Brand Lift uses data from surveys to measure how your ads influence people. For all TrueView in-stream and bumper ad campaigns, you can set up Brand Lift to show surveys to people about your product or brand.
In order to accurately detect your lift, a certain number of survey responses is required.
This article is about how Google Ads will show you an estimate of your Brand Lift for a given metric using your response count, even before a number is reported.
How Google Ads measures your Brand Lift
Google Ads can narrow down how much lift your Brand Lift metric generated based on the difference between the amount of positive survey responses received from people who have seen your ads and those who were withheld from seeing your ads. Generally, more responses are required in order to accurately detect smaller amounts of absolute lift. Before your lift is detected, you will be able to see an estimation of it based on your response count.
When to expect detectable lift
View the following guidelines about the number of responses required to detect your lift:
- For high-performing campaigns, you can expect to detect lift once you receive about 2,000 responses per lift metric.
- At the recommended budget minimum, you can expect to detect lift once you receive 4,100 responses per lift metric.
- If your campaign has not shown any lift after reaching 16,800 responses per metric, you may not be able to detect your lift.
Required total responses for measuring Brand Lift
In order to measure Brand Lift accurately at various levels, the total response count must be within a certain range. The smaller the absolute lift, the more survey responses are required to ensure accuracy. The table below shows the required total response, given a detectable absolute lift:
|Detectable absolute lift||Required total response count|
|> 4%||1,200 ~ 2,800|
|3%||2,800 ~ 5,000|
|2%||5,000 ~ 11,000|
|1.5%||11,000 ~ 20,000|
|1%||20,000 ~ 45,000|
|0.5%||45,000 ~ 180,000|
|< 0.5%||> 180,000|
For detectable absolute lift percentages not mentioned in the chart, you may need to estimate to find the total required response count.
Let’s say that you have .75% absolute lift and want to know the number of responses that you need to detect the absolute lift. 45,000 responses would be more than what you need (since the minimum requirement to detect .5% absolute lift is 45,000 responses), while 20,000 responses wouldn’t be enough (since the minimum requirement to detect 1% absolute lift is 20,000 responses).
Since .75% is halfway between 1% and .5%, you would need roughly between 20,000 and 45,000 responses to get .75% detectable absolute lift (or about 33,000 survey responses).
If your Brand Lift metric’s absolute lift approaches 0, more survey responses are required to accurately measure absolute lift. This is because if there's only a small difference between the responses of people who have seen your ads and those of people who have not seen your ads, more responses are required to determine exactly what difference there is.