Skip to main content

Varicent ELT Help Center

Financial Advisor Turnover

Abstract

Predicting Financial Advisors (FAs) leaving helps you focus your retention efforts.

Predicting Financial Advisors (FAs) leaving helps you focus your retention efforts. While it may be tempting to focus on retaining the biggest books or the most churn-prone groups, this may not be the most effective strategy.

This blueprint provides an example with a historical FA data (Data) data set with FA demographic and performance information. This is a snapshot of data collected from a particular date in the past for each FA. The Turnover column specifies whether the FA churned at the end of your chosen observation period, for example, one year after the data for that FA is generated. This period is known as the prediction window and represents how far forward your Varicent ELT model's prediction is valid. You can change the columns in this example data set as needed. The only mandatory components for this blueprint are:

  • The Turnover column.

  • The Total Revenue column.

  • A categorical feature for plotting.

Tools and data flow

The Adapt columns (Adapt) tool is where you can rename columns to human-readable names and change their data types as needed. For this blueprint to work, Varicent ELT expects that the attrition column (indicates whether a FA left) is named Turnover, the total revenue column is named Total Trailing 12mon Revenue, and the column describing the assigned market size is named Market Potential Tier.

The orange tools are the 'brains' of the operation. The Classifier tool is where Varicent ELT determines which algorithm best performs on your data set and builds a classification model. The Expected Loss (Formula) tool calculates the value for retention strategies by multiplying the Churn Probability by the total revenue of a FA and produces Expected Loss, an estimate of how much revenue your organization would lose at the FA level.

The pink tools then Rank the data by Expected Loss, and assign a Priority to retain bin (Binning) label based on this ranking. By default, high priority is applied to the first 50 FAs, and medium priority to the next 100, but this can be changed in the Priority to retain bin (Binning) tool.

The data is then manipulated for ease of exploration. The Reshape for Scatter Plot (Reshape) tool pivots the priority label to help you visualize the total revenue with color-coded priority levels.

The Aggregation for identifying focus areas (Aggregate) tool also groups by Market Potential Tier and Priority labels, and calculates the average Expected Loss and the sum of Expected Loss in each of these cross-groups. By aggregating additional groups (such as geographic region and referrals-self found ratio), you can slice your data into granular groups to build retention strategies around.

How to use this blueprint

  1. Input your data into the pipe.

    1. In the Row viewer, select your historical snapshot data as the data source in the FA data (Data) tool.

    2. On the canvas toolbar, click Build.

      Varicent ELT will build a predictive model tailored to your data.

  2. Enter the latest FA data on all current FAs using the same columns (all Turnover columns can have the value of "No").

    This current data set is where you want to get insights and make predictions.