also agree on the dangers of double-headed teams. you need to have a direction (and of course be agile, the ideal case here being a startup like imvu.com)

The one point I would like to blog about more is the A/B testing here, and how it’s setup. You can do A/B testing for any new or existing feature which should be improved and then track:
- signup rates if it’s a web application
- sales if it’s a ecommerce webiste
- pageviews or stickyness if it’s a media site
- visitor loyalty
- or any other metric

Having clear A/B results will help you afterwards in your development

I know there’s even web software which will change its copy / design following automated A/B tests.

Indeed, I can certainly conceive applications where User-Centric development is appropriate, especially if it ties in to the revenue model in a synergistic way. In the past I’ve worked on a very large and profitable web based application that has a “membership/pay-per-use” revenue model. In this case it made much sense to prioritize based on what users would use most, therefore surveys and usability studies were deeply ingrained into our feature definition process. This proved to be extremely successful, however much of the value in the application did not come from that process. The user-centric process mostly yielded logical “next” features or nice-to-have’s, which help achieve high user satisfaction/retention rates. The internal processes yielded features that were highly innovative and often designed to displace competitors, which help achieve the acquisition of new users.
More importantly, we had to reconcile both processes as to not have a “double-headed” development team. The result was something like an agile process, but with rigorous QA, and very strongly defined architectural guidelines. Another interesting thing to note is that the work was distributed between several offices (Eastern Canada, England, India) so there was always a team of designers, coders, QAs active on the project. Bug reports and usability issue reports always came in no more than 16 hours after the daily build, day after day, which was a huge component of successful delivery.

Anyway, your post gives me an idea for an interesting project…. I’ll sleep on it and if it sticks I may consider it :)

I totally understand what you mean. It’s important for any software project to have a visionary architect that can drive a team.

However, the point of this article is for a web application, when for instance you’re building something and you don’t know the problem to solve.

Take for instance the case of facebook, or any other startup which are experimenting new web features. The guys at Facebook don’t know what works and what users would accept. Same for Google, they have automated user feedback in their Google adwords product, which changes the probability of an ad showing up depending on the ad performance (if it has been clicked or not).

In those cases, you have to build user feedback in your software development methodology.

Of course, this doesn’t apply if you are building a website which has strict specifications, where the solutions are known, and where there is no risks. In those cases, you can follow the standard methodology

What you are suggesting is to drive the entire design and development by leveraging a very large and open public commitee. That is comparable to every client in the restaurant picking one ingredient for the sauce-of-the-day, leaving the chef’s expertise out of the equation.

What is needed to improve the rate of success, quality and innovation is for companies to finally fully acknowledge the role of “visionary architect” as a fundamental driver of software projects. But it can’t be a strawman, it has to be the real deal. The same kind of person as those that nearly-single-handedly built million dollar dot coms out of a wet basement with server hardware duct-taped together. The type of person that can come up with the idea, design the interface, architect the systems, code the whole thing and put it live… with two extras – a fantastic budget and the power of delegation.

That’s how yahoo, google, facebook, digg, youtube and so on came to be innovative software that people absolutely love to use, and make billions.