Add Known Attributes
When analyzing your onboarded deterministic data against probabilistic data in Resonate, we recommend making some edits to the joined data so that the platform uses a common definition (or representation) of your segments. It’s important to remember that deterministic data and probabilistic data will never match 100% but following these best practices will help to reduce discrepancies.
When onboarding segments of your deterministic data into Resonate, we recommend as a best practice that you add the corresponding known attributes from our Resonate Elements attribute values into your audience definition with those onboarded segments before analyzing.
Let’s walk through an example:
Let’s say you’re onboarding a file consisting of affluent male buyers. This is deterministic purchase data coming from your CRM system. You have information that tells you these CRM records are male and affluent (I.e. known attributes). You will need to add these known attributes as part of your onboarded segments for analysis.
For example: In Resonate’s Ignite Platform, in Segmentation Center, add your onboarded segment via Imported Data. Then find Male and Household Income $150K+ in Resonate Elements and add those to your audience definition. The resulting audience will resemble this:
Then you can move on to analyze and view Resonate insights.
Why do we recommend doing this?
When onboarding deterministic data into a probabilistic predictive insights platform, such as Resonate, you may see some data skews as a result of matching the two data sets and differences in how and when your deterministic data is captured vs Resonate’s predictive data. This can result in seeing insights or audience distributions that differ from your internal data file. Seeing some data skews is to be expected, and this challenge is not unique to Resonate.
By adding known attributes to your onboarded data, you’re isolating your known attributes, so that you can use the power of Resonate to discover new insights.