Future of Web Apps

Bradley Horowitz (Yahoo!) on Social Interaction - What the Future Holds

Feb 20, 2007

People aren't 'users', but people. We are all creators, synthesizers and consumers, people participate in all these roles.

On Flickr, there are amazing photos taken by everyday people. Created an algorithm to determine 'interestingness' based on number of comments, views, links etc. There could be no gaming of the system or spam, and it was retroactive, i.e. you can see which photo was the most interesting on the day Flickr launched. Co-occurrance of tags is used to cluster images around themes, e.g. romantic love, symbols of love, pets we love. This turns every user into an editor - their activity on Flickr is reflected back.

Yahoo! Pipes allows you to take any RSS feed and pull it on to one page and combine the data. E.g., take Craigslist apartment listings and Yahoo! Local nursery listings in order to find apartments for rent near daycare centres. Means it won't just be developers who can mash-up APIs.

Werner Vogels (Amazon) on Why and How It's Easier Than Ever Before To Build A Web Business and Compete With Anyone

Feb 20, 2007

Werner said that the old 'push' model of developing a web app or service depended on planning ahead, anticipating demand, top-down design, and is very resource-centric, providing limited opportunities for re-engineering. But it's very hard to predict what resources you are going to need and when with any accuracy.

A better way is to 'pull', where demand is uncertain, design emergent, decentralised, modular, people-centric, and unlimited opportunities for re-engineering. Find resources, innovate on top of them, run the service and reflect on whether the resources you have do what you want, then swap them out if they don't.

There is no point spending money on resources if you're not using them. Both economically sensible and environmentally friendly to acquire resources on demand and release them when they are not needed. In traditional model, 70% of your time is spent on 'undifferentiated heavy lifting', and only 30% is actually spent working on your product.

Amazon provide on-demand storage (S3) and processing (EC2) which can cut costs and facilitate scaling.

Matthew Ogle & Anil Bawa Cavia, (Last.fm) - Lessons from the Building of the World's Largest Social Music Platform

Feb 20, 2007

Matt and Anil shared some of the many lessons from their early growth period:

Don't overextend: scale with your growth, not before
Make sure revenue sources scale with increased usage, e.g. Google ads, user subscriptions
Involve users in your web application's story. If it's a social network, recruiting new users should improve their own experience
Be as open as you can. Bad news is better than no news.

What works when you are six hackers sitting round a table doesn't work for 40 people, so:

Use lightweight processes
Pick simple tools
Radiate information through the company

They had a problem with tag clouds 'spam', but decided that whilst unflattering, these tags were valid. Instead of blacklisting, they used attention data about how much people listened to the track to weight the tags they used to describe it. If you listen to Paris Hilton a lot, you have more of a say on what shows up in her tag cloud.

Ben Holmes on Everything you Need to Know About Funding

Feb 20, 2007

Index Ventures' Ben Holmes gave an investor's perspective on what makes an attractive web company. Here's a summary of his presentation...

VCs are looking for good developers and technical founders more than ideas. But whilst there isn't as much money in Europe as the US, after successes such as Skype it's easier to make the case for investing in the EU. UK investment in web apps and 'Web 2.0' companies is increasing, up to £79m in 2006.

VCs invest money over three to five years. Half their investments will lose money; a third will break even, and one sixth will make money. Unlike the financial services ads say, past performance is a strong indicator of future performance for VCs - if they have been successful before, they are more likely to pick winners again.

Companies should look for VC money when they have a great team (more important than a great idea), excellent development capability, and a large potential market opportunity. It's important to be honest with yourself about whether you are ready for VC funding, because if you are not, you can end up stuck in a company that's going nowhere.

Notes from Tara Hunt on Building Online Communities

Feb 20, 2007

Tara discussed in detail different ways of thinking about and encouraging communities online. Here's a summary of her talk.

There are three levels of community:

Lightweight social processes, e.g. Digg, Last.fm, or Del.icio.us.
Collaborative information structures, e.g. Flickr, YouTube or Threadless.
High-end collaboration, e.g. Wikipedia, Lostpedia or CouchSurfing.

Communities help improve customer loyalty, are self-policing, amplify word of mouth and provide better feedback.

Common themes to successful communities include:

have a sense of fun/play
keep the dialogue going: especially crucial in the early days
"wouldn't it be awesome if…": take an experimental approach to development
create simple platforms for building on
tell compelling stories
reward community members

Feelings of membership arise from:

creation of community boundaries
perception of emotional safety
sense of belonging and identification
use of common symbols, languages
shared emotional connections

You can't create community. It takes time, patience and hard work for a community to emerge, but the rewards are high.

Notes from Mike Arrington on The Future of Start-ups and Web Companies

Feb 20, 2007

Michael Arrington was the first speaker in today's Future of Web Apps. This is a brief summary of his presentation

What makes a good start-up? To be a successful, Arrington says you need a good idea, a business plan, a revenue model and passion. You need to build it cheaply and avoid spending too much money. Solve a real problem, remove friction, invent a market or, more enjoyably, destroy one. If there's no buzz around your start-up, rethink your product, not your marketing.

Of course there are always companies that flaunt these guidelines. YouTube threw away their original business plan, flaunted international copyright law, had no revenue stream (and really still don't), and spent more than $1m/month in bandwidth. Yet they were still successful.

Opportunities for start-ups at the moment include applications that work as well offline and online - Adobe's new Apollo platform allows for the porting of web apps offline; data and service portability; and the mobile arena.

Compiled by Suw Charman