What does limited mean?
Attributes that are limited can only be combined in an audience with other Resonate Element attributes. You cannot combine limited attribute values with a Tag, Behavioral attribute, 1st party data attribute, or 3rd party segment attribute in an audience.
It also means we cannot predict insights for these attributes against audiences that contain any of these types of data: Tag, Behavioral, 1st Party Data, or 3rd Party Segments from LiveRamp. The reason for these limitations is due to the fact that we cannot generate a data science predictive model behind that attribute. Learn more below:
How can I use limited attributes? What restrictions apply?
Limited Attributes CAN:
- Be combined in an audience with Resonate Element attributes
- Be activated and shipped to DSPs
Limited Attributes CANNOT:
- Be combined in an audience with the following data types: Tag, Behavioral attribute, 1st party data attribute, or 3rd party segment attribute.
- Be seen as insights when analyzing an audience that contains any of the following data types: Tag, Behavioral, 1st party data, or 3rd party segment.
- We recommend against using limited attributes in audiences that you wish to use in Audience Crosstab, as limited attributes do not produce results when crossed with a tag.
How can I tell which attribute values are limited?
You can determine which attribute values are limited by looking for the limited "L" icon in the middle panel in Audience builder and in your Audience definition.
Why are attributes limited?
Attributes are limited when we cannot generate a data science predictive model behind that attribute. Attributes are limited for one of the below reasons:
- It's a third party appended attribute - we do not model third party attributes.
- It's an attribute value that does not have enough survey responses for our data science team to develop a predictive model.
- It's an attribute where there are too many single select answer options for that survey question. For example, individual age. The answer options are 18-100. Too many single select options prevent our data science team from being able to create a good model. See below for workarounds.
- It's an attribute where there are single select answers to a survey question, but one answer has *low sample. This makes our data science team unable to generate predictive models for answers to the entire survey question. As best we can, our research team tries to avoid this scenario, but sometimes it does occur.Work Arounds for Too Many Single Select Option Survey Questions
There are ways to work around attributes with too many single select options for survey questions.
- Individual Ages, ex, 18, 19, 20.
a. Age Groups (ex. 18-24) are not limited, use this instead - Employment Industry
a. Employment Status, Department, Role and Company size are not limited, use this instead
You'll also want to avoid using the above attribute values in audiences that you wish to use in Audience Crosstab, since they are limited. However, you can include these attribute values in an audience for media delivery.
*At Resonate, we have established guidelines to ensure the statistical reliability of our survey data. Our reporting threshold requires a minimum of 30 completes per response before publishing. Various factors can influence whether a particular response achieves sufficient sample size, including the rarity of the response and overall sample size. For responses that don't meet this threshold, we will typically combine it with an ‘other’ category if available. We track the incidence of low-performing sample attributes across six waves of our syndicated study. If after six waves the sample requirement is still not met, we'll remove the attribute from the study.
It's important to note that changes in consumer behavior or opinion shifts could result in more people selecting a previously underrepresented response, potentially allowing it to meet the threshold in the future. In cases where immediate insights are needed for low-sample responses, we can explore alternative options such as custom studies to gather targeted data on the topic. Our aim is to provide you with the most reliable and statistically sound data possible while offering flexibility to meet your research needs.
Comments
0 comments
Please sign in to leave a comment.