There's an involved, technical and (for people like me) fascinating conversation in tech about smartphone apps and the web - what can each do, how discovery works, how they interplay, what Google plans with Chrome, how watches affect things, whether the web will take over as the dominant form and so on.
But for an actual brand, developer or publisher wondering if they should do an app or a website, I generally answer that the calculation is much simpler and less technical:
Do people want to put your icon on their home screen?
Do you have the kind of relationship, and proposition, that people will want to engage with enough to put your icon on their phone? If the answer to this is 'yes', then you should have an app - if only because the app store is the way to do that that people understand, and they'll look for you in the app store. Once that app is there, you can leverage all the interesting and sophisticated things that you might do with it, or you might manage the flow of information just like your website, but the app has to be there.
If you don't have that relationship, then all the clever things you can imagine you could do with Apple or Google's new APIs are irrelevant and your strategy should focus on the web (and social). (By extension, of course, there are some people legitimately wondering if they should have their own website - should a plumber be on the web or in Yelp?)
Meanwhile, you should have a website that works well on mobile regardless of whether you also have an app, and that site should give your complete proposition, since that of course is where links from Google and Facebook will take people. Too many companies present the potential customer with a website that says 'screw you, install our app', and then an app store page that says 'screw you, install our app', and then a first-run screen that says 'screw you, create an account/sign in with Facebook'. You do have to earn an install, I think.
In either of these cases - whether you have an app and a website or just a website, you should presume that your customers will engage with you only on mobile. A large proportion of smartphone use happens on wifi and either at home or at work - people use mobile to snack, yes, but they also use it to do almost anything they do on the desktop internet. So your mobile proposition should not start from thinking 'what would people want when they're mobile?, but rather on 'this customer may only see us on mobile, so how do we shape our proposition to reflect that?". Indeed, as I wrote here, it's not mobile that gets the cut-down, basic experience - it's the PC.