Cross-Sell (XS) engine

The Cross-Sell (XS) engine generates profiles of core customers’ expected spending, and matches non-core customers to those profiles. The engine also generates summary metrics for all customers based on certain settings and parameters.

For more information on profile matches, read Outputs.
For more information on summary metrics, read Customer Metrics.

Configuration

To configure this engine, adjust the following components.

Engine Parameters

To view the parameter settings for this engine, go to the Engine Parameters worksheet. Depending on your user permissions, the worksheet will resemble the following image.

The XS engine uses the following parameters associated with an Action Type Id of XS:

• Cluster_Periods—Specifies the number of historical aggregation periods used to calculate this action type’s Cluster time period.
  To view an action type’s Cluster time period, go to the Aggregated Time Periods worksheet. To update an action type’s Cluster time period, run the Prep Data and Update Time Periods worksheet action.
• Active_PG_Volume_Share_Min—Minimum threshold for the engine to consider a product group active and include it in the clustering process. The engine defines the threshold as the ratio of a product group’s total Cluster time period volume (across all customers) to the total Cluster time period volume (across all customers and products).
• Min_Cluster_Overlap—Minimum overlap for the engine to map a customer to a profile. The engine applies this threshold to both core and non-core customers. The engine measures a customer’s total overlap between their actual Cluster time period volume share for each product group, and the profile’s median volume share for the same product group.
• Min_Core_Customers_Per_Cluster—Minimum threshold for the engine to use a customer profile. The engine defines the threshold as a number of core customers assigned to a profile as a result of the clustering process.
• Cluster_Assignment_RBO_Persistence—This parameter applies to the Rank Biased Overlap calculation in the XS engine’s clustering science. It indicates the slope of the decline in weight for each successive ranked list used in the calculation.
  The default value for this parameter rarely needs to be changed.
• Cluster_Assignment_RBO_Max_Depth—This parameter applies to the Rank Biased Overlap calculation. It indicates the total number of ranked product group lists that the engine compared between an unmatched customer and a candidate profile.
  The default value for this parameter rarely needs to be changed.
• Cluster_Assignment_RBO_Weight—This parameter applies to the Rank Biased Overlap calculation. It indicates the relative weight of the RBO metric when the engine calculates the overall score for a customer’s candidate profile match.
  The default value for this parameter rarely needs to be changed.
• Cluster_Assignment_Overlap_Weight—Relative weight of the overlap metric. The engine uses this when calculating the overall score for a customer’s candidate profile match.
  The default value for this parameter rarely needs to be changed.
• Cluster_Assignment_Absolute_Distance_Weight—Relative weight of the absolute distance metric. The engine uses this when calculating the overall score for a customer’s candidate profile match.
  The default value for this parameter rarely needs to be changed.
• Cluster_Assignment_Square_Distance_Weight—Relative weight of the square distance metric. The engine uses this when calculating the overall score for a customer’s candidate profile match.
  The default value for this parameter rarely needs to be changed.

Volume Metric

To view the volume metric for this engine, go to the Sales IQ > Engine Config > Sales IQ Transactions Aggregated worksheet. Depending on your user permissions, the worksheet will resemble the following image.

The XS Volume Metric calculated column defines the volume metric used in the XS engine. The following default calculation uses revenue normalized with any applicable currency conversions. Update this calculation as needed.

(CASE WHEN ([Currency Conversion Rate] > 0)
THEN [Currency Conversion Rate]
ELSE 1
END)
* [Revenue]

Product Group

To view how this engine groups products, go to the Sales IQ Products worksheet and navigate to the XS Product Group calculated column. Depending on your user permissions, the worksheet will resemble the following image.

Edit the calculated column to use any available product attributes that best define the product grouping level for the XS engine. This engine excludes products configured with a null value from engine outputs.

Customer Cluster Segment

To view how this engine defines customer segmentation for the clustering and matching processes, go to the Sales IQ Customers worksheet and navigate to the Cluster Segment calculated column. Depending on your user permissions, the worksheet will resemble the following image.

When a cell under the Cluster Segment column has a value that isn’t null, the XS engine only compares the corresponding customer against other customers of the same Cluster Segment value. For all customers with a null value under this column, the engine assigns them to a single “NULL” Cluster Segment during the clustering and matching processes.

Edit the calculated column to use any available customer attributes that best define customer segmentation for the XS engine. If you do not want to use any customer segmentation in this engine, define the column with a single default value for all customers, or leave the column’s definition blank.

Outputs

To view the results of running the XS engine, go to the XS Engine Output worksheet. Depending on your user permissions, the worksheet will resemble the following image.

This engine generates results for every active Customer Id + active Product Group combination. This worksheet contains the following metric fields:

• Cluster Median Ratio—Median volume share for the corresponding product group across all core customers in the corresponding profile.
• Customer Ratio—Volume share for the corresponding product group against the corresponding customer’s Total Cluster Period Volume.
• Ratio Gap—Ratio of the difference between the Customer Ratio and the Cluster Median Ratio, to the Cluster Median Ratio. Negative values identify a Customer Id + Product Group combination where the Customer Ratio exceeds the Cluster Median Ratio.

• Customer Overlap—The corresponding customer’s total overlap with their assigned profile. The engine measures this by first observing the smaller value between the corresponding product group’s Cluster Median Ratio and Customer Ratio. Then the engine takes the sum of the smaller ratio across all of the customer’s active product groups.
• Customer Potential—Total volume potential of the customer. For more information, read Customer metrics.
• Action Impact—Annualized impact of getting volume back to the Baseline Avg. This is a calculated column defined as the difference between the Cluster Median Ratio and the Customer Ratio, multiplied by the Customer Potential. Modify this column as needed to create the desired Action Impact metric.
• Current—The Customer Ratio. Modify this calculated column as needed to capture the best measure of the “Current” state for any actions created.
• Target—The Cluster Median Ratio. Modify this calculated column as needed to capture the best measure of the “Target” state for any actions created.