It is getting late and my brain is about off. So here are some random thoughts on improving PPC campaigns:
1) Local search is done by IP mapping. Mapping by country is 100% accurate. But state IP mapping is only 50% accurate. Then City IP mapping is only 30% accurate. This is due to many reasons such as AOL traffic all coming from one spot. That type of thing. So the suggestion is to have a general national campaign then have local campaigns to target local traffic. Local targeting is shown to have higher click through rates and higher conversion rates. But if a user does not get IP mapped correctly then the national campaign will catch them.
2) Use “Day-Parting”. This is the idea of figuring out what time of the day you have the best conversion rate and bid high during that time. Then bid low on the poor converting times. Also some days of the week are better than others. All the search engines have the ability to automatically change your bids based on time of day and day of week. The hard part is to figure out when to raise or lower the bids. You can get this info from log files or analytics programs. Companies are seeing great returns by using this technique. Day-parting works well if you have competitive bid conditions, limited sales windows, identifiable demographics by time, business compatibility and if existing stats point to the success of a day-parting technique.
3) Target the long tail keywords. The long tail are those millions of words that have little traffic. We need to do this better at World Vital Records. But you need a good system to help manage this. Long tail volume can really add up. Users here also have very clear intentions (usually more search terms in the query). There is also less competition for these words which means that bids are lower.
4) Steps to take action on. First calculate your click value and what is the probability of conversion. Model your market and track everything for profitability.
5) Don’t put your words into “buckets”. You can always break them down by geography, by product, or by request. It is tempting to group keywords this way. But the best way is to do every combination. IE location-product-request all together. You can do this by tagging your keywords.
6) Use dynamic keyword insertion (sometimes). DKI is a powerful tool that can make things easier. But it can also decrease your CTR. It is not the solution for everything. If you are using DKI use phase matching and not broad matching. It is good for phrases that have 1 dominant keyword.
7) Think of keywords as investments. Don’t get emotional and don’t get attached to keywords. Either they work or they don’t. Set rules and objectives for each keyword. If these objectives are not met get rid of them. 
Diversify your keywords. More keywords leverage the long-tail.
9) Lowering bids or turning off your ads does NOT affect your quality score. I have always wondered about this. But everyone agreed that Google understands differences in bid placement and does not penalize you for adjusting your bids.
no comment untill now