Actual CPA for retargeting – One example of how to think about it
Posted: Wed Dec 11, 2024 5:43 am
Among users who converted through retargeting, there is a possibility that there are users who would have converted even without retargeting . Of course, the degree of this will vary depending on the design of the campaign, but retargeting has a higher probability of this happening than other measures.
I conducted an experiment that may provide some insight into the extent of the so-called “ incremental lift ” in retargeting , so I’d like to share it with you.
>>Download the document
set (service documents and case studies) to learn about Ollie's
Table of Contents [ Hide ]
Are you prioritizing short-term gains and spending too much of your budget on retargeting?
Are you advertising too much to users who would convert without retargeting?
This led to a discussion:
Therefore, we decided to think as much as possible about the "CPA of CVs that could only be acquired through retargeting."
What we did
I have tried the following methods:
Remove the Criteo "Product Detail Page" tag for a period of time.
For the excluded period, look at the difference in CVR between the "Home Page" and "Product Detail Page"
However, with this method,
It doesn’t take into account the users who visited each page (so you can’t compare them independently)
Because the CVR difference is measured over time, the impact of time cannot be taken into account
Although this was a major problem, we approached the issue with the policy of "if there are any trends, we will use them as a reference for our allocation plan."
result
I conducted an experiment that may provide some insight into the extent of the so-called “ incremental lift ” in retargeting , so I’d like to share it with you.
>>Download the document
set (service documents and case studies) to learn about Ollie'sTable of Contents [ Hide ]
Are you prioritizing short-term gains and spending too much of your budget on retargeting?
Are you advertising too much to users who would convert without retargeting?
This led to a discussion:
Therefore, we decided to think as much as possible about the "CPA of CVs that could only be acquired through retargeting."
What we did
I have tried the following methods:
Remove the Criteo "Product Detail Page" tag for a period of time.
For the excluded period, look at the difference in CVR between the "Home Page" and "Product Detail Page"
However, with this method,
It doesn’t take into account the users who visited each page (so you can’t compare them independently)
Because the CVR difference is measured over time, the impact of time cannot be taken into account
Although this was a major problem, we approached the issue with the policy of "if there are any trends, we will use them as a reference for our allocation plan."
result