FOWA NOTES ---------- Following is a big, 4000-word summary of the proceedings at FOWA London 2009. If you think this is hard to take in, it was even harder for me, I was there trying to make these notes :P ------------------------ Taking your site from one to one million users - Kevin Rose (digg, etc) 1. Ego Does this new feature stroke the users' egos? Leaderboards and rewards (like badges, ranks, etc) drive usage Twitter's 'followers' gives you social cachet 2. Simplicity Concentrate on 2-3 features for the whole site, and do them really well Ask yourself - is there anything I can remove from this feature? Keep checking site usage and remove dead wood. (like our jobs feed :) 3. Build and release Learn from users' activity. Don't try to predict what they want next. Iterate... 4. Hack the press Invite only systems generate a buzz - send invites to bloggers to get free publicity Talk to junior bloggers, it's much easier to get their attention Attend parties for events you can't afford Network, but bring a demo! 5. Connect with your community Start a podcast Throw a launch party, then yearly/quarterly events, but don't tell the bar! (they will charge you if you tell them you are having a party) Be active in your own community 6. Advisors What problems are you going to face? If you're a developer, marketing could be something you need advice on. Give them stock compensation instead of a place on the board of your company ...I missed a couple of numbers here Farmville makes people feel indebted to their friends, so they feel obligated to play more 9. Analyse traffic Entrance sources Paths through the site Exit points 10. The entire picture Remember to step back from time to time to look at the big picture and see how things are working. ------------------------ Three vital marketing systems for a successful web app - Mike McDerment (freshbooks) Marketing 101 'The funnel' 100 > 10 > 1 (100 visitors, 10 sign up for trial, 1 becomes paying customer) Tracking, Storage, Reporting are three important things you should worry about Google Analytics does tracking Store marketing info (eg where they came from, when they signed up, etc) with the user account details in database (it's apparently that important) Reporting - make it easy for marketing people to get data, so you don't have to keep writing queries. Write a data search app. Question - what makes visitors sign up? How can you get more people to sign up? Answer - you can get more to sign up but they are less likely to become paying customers, so it's not really worth it(!) ------------------------ The future of javascript design patterns - unleashing full object-oriented capability - Dustin Diaz (twitter) JavaScript is like corn. Frameworks are different ways of eating corn (he listed them. YUI is creamed corn because a lot of people don't like it, jQuery is like corn syrup because it's in everything and make people want more) Frameworks "frame" the way we "work". So it's important to choose the right one. jQuery is like cocaine - one line and you're hooked. Or we can treat frameworks as tools and use them to accomplish discrete tasks, eg Date JS No frameworks! Compose small tasks Keep business logic and look-and-feel separate (modals and views style) Use classes Good examples: Dean Edward's Base http://code.google.com/p/base2/ Dan Webb's Low Pro (based on jQuery) http://www.danwebb.net/2006/9/3/low-pro-unobtrusive-scripting-for-prototype (actually it seems to be based on prototype) Don't bloat your code with error- and type-checking. ------------------------ How we built HelloApp - Scott Guthrie (microsoft), Matt Lee (redgate), Keir Whitaker, Ryan Carson (both carsonified) Boring massive plug for ASP.net. Yawn. ------------------------ Passion and paychecks: Open source lessons - Addison Berry (lullabot) There is money to be made in open source! If you don't make your code public, people will just write their own anyway She has a really great job, but I'm not learning much... ------------------------ Introducing Atlas: A visual development tool for creating web applications - Francisco Tolmasky (280 north) The web can do the job of most desktop apps now, the technology is there, but the tools to develop the apps are missing Cappuccino aims to do that (framework for developing web apps, cocoa on the web) http://cappuccino.org/ Example: 280 slides (looks pretty cool) http://280slides.com/ So they tried to make a web app that develops web apps, but people didn't like it, they didn't trust the company with their code So they made it into a desktop app - Atlas (280atlas.com became 280atlas.app) Demos making an RSS reader app in about 5 minutes, impressive but all the features he uses are built in already... It also makes desktop apps though, you can make the app once, and save it as a web app, or a (mac) desktop app Public beta Nov 15th and will cost $20. It looks pretty amazing. ------------------------ Get niche, get rich and go mainstream (this works better with the american pronunciation of niche, which rhymes with rich) - David Prager (revision3) Lets look at the definitions of these words. A niche is a subset of a larger community Rich means the quality and depth of your content Case study - city sourced Capitalise on community Shows a diagram comparing vodafone with revision3. Vodafone pay a lot to