AB Treatments Tool Icon

AB Treatments Tool

Version:
Current
Last modified: December 27, 2019

The AB Treatments tool assists you in determining which group is the best fit for AB testing. This is appropriate in cases where the test will be carried out on a set of test units from the same group, based on user specified selection measures (between 2 and 5 measures can be selected). For a number of common test situations (particularly those that make use of broadcast media such as radio or television), a "group", such as a designated market area (DMA), needs to be selected to implement the test, with the units (e.g., stores, customers) in the group being the treatment units in the test. The tool returns a report with a rank ordered list of groups based on how well the average unit in the group meets the selected criteria based on a Euclidean distance measure.

This tool uses the R tool. Go to Options > Download Predictive Tools and sign in to the Alteryx Downloads and Licenses portal to install R and the packages used by the R Tool. See Download and Use Predictive Tools.

Connect an input

A Designer data stream with a unit identifier, a group identifier, and two or more measures to compare different groups to the overall average for those measures.

Tool configuration

Use the Configuration tab to set the treatment controls.

Minimum number of treatment units for a group: The minimum number of units that need to be in a group in order to have that group considered (a filter). A rule of thumb is to have ten or more units receive the same treatment in order to provide a reasonable level of power for a test.

Use the Data Input tab to configure the incoming data stream for treatment.

  1. Choose Field: Object ID (String): The field with the unit identifier (e.g., store, customer). This field must be some sort of character based type such as String or V-String.
  2. Choose Field: Group (String): The field with the group identifier (e.g., geographic region, segment). This field must be some sort of character based type such as String or V-String.
  3. Choose Field: Variable 1 (Double): The first numeric field used to compare the average unit in a group to the overall average of all units.
  4. Choose Field: Variable 2 (Double): The second numeric field used to compare the average unit in a group to the overall average of all units.
  5. Choose Field: Variable 3 (Optional) (Double): The third numeric field used to compare the average unit in a group to the overall average of all units. Specifying this field is optional.
  6. Choose Field: Variable 4 (Optional) (Double): The fourth numeric field used to compare the average unit in a group to the overall average of all units. Specifying this field is optional.
  7. Choose Field: Variable 5 (Optional) (Double): The fifth numeric field used to compare the average unit in a group to the overall average of all units. Specifying this field is optional.

View the output

  • D anchor: A data stream that contains all the information contained in the report for all eligible groups (those that meet the minimum number of units criteria), along with the average, minimum, and maximum index value for each measure used to assess a group relative to the overall average. These additional measures are index numbers, so a group that has a value of one on a measure has an average for that group that exactly matches the overall average for that measure. If the index value is less than one, then, on average, that group has units that have an average value on that measure which is below the overall average, while an index over one indicates that the reverse is true. The minimum and maximum value for the index number of each measure is provided for each group, so you can understand the variability in each measure for members in each group.
  • R anchor: A set of report snippets that includes the group identifier, the number of objects (units) in the group, and a score for each group that ranges from 1 (the closest to the overall average) to 0 (the furthest from the overall average) for the best ten groups based on the score measure.
     
Was This Helpful?

Need something else? Visit the Alteryx Community or contact support.