One of the most over-looked and common mistakes in ppc, or any real-time bidding environment and ESPECIALLY on mobile ad networks, the thing I see the most in the testing phases of campaigns is improper structuring of non-converting keywords. Meaning to “prune” your non-converters each day/end of your testing phase. For example, here’s a small mobile campaign we are testing out right now:
Now a few you can tell I didn’t get traffic on. However, my “green”s ROI is easy enough to tell me quickly right to cut out my “non-converters. In this instance, I’m going to look at my spend and determine which keywords are worth cutting. The ones I broke even I, I’ll keep testing these.
Now generally, I’ll run 3 ads at a time, pick the highest CTR ad from each group. Remake group, all 3 ads against each other. At this time, I setup two new ads to split against each other and compare CTR’s and CR’s against my top 3 performers. All the while, you have to run this keyword report for each campaign you got up to drill down to the profits.
We still profited from this small test run of about $189.10. Now tomorrow we will run only our converting ads that produced positive ROI’s plus the ones that broke even. The second day, I typically cut “break evens” off entirely. (Just the way I am).