How we link survey respondents to your campaign delivery
Some of your Brand Lift results can be broken down using campaign delivery data. For example, how many times a respondent was exposed to your campaign or where those exposures occurred.
To make this possible, Happydemics links campaign delivery data with individual survey respondents.
How does it work?
When your campaign is measured with a retargeting pixel, Happydemics receives an exposed pool containing the impressions delivered during your campaign and the information associated with each impression.
Separately, we collect responses to your Brand Lift survey. We then match these two data sources using the respondent's IP address.
When a respondent's IP address is found in the exposed pool, we can associate the delivery information available for that IP with their survey response.
This allows us to identify, for each matched respondent:
Exposure frequency: how many times their IP appears in the exposed pool
Exposure location: where those impressions were delivered, based on the location information passed by the DSP
The matching is performed at respondent level. This means each respondent has their own exposure profile rather than being assigned an average campaign-level value.
What does this data represent?
This information comes directly from your campaign's delivery data.
It is not declarative: respondents are not asked in the survey how many times they saw the campaign or where they were when they saw it.
Instead:
Exposure frequency is based on the number of recorded impressions associated with the respondent's IP.
Exposure location is based on the geographic information passed by the DSP at impression time.
This delivery information complements our Brand Lift methodology, which continues to use ad recall to distinguish respondents who remember the campaign from the control group. Exposure data can be used as an additional source of analysis when available.
How do we ensure reliable breakdowns?
We apply several rules before displaying delivery-based results.
IP addresses appearing more than 100 times are excluded
If an IP address appears more than 100 times in the exposed pool, it is unlikely to represent the activity of a single individual.
For example, it may correspond to:
an office or shared network
a public Wi-Fi or hotspot
a mobile network
a VPN
In these situations, neither the exposure frequency nor the exposure location can reliably be attributed to one person. Respondents associated with these IPs are therefore excluded from delivery-based breakdowns but are still included in the overall results of your Brand Lift study.
Respondents without usable delivery data are excluded from the relevant view
A respondent is only included when the information required for that breakdown is available. For example, if we can identify a respondent's impressions but no geographic information was passed for those impressions, the respondent can be included in the frequency analysis but not in the location analysis.
Why is there a 50 exposed ad recallers minimum?
A group must contain at least 50 exposed ad recallers to appear in a delivery breakdown.
This rule applies to both:
exposure frequency groups
geographic areas
Below this threshold, results can become too unstable to interpret reliably. As a result, the groups displayed in your report are created dynamically for each campaign. We show the most detailed breakdown supported by the available sample rather than applying the same groups to every campaign.
Availability
Delivery breakdowns are available when:
your campaign is measured using a retargeting pixel
your Brand Lift study is completed
👋 If you don't find your answer, feel free to reach out to support@happydemics.com for further consideration.
Last updated