sponsor man utd. Big reach, but niche audience (vodafone customers) = low return vs Revision3's app judgment tv show is made for a niche audience, but has a broad reach = larger return for lower cost Niche audiences are more likely to evangelise your product Build out from niche to other niches, then to mainstream, eg: facebook - havard => other colleges => everyone ------------------------ Payments innovation will unlock the web's potential - Osama Bedier (paypal) Assertion: legacy infrastructure hurts. Illustrated by a nice history of railways that basically boils down to: the range of the space shuttle is determined by the width of two horses' arses. Next analogy: electricity's killer app for first 30 years was lightbulbs. Then toasters came along, but you had to screw them into lightbulb sockets on the ceiling! Then power sockets were invented and paved the way for innovation and an explosion of uses and users of electricity (nice exponential curve) The internet is at the pre-television stage on this scale (at the bottom end of the curve basically) "Cash is dead, good riddance" Cashless payments enabled more commerce Ebay use went up enormously when paypal was introduced (as people didn't have to wait for cheques to arrive in the post, clear, etc, before sending items) They are opening up paypal. https://www.paypal-communications.com/innovate2009/ They've also got a weird and very amateurish-looking blog at https://www.x.com/blog/ (yes, X.com) Amusing promotional video to end with, strongly implies that paypal promotes sexual promiscuity http://www.youtube.com/watch?v=rgT7gGciQrg ------------------------ Advanced web app marketing strategies - Chris Abad (spymaster) Spymaster is a twitter-based game that got a lot of attention when it first came out You can introduce some of the game elements into other apps How? Rule 1: build a great product Build passionate community of raving fans Go where the community is instead of making your own Incentivise users to market for you Give rewards for sending out twitter messages (eg: more money in the game for each public message they tweet) Incentivise users to recruit their friends Eg: power-ups depending on how many you invite Find key viral channels to maximise exposure: Twitter public updates - use provocative wording! (eg: I have just transferred money from my swiss bank account/bought a shotgun/attempted to assassinate Joe Bloggs) This will draw peoples attention Include your hashtag though, so you can be a twitter trending topic (and people know their friends haven't really bought a shotgun) Twitter direct messages Very visible to user, often set to come through as SMS messages to phone Example is to send them when a friend does something in the game, which draws you back in (similar to kevin rose's farmville anecdote) He seemed quite shy about answering a question about the ethics of getting naïve users to spam their friends. ------------------------ Unveiling the official web app survey results - Ryan Carson (carsonified) Anonymous survey of web apps was carried out on techcrunch.com http://webappsurvey2009.techcrunch.com They had 65 respondents and most were very "small time" - 0 users, $0 spent on marketing/development, etc Basically it was pretty useless data. How the survey is 'official' was never explained. ------------------------ Going global: the future of facebook connect - Cat Lee (facebook) Social elements are everywhere. You can use facebook to bring it to your site. Examples of news sites with a live feed of twitter-like messages next to live video stream of big events like Obama being sworn in Building blocks of facebook connect: - Identity - Connections - Activity Basically another product shill, but a lot of take-away points for the university. Possibility of sharing activity back to facebook profiles to encourage people to use glamlife more (stat on slide said it would increase referrals from facebook by more than 300%) Could be used to look at a user's network on fb and suggest friends on glamlife (friend finding is what they call it) Profile pics could replace gravatars for people who link their accounts. There's a lot we could do with it. ------------------------ Future of social web apps - Ed Anuff and Mike Malone (six apart) "A new way to build social applications" (side note, everyone so far has been American. British web app scene is obviously not too hot) Title very misleading, the whole talk is a big TypePad plug - borrring! Motion for building social apps, example: http://fowa.typepadmotion.com/ Doesn't look anything particularly special. Stuff at - http://developer.typepad.com/ ------------------------ How the Guardian is using APIs, frameworks and tools to build a mutualised newspaper - Chris Thorpe (guardian) Same stages in traditional newspaper model and online news Mutualisation - anarchistic way of making users into contributors (big definition somewhere) Government has made data available: http://data.hmg.gov.uk And there's a google group. Looks interesting. Basically the talk is about how they can get users to contribute to the website, and they'll save money that way instead of by charging a subscription fee Some very interesting ideas about how to live with the changes that are happening in online news They've seen the difficulties other industries (music) have