Hello, I have a method for keyword bidding and mining. It's not perfect, but if you're new, it gets you past the paralysis of inaction and helps increase you're click through rate. Tools: Website Landing pages A goal (sales, subscribers, etc) CPVLab tracking software 1. Do a little bit of research in your niche. Find out the pain points that hurt your audience. Review CB sales letters and videos in your niche. Manually Compile a list of keywords that, based on your research, you think would be relevant in your niche. To Make things easier, you can ask the vendor(s) of your market if they can provide a list of keywords (you would approach them as an affiliate of course). 2. Plug these keywords in Google keywords planner Generate keywords from Google, save them in a text file. Delete keywords that you believe aren't relevant. This text file is named "phase 1 kw batch". I will have about 500 keywords in this batch. 3. Have CPVLab set up on your website CPVLab has the ability to tell you which keywords and query strings generated subscriptions and sales (won't tell you clickbank sales, but you can cross reference the time of transaction in clickbank and time of keyword click in CPVLab) 4. Run "phase 1 kw batch" in a $20/day campaign 5 days I bid high on the first day or two so I can quickly see which kws are getting attention from the market and which might be time wasters. Run all keywords on broad match. Gather data. kill keywords with low ctr. Make a new text file and put keywords that generate sales/subscriptions in this new text file. This text file is called "phase 2 kw batch". If you followed step 3 (which you should), you will also discover the specific query that ultimately generated the sale or sub (that's why you will run all keywords on broad match). Example: The keyword "tasty fast diet" got clicked. I got a sale because of that click. Because I have CPVLab, I am able to learn that "cookie diet" was the search query that ultimately generated that sale. Now "cookie diet" is going in the "phase 2 kw batch" text file because it showed sale/sub potential. Make sure your campaign is set to run all day. You are collecting data. If you're adspend runs out at 2pm but you're in a market that mainly responds at night time, then you're assed out. Make sure you're data isn't handicapped. 5. While doing step 4, you will split test 6 ads in your $20/day campaign Your ads = 3 headlines and 2 body copies equal 6 ads in total. Should take 2-3 days to eliminate the bad eggs from the good. Discover you're winning ad. CPVLab will match the sales/sub with the ad (again, it won't tell you which ad generated clickbank sales, you'll have to figure that out manually or find an expert who can link CPVLab and Clickbank). Using a decent ad with decent ctr helps decrease cost and discover more keywords with sales/sub potential. 6. Run a new campaign. Use the winning ad in this campaign. Use the "phase 2 kw batch" keywords you generated from the the phase 1 campaign. This time, use phrase match bidding (and not broad match, which we did in the first campaign) for you're keywords. You might make exceptions. You might want to do exact match for some keywords based on their performance. There's no hard or fast rule here, you'll have to decide based on your analysis. Also, ultra long keywords I simply set to broad. Start getting sales. Start getting subs (you are collecting emails right? ). Keep scrubbing, refining, optimizing. Use the increase traffic to split test you're Landers. Split test your offers.