tedster - 12:35 am on Jun 25, 2010 (gmt 0)
It's important in doing this analysis to be very precise, so I'd like to summarize some ideas.
1. If your rankings have improved on an important keyword but your search traffic went down for that same keyword, then maybe your rankings are not really up.
2. You may be getting customized or personalized results when you check rankings.
3. If you use an automated rank checker, you may be getting red-herring data (some people have reported this happening.)
4. Rankings may be up only part of the time - Google seems to be doing more dayparting than ever before. So take a day and do some manual hourly checks and see if the ranking is regularly where you thought. Check your Webmaster Tools and see what kind of impressions and clicks Google is reporting for your terms.
5. Rankings may be up but not for your main geographical target. Check the IP addresses of your search traffic.
6. Maybe your ranking went up, but the Suggestions drop down changed and is taking some of the previous traffic off to a different search.
7. If you're comparing data from this year to last year, remember that the Google interface changed seriously in May 2010. I've seen some rough data that suggests that #1 on today's search results page generates up to 30% fewer clicks at #1 than a similar #1 would have received last year. Now don't take that and run with it just yet - I said the the data was "rough".
I'll bet there are other possibilities, too and would love to hear more ideas. But what I'm saying most of all is to zero in on a particular search term. I don't think that insights will come from looking at high level, aggregated data.