/Opinion/ The key to success is failure
12/12/2007 | Filed under Discover > Opinion
Hannah Donovan, Last.fm’s head of creative, provides five tips for building a successful web application. Her top tip? Make mistakes
Last.fm has seen quite a bit of press lately, and has even been dubbed “the most successful UK Web 2.0 start-up”. Thinking about what makes a successful website, and how we’re trying to do it, I realised that success is just finding out what works and doing it consistently. How do you find out what works? Process. For us, and probably many other start-ups, this involves trial and error. We’ve screwed up a lot, and that’s good! Without failure, we would never have come to some of the key realisations that have shaped our web development and design.
I’ll begin with a brief history. Last.fm was founded in 2002, and it didn’t hire its first professional, front- end developer until 2005. A year later, it hired its first designer (me). This seems like the wrong way to build a good web product, and it sort of is. However, where Last.fm failed in launching a usable, well-coded website, we succeeded in getting our idea out there.
Put perfection behind you
At the time, Last.fm did something that no other website did. Timing is important, and we focused on getting the idea into circulation; this was a critical realisation, and it’s a guiding principle that we still apply. It’s often a challenge, especially if you’re like me, because I like things to be perfect. There isn’t always time for perfection. It’s better to get the idea out there and refine it later. MySpace is an excellent example of this (and where MySpace has failed is in refining the product).
My second piece of advice is don’t release new visuals without new functionality. This is the important lesson we learned when refining the features we had released quickly. If the design needs an upgrade to make it more usable, there’s a chance that the functionality does, too. It’s much better to release these at the same time, and it will keep your users much happier. Nobody likes interface changes that appear to be unnecessary.
It’s often useful to expose functionality (within reason). We have a particular problem with some of our features seeming like they “work by magic”. Ensuring that the user doesn’t have to make too many cognitive leaps, and exposing some of the technical process, are good ways to turn something that’s complicated in one step into something that’s less complicated in two or three.
My third tip: remember that designers and developers work on the same team. It became necessary for all teams at Last.fm to internalise the interfaces and technical processes. The easiest way is to have designers and developers working together seamlessly. Communication is key. It’s important that everyone speaks the same language to maintain visibility and accountability. Transparent work processes, as well as quick daily meetings, also help.
We use three different communication tools: Basecamp for management and conversation about a project, Trac for bug tracking and outstanding tickets, and irc, our “osmotic communication” method. We also use an altered version of Scrum development and do five-minute catch-ups at the end of the day (for designers and developers alike). So far, this has been easily scalable.
Hitting the hard stuff
Tip number four: do the hard stuff first, and use iterations. We were working well, yet still missing deadlines. To stop “feature creep” (when a release is delayed because new features keep being added), we followed this rule. Now, should we have to pull anything out before a release, the important work is already done. This also makes it easier to put perfection behind you. We consider the “hard stuff” to be anything along the lines of creating forms, editing, generating, configuring and searching.
My final tip is use broad brushstrokes. We managed to get this right when we internationalised and picked the most difficult language first (Japanese) – we worked big and broad, and tackled the fine details later. With some projects (and ours included 12 languages, a fluid layout and cross-platform software), thinking in broad terms equals sanity.
Not all of these processes will work for everyone, of course, but finding out what does can’t happen without mistakes. With the greatest errors come the best insights about what works, what doesn’t, and what can work consistently.
Originally from Canada, Hannah Donovan, head of creative at Last.fm, obtained a Design degree from the University of Alberta, where she cut her teeth on the web instead of going outside. When she’s not thinking about the web, she likes to draw monsters.
Comments
Philipp / 18/12/2007 / 12:21 / http://www.rechtsanwalt-blawg.de/
Thanks for that tips. I would want to point out one of them (the most important from my experience).
Don't try to do the things perfect. You can't - but you waste money. First there's a rule written in stone: The last 5% of your work take up 90% of your time. So build your product up to 95% then release. The other thing is, that a perfect work is just perfect for you. What's with your customers?
Do they value the last 5%? Really?
Diptanshu Sharma / 25/12/2007 / 06:55 / http://www.dataworldindia.com
I live by this philosophy. One must not be afraid to make mistakes. It's a learning process, and mistakes are a part of life. I too like things perfect, but that's not how life always is. The important thing is to do it, rather than slog trying to make it perfect.
Timing is vital for survival on the Net, and for us in 1998 this was why we were in the news. Our site went through three versions; the first version was extremely basic because the idea was to get the message out asap. I always encourage people to not be afraid and 'just do it'. Perfection can come later. Perfection and design are relative things, so no point focusing on issues that don't seem 'perfect. More important is to enjoy than worry that someone will find fault in your design. Focus on what's important. It's easy to let creativity and perfection be our own enemies.
Gary Rice / 28/12/2007 / 14:44 / http://www.Vlook.co.uk
I agree with you on not creating the perfect design, because its a waist of money and time. Its good to sit back and see what needs changing and what has been missed in the original design process
Roger Atkinson / 29/12/2007 / 09:15 / http://www.the12voltshop.co.uk
I think that the info is really good, but she's just saying what we've all been doing for years. Now I'm an old git (55 years worth) that has been messing around the net designing the odd website since day 1 and doing exactly what she says. Half the fun of building websites is the 'suck it and see' part. I'm just finishing my latest project for a friend. Over 400 pages of frames and straight HTML, when it all works and looks like we want it then I'll think about using a database or cms or something, but half the fun is building, chopping and changeing. (And she's selling stuff so something must work). Great mag keep up the good work.
Harry Selent / 24/03/2008 / 20:09 / http://www.medicalcharting.com
Reminds me of the quote from Peter Drucker, famous mgmt consultant who said you get more bang for your buck by doing the right things, rather then doing them right. Its more important to do something, than sit around and wait for perfection before you start.
Chad Patton / 27/03/2008 / 01:05 / http://www.iondetoxpro.com
I can appreciate your advice especially lately. My company put me in charge of designing and getting a website up and running. It was frustating because I felt I was learning a new language. It was also exiciting seeing your creation come to life. I did have to get over the urge to make it perfect before you get it our there. I am constantly tweaking things so I found it was better to get it 95% were we wanted it and then just keep working on it and letting it evolve.
Collin LaHay / 05/04/2008 / 21:10 / http://mixedmarketarts.com
It is said that people learn more from their companies that fail than the ones that make millions.
