Found this interesting article via sphinn.com http://www.ppchero.com/determine-keyword-quality-score/ What was particularly interesting I thought, was what he's saying about new keywords having a Quality Score assigned to them based on what their ctr has been for other advertisers - and that keywords with a bad Quality Score to start with might ring alarm bells for you that these keywords just never perform well for anybody
New keywords having QS based on what their CTR? But if they are new there wouldn't be much data to based the QS on, would there?
Keywords that are new to your campaign have a Quality Score based on what their ctr has been for other advertisers.
What the guy who wrote the article is saying, might be true, but it's not all that informative. He is saying a keyword is working well for advertisers because you happen to get a high QS for that keyword right from the off. How can one gage how well a keyword is doing just by going off the QS/CTR? You can't. I have some keywords that are low cost per conversion but the QS is 6/10 (low CTR compared to the rest). I have others that don't convert well but their QS is 10/10 (Very good CTR compared to the rest). It all boils down to the one thing - CTR is over hyped and meaningless compared to ROI/Cost Per Conversion. I often see people get excited because their CTR is superb (very high), yet they are getting too many clicks from the wrong sort of people. A low CTR can often get more targetted clicks - a higher Conversion Rate and better profits, even though the QS suffers due to this low CTR.
muchacho79 has summed it up very nicely. Look at conversion rates and ROI before worrying about QS. If you want improved QS, either tweak your landing page, or set up custom landing pages per ad group, using standard seo practices such as title, h1's, keyword density, etc. Don't overoptimise, just optimise, keeping in mind great QS may have a few pence/cents on cpc, but if it's at the expense of your visitor engagement, then you've just shot yourself in the foot...