I use overture as an index or guideline and never take the actual number literal; at best you could use several values of keywords you are familiar with and put them into relationhip to each other - sort of calibrate them.
How about if the Overture data was exactly right, but of course has no way of predicting anything about what happens on Google searches, and your SERP CTR is 1% anyway?
Seems like tools that provide any numerical displays whatsoever irresistably draw webmasters into thinking that human search behavior can be predicted with great accuracy and plugged into a spreadsheet.
From my understanding here are the problems using overture as a keyword tool.
- removes plurals - the ordering of keywords is skewed (maybe thats wt, or gks can't remember to lazy to look) - in highly competitive keywords, webmasters pinging the search term raises the value and gives you an inaccurate estimate of the type of volumn you may receive
Example: I chose a highly competitive keyphrase in the legal profession. It has relatively high CPC of 2+ dollars. Overture said about 10,000 searches a month. I maintained the #1 position in MSN and Yahoo for about 2 months and got hardly any traffic. If you have looked at heatmaps, then the two things don't add up.
This was already said, but the Overture tool is just an estimate. It is not an exact science, and it only covers searches within Overture. It doesn't tell you how much trafic to expect. Just out of curiosity, what kind of CTR would you expect from a low-search volume phrase? I think 1% is pretty good.