with ignoring a changing marketplace Stark contrast to the Murdoch model ("give me more money" subscriptions for online news) They're very proud of their data, and it's all available. http://www.guardian.co.uk/open-platform ------------------------ How people will use the web in the future - Aza Raskin (mozilla) What's coming next for browsers? YOUcentric 1. Identity Browser will know who you are, so they can register/log you into sites without you having to do anything except press a button "log me in" Knows your credit card details so you can make payments with a button "yes, make payment" 2. Contact details and friends Browser knows your friends and their contact details If you want to email a friend, your browser will know you use gmail, and the friend's email address Or it might know that you use twitter more often, so will default to sending messages on twitter 3. Integrated Moving towards the 'web OS' GPS USB support (plugging in a camera might automatically direct you to flickr upload) Music library access etc Security is a problem, however You can pop-up a warning but people don't understand and just press the 'whatever' button So knowing that some of your friends are already using a service, the browser can suggest that it's probably ok, and so on 4. Make the web revolve around you Ubiquity - quicksilver for the web Taskcentric Lots of examples of having to break your train of thought to do things - something like 12 steps to translate a sentence on a website He shows using ubiquity - select text, bring up ubiquity, type "translate", press enter, and the text is replaced on the page "Think about what you want to do, not how to do it" Edit the web, example of deleting google's logo and typing 'Aza's search' as a replacement. This will then be saved locally. ------------------------ Day two ------------------------ The future of frontend engineering - Learning from Twitter - Brit Selvitelle (twitter) Asks the question: What makes Twitter developers tick? What makes developers tick? He hates email, but gets excited by gmail's 'new' icon, because he likes to play with new features. Tweetie 2 has a cool feature where when replying to a tweet, you can drag down to reveal the message you're replying too (in case you forget the context) Makes announcement that a 'Twitter labs' thing is coming (gmail labs style feature rollout) All a bit of a ramble, but gets some information in with: An update on really obvious things: Scale javascript like you scale your website (premature optimisation is bad) Release alpha 1 with an API Be optimistic! (Lars Rasmussen quote) Optimism is often characterised as a bad thing. What do you want? ------------------------ The future of the cloud - Simon Wardley (canonical) I've seen this presentation before at one of the cardiff web scene events. Lots of slides, pretty funny, actually quite informative. Basically this: http://www.youtube.com/watch?v=okqLxzWS5R4 But with more slides (248). Actually it's mostly word-for-word the same. He's a one-trick-pony, albeit it's a good trick. What is the cloud? Google has 67 definitions. He goes into some of them which reveal more definitions. Basically there is no definition. Because it isn't a thing, it's a transition. Trend from innovation to commodity status. Cloud is very much like sub-prime lending (in that it will lead to collapsing and general pain) The future of the cloud must be open source or we can kiss goodbye to our internet freedoms. ------------------------ Rails and the future of Agile - Yehuda Katz (merb and rails) Could be called 'Agile and the future of Rails' Merb => Rails 3 Quick Agile recap. "Defer decisions until the last responsible moment" Ruby is good for abstracting any code into methods. Which helps with deferring decisions. What database should I use? Don't have to decide yet, just use sqlite until you know which is best for you. Which model layer should I use? Don't have to decide that either. And so on, for caching, security, and performance. Rails gives you the most commonly useful one out of the box, and you can swap in your own choice later (the merb influence) New to Rails 3 - background tasks Rails has an agile culture (Dave Thomas first person to write a rails book, etc) This leads to innovation and contributions back to the wider development community (and these feed back into rails) Testing enables refactoring (enabling you to defer decisions), leading to cucumber Version control => github Experimentation => couchdb, mongodb, etc, which feeds back into ActiveRecord Ruby leads to Rails, which leads to Agile. ------------------------ Your app + mobile widgets = awesome - Joel Moss (codaset) and Sanj Matharu (vodafone) Widgets are something you find in a can of guiness. Mobile widgets are the same as apps. iPhone apps are great but require you to learn a new language - Objective C These ones are multiplatform, and built using HTML, CSS and JavaScript They can access network resources, local storage and device APIs You can test them on the desktop using Opera He shows off a phone that bears more than a passing resemblance to the iPhone. The widgets looks interesting though (might be a good, and easy way to reach non-iPhone-owning students) I wonder what phones this actually works on though? Apparently there's an 'app store' coming soon More