Hello. I have recently made a major change in the attribution model from the last click to time-decay as I did the same change on another account which resulted in better performance. I tried it again on one of the other accounts that I work with and it made things worse in terms of performance. I wonder if that specific change couldn’t be the cause of hindered output of my smart shopping campaign? The amount of conversions per selected time frame is the same. I am always looking for ways how to improve my campaigns so I thought default settings aren’t the best and gave it a shot but now I can’t properly evaluate the sudden decrease in “revenue” from PLA as it was almost 1/2 higher last month. Maybe it might be a product of other external factors as well? Could someone elaborate? I will be most happy with any information.
When you switch to a model that’s backwards-looking (anything other than last click) it won’t track retroactive clicks as conversions. In a way, you’re starting from scratch.
If you typically need 1-week to attribute all your conversions you’ll have 1-week of down performance. You can get an idea of your typical attribution from those reports in Google Ads or Analytics.
Also, importantly, you need to use a longer look-back period to evaluate performance. If it takes a week for most of your conversions to be realized you shouldn’t be looking at the most recent 7 days of conversions.