Jim:

While it's smart to look at popular [historical] search terms that search engines are capturing, have you considered the possibility that popular terms are (by definition) terms that don't lead to actionable business?

Consider how people use Google for example - they typically start with just a few keywords that are ambiguous and generally produce lots of hits that are mostly unusable. These are the popular terms because they represent where people commonly begin for any search effort, but not necessarily because they represent anything specific that people want to look at. Terms like this typically don't produce recommendations that are good enough to cause the searcher to *stop* searching. As the searcher attempts to find useful recommendations, the next query includes additional terms to try to focus on exactly what information will help them. As they refine their search, they get close(r) to what they really want and further from popular terms the like of which are recommended by tools such as the one you cited.

As consumers close in on results that are useful, their phrases become less and less common -- e.g., less popular -- until they *stop* searching. When they stop searching, one of two outcomes are possible -

1. They gave up because they didn't find what they wanted;
2. They found exactly what they wanted.

In my view, it's more important to know the last thing someone searched for, not the most popular thing they searched for. Unfortunately the search industry is not geared to provide that - they would prefer that PPC budgets are spent continually bidding up prices on the popular terms.

Based on this simple understanding of how consumers use search engines, and how search tools focus on popularity -- I have a hunch we're not assessing the topology of customer behavior in search. It may be that the query terms used when they *stop* searching are more valuable than the popular terms recommended by research tools that focus solely on popularity.