info: http://jil.org/vodafone/appstar ------------------------ How to increase the accessibility of your web app - live demo of screenreader issues - Robin Christopherson (abilitynet) First, some facts and figures: 54% of over 65s have disabilities (20% of the general population) Disabled adults spend twice as much time online as the rest of the population Legal and General accessible relaunch led to a 90% increase in use (all use!) SEO and accessibility overlap (no percentage given for this one :P) Then follows a demo of screenreader (jaws) interpreting facebook. It's not pretty. At one point some PHP code gets read out. So they have to use the mobile version, which is better, but severely limited. HTML5 promises a lot. (http://blog.gingertech.net/2009/08/03/aspects-of-video-accessibility/) MVDA (NVDA?) is a free screen reader that is suggested as useful for testing. (we should probably do this) Demo of HTML5 video: not great If the controls are displayed, it reads out the percentage complete every second or so, so you can't hear the audio description. Google Chrome Frame is mentioned. It's an accessibility black hole (http://www.paciellogroup.com/blog/?p=444) Demo: It breaks screenreaders completely. You get nothing. Drag and drop keyboard interface using ARIA is demoed. Looks cool. CAPTCHAs are bad! Someone did a 3 month on, 3 month off experiment with them. Found that turned off, they had a 7.3% increase in conversion. Logic questions are better (for example 'is fire hot?') Enlightening demo, I've never seen it before. I wonder how our sites fare (probably not great) ------------------------ Marketing your web app - the future of brands online - Alex Hunter Why should I care about marketing? "Marketers are the cancer on the nutsack of creativity" However, you can't afford to ignore your brand. Example: who do you love? Coke vs Pepsi - they taste the same, but people feel an affinity for Coke, due to the strength of their brand. How will you get users to love you? You have to work your ass off. Marketing used to be monodirectional. There were billboards that you can't argue with, they just tell you what to think. They used to get away with anything ("Guiness is good for you", "More doctors smoke camels", etc) Now we have very finely tuned bullshit detectors. Now marketing is bidirectional. People can talk back. Brand now = YOU People don't want to connect with brands, they want to connect with people. Examples: Apple = Steve Jobs, Dig = Kevin Rose, Carsonified = Ryan Carson These people stake their professional reputation on their product. Are you willing to risk your reputation for your product? If not, give up now. Anecdote about JFK visiting NASA, he asks a janitor "what do you do?" and the janitor replies "I'm putting a man on the moon" Every member of the team should feel a part of the whole, and that is reflected in your brand. Be HUMAN. Example - digg They have the best corporate blog - everybody blogs! Not just the CEO/founders: developers, database guy, receptionist, etc http://blog.digg.com/ They have a TV show - diggnation. They have fans! It's all about identity with the people behind the brand. Don't just fudge it together, you have to put time, energy and thought into the brand. It's not about the name, stop dropping vowels from words and picking cute names, it's about values. Everyone says they are 'fun', 'ethical' and 'challenging'. What do these words mean? You have to take these things and show you are them more than other people. Fun = burrito eating competition Ethical = treat users with respect, example of a music site that gives their artists 70% of all the money they make (a huge percentage) ------------------------ The Future of Print - Lynne D Johnson (This is not about print, just print journalism) Print is not dying, but it is going to become more 'elite'. People who want it more will pay more money for it. Radio and TV didn't kill of print media because the same model worked there - 'CPM'. CPM = cost per thousand users (paid by advertisers) The Internet is different. They went to a CPC model (cost per click), but nobody clicks ads. Nobody even reads them. Google aggregates news so you don't even have to visit the original site. Craigslist does classified ads better - a huge revenue stream for newspapers, gone. Shops can email their customers directly to announce a sale. Social media disrupted it even more. People get their news (and ads) from their friends. Huffington Post beaks news before newspapers do. CNN iphone app is "amazing". I lost interest after a short time, it didn't seem very relevant and it was never really impressed on me why I should care. ------------------------ Musical interlude involving some guy named Peter Gregson playing the skeleton of a cello. He was pretty good, but the giant screen showing live feedback from his twitter, which we encouraged to send messages to, didn't work, rendering the whole exercise pretty pointless. ------------------------ Startup metrics for pirates: AARRR! - Dave McClure (founders fund) This guy had the most horribly designed slides I've ever seen. But he made a valid point that because of that, we were likely to remember them. His first slide was basically all about how the chances of making it as a startup are tiny and that "YOU WILL FAIL". It continued in much the