The promise of the 1:1 future is here but it's not always easy to execute.
One of the fundamental tenets of direct marketing is that audience selection is the most important driver of response. So when people talk about and do "data-driven" marketing, they are usually talking about using data to select audiences.
Two recent emails that I received each demonstrate how marketers use data to drive (what they think are) relevant customer communications. They are both fairly simple in concept and are great examples of how data can be used to drive relevancy and results.
Staples wanted to promote its Copy and Print Center, but on a specific location basis. So they sent out the following email:

Whether they were only intending to support the Natick store or they were simply trying to match my record with a specific store, they blew it here. Assuming I was the right prospect to have a sign or banner made, I don't live anywhere near Natick, have never been to that store and, thus have no purchase record there.
What likely happened is that they had a data error of some sort, which could have been limited to my record or could have affected the whole campaign.
Now for another example.
Look at how United uses my actual Mileage Plus® balance to show how many miles I'll have once I take their Visa card:

Not only are they making this exactly relevant to me, they are probably varying their offer based on my current balance. While I might not respond, it's a much more relevant communication for one primary reason: the email is based on actual purchase activity, as illustrated by my balance.
In each of these cases, you can see that data-driven personalization is not easy to execute, must be thoroughly tested and, when done right, can be very impactful.