Change your landing page, ads. There are tons of variables that go into it. You have to keep experimenting to find the best mix.
Give more on your landing page. Don't just sell, educate. Give people a reason to stay on your landing page. If you don't capture someone's attention in 6 seconds, they leave, indicating a bad experience on your site, hence a low quality score. Change something, test, change something else, test again, rinse and repeat until successful.
Tom, thanks for your advice! I didn't think QS had anything to do with the landing page. If you're correct, then it's very similar to trying to rank well in organic search. It would be very easy for Google to implement too - just use the same technology/part of the algorithm.
Quality score has a lot to do with your landing page. Make sure it's fully optimized for your keywords. If necessary use multiple landing pages with keyword insertion on them. Adwords requires you do a bit of seo work in order to get high quality scores. It's a proven fact.
You are correct, partly. About two thirds of QS is a comparison of your CTR to competitors for that same keyword. Another 25% is keyword relevancy. But this should be easy. If your page is about dog collars, you don't use cell phone keywords. While Tom's suggestion is good (you need to grab their attention right away if you'll have a chance at selling) and you should test different page versions, there is no metrics that Google uses to calculate QS once someone clicks an ad. Unless you get a QS of 1, which means you are not following one or more of the Adwords rules, there is nothing you can do on your page which will increase your quality. It's mainly about your CTR vs that of others.
Google tells use that Quality Score is about 70% CTR based, 20% Relevancy and 10% landing page. However we all know there are several other "unknown" variables in there that Google will never disclose.
The other thing not to forget is that Google also give 'weight' to your past Adwords ability. So some seasoned adwords users will get a better score while less experienced will have a 'handicap' of sorts to get past first.
It's because campaign and account history very much factors into ad rank. http://adwords.google.com/support/aw/bin/answer.py?hl=en&answer=10215
I just knew someone would bring that link up. The relevant part Eschatonic is referring to is under Quality Score Formulas. The second bullet point says: "Your account history, which is measured by the CTR of all the ads and keywords in your account" First, let's assume that the first bullet point is likely more important since it is first. In other words, it has more weight in the QS calculation. The question is, how much weight does this second point have? Of course, no one knows for sure except Google. However, it can't have very much weight. I base this on two things. First, I've taken very poor QS accounts across the board (2s and 3s) and brought them up to 7s and sometimes 10s within days. So, either the weight is very small or the statement is incorrect, a leftover relic of days gone by. My point is, if I was able to bring QS to high levels quickly on languishing accounts, the weight is either extremely small or it doesn't exist and the statement is false. Also, logically, why would you punish someone for not creating good ads in the past? This would mean, or make it appear that, forget about ever getting a QS of 10. As I've said, that simply is not the case as I've done it more than once. Again, that puts doubt into the statement or the weight is very, very small. Furthermore, an account can be promoting many different products and services. It doesn't make sense to use the CTR of these different products and compare them. It would make more sense instead to compare to the previous history of the *group*, not the account. In fact, the third bullet point basically says that. Maybe that's what they meant and the wording wrong or purposely so to throw off people trying to figure it out. But even if the wording is correct, I would not put much weight on it. I could suddenly figure it out and create the best ad ever. Punishing me for the previously poor ads should not be taken into account which is what that statement says. A further clue is the seventh and next to last bullet: "Your account's performance in the geographical region where the ad will be shown" I'm sure they meant the group's or at most, the campaign's performance since geotargeting is set at the campaign level. You can have campaigns targeting different regions. Poorly worded or done so on purpose? Whatever the case, I still think that previous *account* history has little or no effect. It should not even be considered in the calculation.