Trellance - Building a Next Best Product AI Model for Credit Unions

AveryTrellance
AveryTrellance Registered Posts: 1

Team members:

Avery Swiontek, Product Manager - Predictive Analytics:

  • Rewati Sinha
  • Amitoj Lotey
  • Prachi Dave
  • Savan Hirpara
  • Vishnu Thanki
  • Haritha Ginjupalli
  • Sravya Vangala
  • Sumi Balarama
  • Sangmeshwar Swami
  • Akhilesh Prajapati
  • Vikram Bhattraj
  • Mwafaq Abu-Shanab

Country: United States

Organization: Trellance

Trellance is the leading provider of data analytics and business intelligence solutions, professional services and consulting for credit unions. Our solutions and services, together with the patented common data model, are used by credit unions to find actionable insights, improve member experience and achieve portfolio growth.

Awards Categories:

  • Best Moonshot Use Case
  • Best ROI Story

Business Challenge:

MSU Federal Credit Union, a $7.3 billion asset credit union with over 340,000 members, was looking to provide targeted offers to members to increase investment in certificates of deposit in order to maintain liquidity during this time of rising rates.

To achieve this, they turned to Trellance’s Predictive Analytics team for a solution. The Analytics team recommended Trellance’s Next Best Product AI to identify members most likely to respond to marketing efforts around certificates.

Next Best Product AI runs an algorithmic analysis on current members, reviewing the products they currently have, how often they use them, and other personal data on file to identify the products they are most likely to respond positively to in a marketing campaign.


image

Business Solution:

Trellance ran a Next Best product model for MSUFCU. We utilized transactional data, product holding, account balances, etc. for the members and ran multiple models using Dataiku. We then picked the next best product for their membership and provided the results to them using Dataiku visualizations.

Once a group was identified, MSUFCU decided to run A/B testing in order to assess the effectiveness of the Next Best Product AI. The Trellance predictive model selected a targeted group of approximately 4,000 members to be included in the marketing campaign, Group A. The credit union used their normal selection process to randomly select a second group of 4,000 to be included in a separate marketing campaign, Group B.

Over a 30-day period, members in both groups received emails and a direct mailer promoting the certificate of deposit program. For those members included in either campaign, the marketing they received within MSUFCU’s mobile app was also focused on certificates of deposit. Content was the same for both groups and was sent out simultaneously.

Day-to-day Change:

MSUFCU was beyond happy with these results. Tim Williams, an Analyst on MSUFCU’s Business Intelligence Team and who uses analytics to help marketing better target members for campaigns, said “The Predictive model performed way better than a random group, and I think above average on a targeted group.” He indicated that the MSUFCU would continue to use Next Best Product AI to look for opportunities to increase auto loans and credit card usage among their members. “We are definitely looking to move forward and continue to integrate this with our marketing programs.”

Business Area Enhanced: Marketing/Sales/Customer Relationship Management

Use Case Stage: Proof of Concept

Value Generated:

The group selected by Next Best Product AI had an average account balance of $90,000 and opened 160 CDs with a total revenue of $7.4 million. The group selected by MSUFCU had an average account balance of $7,000 and opened a total of 53 CDs with a total revenue of $666,000.


image

Value Brought by Dataiku:

Using Dataiku for predicting the next best product at a credit union can bring significant value by leveraging advanced analytics and machine learning techniques.

By analyzing historical member data, transaction patterns, and financial behavior, Dataiku can identify relevant trends and patterns to predict which product a member is most likely to be interested in. This enables the credit union to personalize their offerings, leading to higher member satisfaction and retention rates.

Moreover, the predictive insights help the credit union proactively target potential customers with tailored promotions, thereby increasing the cross-selling and up-selling opportunities and ultimately driving revenue growth.

Value Type:

  • Improve customer/employee satisfaction
  • Increase revenue
  • Reduce cost
  • Save time

Value Range: Millions of $

Comments

Setup Info
    Tags
      Help me…