I'm aware of this test* (I imagine most planners have come across it at some point) It apparently was partly the basis for AMV rescuing the Sainsbury's account a few years ago. They discovered that people sleepwalk through supermarkets and do not notice any of the new and exciting things that supermarkets do to entice them to spend more money. People spend less than a second scanning the shelves, look for their usual brand (this counts as "loyalty" by some measures) and buy it. If you are right in front of their eye for that 0.7 of a second, you might, might get them to switch. They illustrated this to the client using the same monkey experiment (or a variant thereof) mentioned above.
People don't see things if they are engaged in a different, directed task. Hence "Try something new today". (I think it is a genius strategy, by the way.)
For me, there is a huge irony here - the experiment is basically an illustration of the weakness of the traditional advertising model, and in particular the idea of interruption or persuasion. It isn't that people don't want to see your ad. It's just that when doing something else, they are so good at editing that they don't even notice that they are ignoring them. Making a TV ad from an experiment that demonstrates why a lot of ads don't work is a piece of surreal genius.
*I knew what the ad was about. I knew it was coming. I knew something would happen. And I still missed the bear. What a dick. More importantly, when describing the ad the next day, I totally forgot what it was for. It took me ages to remember it was about cycling. As the prof pointed out in Scamp's post above, if they had asked him they could have done something with bicycles. I think that would have made a stronger ad.