much work, post shorter updates, and then Twitter said you have 140 characters so keep it brief, and finally we have Facebook and Pinterest saying you just click a single button to register your current feeling about something.\n\nHere is the trail of bodies being left behind by our constant appetite for innovation (progress)\n
• #29: I see this pattern in other industries as well. Look at the world of car rentals. Companies like Hertz dominated for decades by offering the simple (overpriced) service of renting a car by the day, \n
• #30: then zipcar comes along and says you can rent cars by the hour, shaking the industry up. Zipcar’s reign lasts several years and they got to be so popular that I saw Hertz introducing by-the-hour rental services trying to compete with Zipcar.\n
• #31: Of course anyone that lives in Portland knows there’s a new king in town. Car2Go goes beyond zipcar by saying you can rent a car for mere minutes, and you don’t have to worry about parking anywhere in the city.\n
• #32: and already on the horizon I see things like RelayRides, which are like a massively distributed zipcar service that let anyone rent their personal car to anyone else by the hour.\n
• #33: I'd argue that this pattern of something getting big, fading away and getting replaced seems to be going faster given so many tools available today. In the past you had to write everything from scratch but now there are mature programming languages, design pattern libraries, and lightweight scripting frameworks making it easier than ever to go from idea to alpha in a short time.\n\nThis kind of open knowledge sharing means everyone can build on each others work quicker, but also makes competition more fierce.\n
• #34: The real problem I'm facing today is how to keep MetaFilter relevant in 2012 and beyond, given I started putting it together 14 years ago. My greatest big-picture threat is irrelevance. \n
• #35: Will MetaFilter even exist in 2020? If so, what form will it take?\n\nIt never hurts to ask yourself this question about your own work.\n
• #36: In the short-term I've also noticed something as a more immediate threat. \n\nI don't fear Yahoo Answers or Quora, or any of the other Q&A sites that have popped up, I don’t think asking questions on the internet is a zero-sum game, where if someone asks a single question on Quora that means one less question for Ask MetaFilter. That’s not how it works.\n\n
• #37: The biggest source of change for MetaFilter these days is hardware.\n\n
• #38: Here is a graph of mobile clients accessing MetaFilter's servers over the past few years. In 2009, it was 2%, then the following year it is 12%, it shot up to 20% last year, and earlier this year I looked at our stats and saw it approaching 33% of all devices accessing our servers. This is significant because I have a giant content site people read and it has a few ads on it to pay the bills. \n\nWe’ve had a gradual, but significant reduction in revenue as mobile usage has gone way up. Traffic overall is steady, which is why I think this is the cause.\n
• #39: I'm not a guy that runs adblock and hates advertising, I myself click on relevant ads several times a week on my desktop computer I actually like targeted advertising that works, but I've only done it a handful of times that I can recall in years of using an iPhone. \n\nHere's what the NYT looks like on a mobile device, \n
• #40: but here's how I read articles there, zoomed in. A couple taps, and no more ads.\n\nAnd even if I did see an ad, the likelihood of me clicking it, getting redirected in a new window to it while I’m trying to read an article is very small. It’s much easier to do this stuff on a desktop. It’s clumsy on a small device and I end up ignoring all ads on it.\n
• #41: Strangely, I see this affecting almost every content site out there. This past spring I visited NYC and talked to people at all the following companies shown here and everyone was kind of freaking out over their stats showing mobile use approaching 30%.\n\nSome where seeing revenue go down, others were just wondering what they’ll do next given these new conditions.\n
• #42: I don't know if the future of the web is that we all have to release mobile apps specific to our service, but if you can do something better in an app than you can using the web you should probably consider it given a third of your audience is on those devices today. \n
• #43: \nI also think this is a tremendous opportunity for someone to come along and basically do Google's job of selling and placing ads better than they are, because even after five years of iPhones no one has come up with mobile advertising that works. The advertising research groups talk about the “mobile gap” and I think it’s a real opportunity, because display advertising on a tiny 4” screen is going the way of the dinosaur.\n\nI am sorry to the two owners of large ad networks that are in this room. Though if you could come up with a mobile solution to this small screen challenge, you’ll do very well in the future.\n
• #44: Mobile is a real revolution, the official Olympics website reported 60% use by mobile as people accessed schedules and results while on the go, or checked out the site on an iPad while on the couch watching it live. \n\nAnalysts have long predicted 50% mobile usage of most sites by 2015 but I think that's too conservative of an estimate. I think we’ll be there by 2014 at the latest, since we’re seeing a couple percent mobile usage growth month-to-month.\n
• #45: And of course, I think the next obvious step after mobile is ubiquitous computing, or UbiComp. An “internet of things”\n\nThis next wave that could supplant mobile is tiny devices everywhere.\n
• #46: Things that are part of our everyday lives that are connected to the web.\n
• #47: I don't know what this means for content sites but it is clearly where things are heading.\n
• #48: I run a big content site and provide a space for people to discuss things, if devices are going out and getting answers and boiling it down to one blink, what does that mean for my service? I mean, don’t get me wrong, I love Mike Kuniavsky.\n\nI have no idea where MetaFilter fits into the world of 2020 just yet, but it's a very exciting time to be around.\n
• #49: \nMy time here is short, and we've had an amazing conference so far, so I'll just leave you here with this. \n
• #50: Embrace failure and learn from its lessons, refactor whatever you are doing to incorporate failures as feedback. \n
• #51: Don't waste too much time freaking out over money because there are always better ways to spend your energy,\n
• #52: and finally, the future is a crazy uncertain place, and while it's fun ramping up to the top, beware of the descent on the other side. Stay on your toes. Always be trying new things.\n
• #53: \nThe future is indeed bright and exciting as hell.\n\nI'm glad that I get to share this crazy bright future with the all the great minds gathered here today. \n
• #54: \nThank you very much.\n