The thing is, if you promise very little, you don't get a chance to deliver because I'll ignore you. And if you promise too much, you don't get a chance to deliver, because I won't believe you…
Hence the paradox. The more you promise, the less likely you are to achieve delight and the less likely you are to earn the trust to get the gig in the first place. Salespeople often want you to allow them to overpromise, because it gets them through the RFP. Marketers, if they're smart, will push you (the CEO) to underpromise, since that's where the word of mouth is going to come from.
I have worked with someone who is very good at the promising part. She enjoys it. And when the promises don't work out, she's always ready with the perfect excuse. This is a great strategy if you have a regular job and the excuses are really terrific, but if you need internal or external clients, it gets old pretty fast. It certainly doesn't lead to the sort of word of mouth one is eager to encounter.
Surgeons have this problem all the time. They promise a complete, pain-free recovery and work hard to build up a positive expectation, particularly for elective surgery. And the entire time you're in bed, in pain, unable to pee, all you can do is hate on the doctor.
This is one reason why recovering from failure is such a great opportunity. If you or your organization fail and then you pull out all the stops to recover or make good, the expectation/delivery gap is huge. You don't win because you did a good job, you win because you so dramatically exceeded expectations.