same fashion. Web 2.0 is great, you can measure user behaviour in real time. What is the best attitude? Long-term: audacious, creative inspiration Short-term: humble, analytic rigour He mentions a blog post 'to the smartest person in the room' but there are so many with this title I couldn't find it. Optimise your product for the maximum happiness of your users. Only if they are happy will they pay you. Progress ≠ features (less = more) Focus on user experience (and distribution) Measure conversion - compare 2 or more options and see which one gets you better conversion rates Fast, frequent iteration (and feedback loop) Keep it simply and actionable (at this point I noted that we should ask students what they like about glamlife now, and what they would like from it, and kill the things that they don't like) http://www.youtube.com/watch?v=irjgfW0BIrw is Dave's video of startup metrics that he says you should watch http://startup-marketing.com/the-startup-pyramid/ is mentioned 3 key problems for startups: Management - What to measure Product - Building the right features Marketing - accessing 'web 2.0' channels Management: Which metrics to use? Why? Focus on a critical few actionable metrics. If you don't use the metric to make a decision, it's not actionable. Hypothesise customer lifecycle Target 3-5 conversion events (less is more) Test, measure, iterate to improve Product: What features to build? Ones that are easy to find, fun/useful, unique Features that increase conversion Measure with A/B testing, iterate FAST (daily/weekly) Optimise for conversion improvement 80% optimising current features 20% developing new ones Kill a feature! Stop adding features. Find the one thing that users love. Kill a feature each week, see how many people scream. When people scream loud, you know you've found your killer app. Bring it back, but better. Marketing: What channels? Which users? Why? High volume, low cost, high conversion Low-hanging fruit: blogs, SEO/SEM, landing pages, automated emails One step at a time: 1. make a good product: activation and retention 2. market it: acquisition and referral 3. make money!: revenue and profitability ------------------------ Practical advice for managing the growth of your web app - Chris Lea (virb) Following on from Dave's caustic sermon, this one took on a similar tone. Scalability vs efficiency They are not the same! Eventually your app will be I/O bound on *something* (cpu, database, disk, network, etc) Scalability is when overcoming I/O boundaries can be expressed in financial terms. Requirements: Must understand what your I/O boundaries are. Engineer your app so that you can add I/O stuff when you hit the bounds Know how much it costs Efficiency is when your cost of scaling is small Working on scaling problems is almost always beneficial Working on efficiency problems is sometimes beneficial Premature optimisation is the root of all evil. Steve Sounders' 14 rules: http://stevesouders.com/hpws/ do them! Useful things to know: When you are smaller, focus on scalability When you are bigger, focus on efficiency Scalability example: virb Aggregation of feeds was cpu bound Can add more hardware to speed it up (and know how much it costs) Efficiency would help but resources are best used elsewhere for now Efficiency example: facebook photos Storage bound It's easy to add more storage but they were adding tens of TB of storage per month, and it was getting very expensive They optimised their storage and saved a lot of money. Cal Henderson's book "building scalable web sites" is good. ------------------------ 3 startups launched on stage! First one was Aware Monitoring http://beta.awaremonitoring.com/fowa Beta code: FOWA2009 Broadersheet, a nice-looking iphone app for optimising your news reading. http://broadersheet.com Go test it, a cucumber-like testing thing, but you can set up the tests automatically by clicking your way around the site. Looks pretty cool. http://go-test.it/fowa ------------------------ Ryan Carson couldn't be bothered with the 'what is the future of web apps' panel, so: Bruce Lawson returns, he's doing the 'watch me make a html 5 page' thing, which I saw him do at FOWD. So I was kind of bored by it. Then Francisco Tolmasky returns, to show off some more stuff about cappuccino and atlas. Mentions that atlas is not ready yet, but you can do some similar stuff in interface builder and cappuccino with the nib2cib conversion tool. ------------------------ Gary Vaynerchuk He's a funny guy, and very charismatic. Makes a few points: People ask him why he expects to sell a book when all his content is online for free. He thinks it's to do with the 'thank you' economy. People want to thank him, so they buy his book. He loves how new media has made everything much more transparent: "If you're a dickface, people are going to find out." "If you don't care about your users, you're going to lose" Apparently very few people do. He knows of "lots of shit products" that are successful because their owners care. ------------------------ Then I went back to the hotel to nurse my cold, by eating organic fruit. All in all, second day better than first day, quite a lot to take away. If anyone wants to have a go at summing up some of the notes here, I'd love to read it! I missed the 'Kevin and Gary show' was it any good?