How a Growing Bank Leveraged Machine Learning to Increase Sales Pipeline Predictability

Our Goal

Define the Ideal Customer

First, we needed to define exactly who the stakeholders believed was their ideal customer and determine aligned attributes they identify with their target audience. This insight would allow us to select high-value use cases for the machine learning models.  

Our Solution

• Sat down with stakeholders throughout each department of the organization to discuss characteristics of their ideal customer. 
• Brought stakeholders together in a workshop to get everyone on the same page as to how they would identify their ideal customer
• Selected high value use cases

Our Goal

Understand their Data

Machine learning initiatives are only as successful as the quality of the data. We needed high quality, valuable data that would support the questions we were asking and give us accurate, reliable predictions we could use. So, before we developed the models, we needed a solid foundation in place. 

Our Solution

• Identified all available data and performed an analysis to assess data quality and completeness to support the defined objectives. 
• Identified where to find the best sources for data
• Analyzed information to determine where data needs to be improved or where gaps are located

Our Goal

Develop a Machine Learning Model

With characteristics identified and clean data, we developed a model that leverages the significant characteristics for use against prospects or existing customers. During this step, we allowed customer data to speak for itself. 

Our Solution

• Identified data elements that should be input for the machine learning model based on the data profiling
• Provisioned a cloud environment and developed data ingestion
• Defined and developed machine learning predictive models that supported the defined use cases
• Executed the model against real data and assimilated the output to graphically show customer segmentation

Our Goal

Finalize the Ideal Customer Definition

We used stakeholder inputs, data profiling outputs, and machine learning to let data and actual outcomes influence the definition of the ideal customer.  Our team explained what the model said and the characteristics of the target customer and how that compared to the stakeholder's thoughts. Using our information, we could create criteria or attributes and the bank could use those to build customer personas. 

Our Solution

• Took data from the model to share insight into the true characteristics of their ideal customers. 
• Built machine learning models that would score customer and prospect lists against the customer model
• Operationalized the model to use in marketing campaigns