We have recently been looking at testing websites on mobile devices. While most sites are still in Jakob Nielsen's 3 classes of mobile user experience i.e. horrible, bad or impoverished, where site owners have taken the trouble to develop mobile sites they can be quite good (e.g. BBC, Facebook, Google Maps, etc.) However, to have an effective mobile website, usability is only half (albeit an important half) of the story.
What our research shows is that users can often have very different goals that they wish to achieve when using their mobiles than on a desk or lap top. Virtually nobody is going to use a mobile to look at a web site if they have a larger machine handy. So you use your mobile when you are out and about and you use it for very different things. We did some work testing a financial services site on mobiles and unsurprisingly no one wanted to use their mobiles to research these products, but they were keen to check the value of their portfolios when on the move. Often mobile users will want an organisation's contact details which can take ages to find even when a site has been designed for mobile use.
Our belief is that organisations looking to build their first mobile web site should identify a very small number of user goals, possibly only one, that the mobile site is to support and concentrate on doing this well. For some sites, like the ones mentioned above, the goals of the user on the move are fairly obvious, for most sites they are not. In which case, research with the target audience to work out the goals to be supported is essential.