Big methods
So today the great man of usability says: High-Cost Usability Sometimes Makes Sense Summary: Computing the net present value (NPV) lets you estimate the most profitable level of usability investment. For big projects, expensive usability can pay off.

I don't really agree Jakob, much as I adore you and your conferences and your fashionable glamor publicity shots.

Well, ok. I sort of don't agree. I am a strong advocate of the right tool for the job but simply advocating for a single large, monolitic usability service effort is like harking back to a single mainframe for the hardware world. Very appealing in a marketing sense. One machine in one room. Everybody obeys. We move on....hmmmm...

The real world though these days (and it IS an ever more agile or as I like to say "humane design" world) demands fast, powerful services that inter-relate into a community of services and deliver a networked effective value.

So I am for the aggregation of multiple, appropriate, guerrilla or light-weight methods that can be iterated rapidly, assessed, converted into coded solutions and repeated. We still do all the up-front thinking and planning but when it comes to getting it done - Let's just go do it! Learn from any mistakes and Do it again.