It's descriptive, not prescriptive, and is sort of a survey of psych experiments that deal with perceived happiness, remembered happiness and predicted happiness. What you want to do with this information is up to you.
The book is well written and very colloquial, especially considering the number of footnotes there are. Here are some of my favorite "bits" from the book:
We tend to rationalize misfortune. This only works, though, if our misfortune is bad enough (p. 181). For example, if I stub my toe, I'll be annoyed. If I get in a car accident, I'll be grateful for having survived. Gilbert cites data on people with cancer who can always find people who are in worse situations than themselves and thus tend to rate themselves happier than people who are not sick at all.Lots of interesting experiments here, and worth reading.
No one likes to think of themselves as normal. The most extreme stats he cites are that 90% of motorists consider themselves better-than-average drivers and 94% of college professors consider themselves to be better-than-average teachers (p. 229). At those rates, how bad a teacher do you have to be to think you're worse than average?