I think this method of measuring the performance of retargeting is wrong. When your visitors later see your ads and convert, the value may not all be attributab

I think this method of measuring the performance of retargeting is wrong. When your visitors later see your ads and convert, the value may not all be attributable to the ad. I would expect some visitors would have later converted even if they hadn't seen the retargeted ad. A better approach would be a controlled experiment. Divide your non-converting visitors into two groups, and run ads for only one. Then compare the results.