There are a couple things that I am skeptical of, however.
Let’s, for example, consider this concept of “breakout” status. If, using your Kate Spade search as an example, “Kate Spade sale” reached breakout status recently, can’t we assume that the search term is a “new” one, and has reached its status simply from going from 0% to being in existence, in which case its growth is essentially infinite at the start?

Let’s say Kate Spade had a summer sale, and sent out postcards to all of its registered clients. For the first time, some those clients hopped online and searched for “Kate Spade Sale”, creating the breakout status listed. Let’s say 100 people did just that.

However, some of those same clients probably also searched for “Kate Spade purse.” However, if 1000 people did that, but “Kate Spade purse” was already receiving 10,000 hits, this hit would only report a 10% growth, with “Kate Spade Sale”, a new search, has an infinite, or “breakout growth.”

In this style of situation, which is all too feasible, we can see the potential for the growth per volume number to differ largely from the actual volume of search.
Here, as a marketing manager, I would probably be included to invest more media in “Kate Spade Purse” than “Sale,” even though the growth purse volume of the latter is higher.

What this tells me is that the use of Google Insight needs to be coordinated with a tool that tells you actual market volume.