AI-Powered Churn Prediction Saves $150,000 in AUM

AI-Powered Churn Prediction Saves $150,000 in AUM

Executive Summary

Precision Financial Group, a growing RIA firm, faced challenges in proactively identifying clients at high risk of attrition, leading to unexpected AUM losses. To combat this, they implemented an AI-powered churn prediction tool from Golden Door Asset, integrating it with their existing CRM. The AI platform analyzed client account activity, market conditions, and demographic data to identify at-risk clients. This proactive approach allowed Precision Financial Group to retain $150,000 in AUM within the first six months of implementation.

The Challenge

Precision Financial Group had experienced significant growth over the past few years, expanding its client base by 20% annually. However, this growth was being partially offset by unexpected client attrition. Their previous method of identifying at-risk clients relied on lagging indicators, such as observing a client's decreasing account balance or, worse, only noticing churn after a client had already terminated their account. This reactive approach proved inadequate in today's competitive landscape.

Specifically, they were losing an average of 1.5% of their AUM each quarter due to client attrition. One particularly painful example was the loss of a high-net-worth client with a $500,000 portfolio. The client, a retiree, had become increasingly anxious about market volatility but hadn't explicitly voiced these concerns to their advisor. By the time Precision Financial Group realized the client's distress, the client had already transferred their assets to a competitor promising "safer" investments.

Furthermore, Precision Financial Group's existing CRM system lacked the sophisticated analytical capabilities needed to identify subtle patterns indicating churn risk. Advisors relied heavily on intuition and personal relationships, which proved insufficient in managing a rapidly expanding client base. They estimated that for every 10 clients they lost, they spent approximately $2,000 in marketing and acquisition costs to replace them, further impacting their bottom line. The lack of proactive churn prediction was directly costing the firm an estimated $30,000 annually in lost AUM and increased acquisition expenses. They needed a way to identify at-risk clients before they initiated the account termination process.

The Approach

Lisa Tanaka, Head of Client Relations at Precision Financial Group, spearheaded the initiative to implement an AI-powered churn prediction solution. Her strategic thinking centered on shifting from a reactive to a proactive client retention strategy. The decision framework involved several key steps:

1. Needs Assessment: Lisa conducted a thorough assessment of the firm's current client retention processes, identifying key weaknesses and areas for improvement. She interviewed advisors, client service representatives, and management to gather diverse perspectives on the challenges of client attrition.

2. Solution Evaluation: Lisa researched various churn prediction tools available in the market, evaluating them based on factors such as accuracy, integration capabilities, cost, and ease of use. Golden Door Asset's AI-powered platform stood out due to its ability to integrate seamlessly with their existing CRM and its demonstrated track record of success with other RIA firms.

3. Data Integration Strategy: A crucial part of the approach was developing a robust data integration strategy. This involved identifying the relevant data points within their CRM (e.g., account activity, client demographics, communication history) and mapping them to the AI platform's data input requirements.

4. Algorithm Customization: Recognizing that each client is unique, Lisa worked with Golden Door Asset's data scientists to customize the AI algorithms to reflect Precision Financial Group's specific client base and investment strategies. This involved training the AI model on historical data to identify patterns and correlations associated with client churn.

5. Training & Implementation: Lisa led the training and implementation process, ensuring that all advisors and client service representatives understood how to use the AI tool and incorporate its insights into their daily workflows. This included developing specific protocols for reaching out to at-risk clients and addressing their concerns.

6. Monitoring & Optimization: The final step was to establish a system for monitoring the AI tool's performance and continuously optimizing its algorithms. This involved tracking the accuracy of churn predictions and making adjustments as needed based on new data and insights.

Technical Implementation

The technical implementation of the AI-powered churn prediction tool involved several key steps:

1. CRM Integration: Golden Door Asset's platform was integrated with Precision Financial Group's existing Salesforce CRM using a secure API connection. This allowed for real-time data synchronization between the two systems. The data extracted from Salesforce included:

   • Account Activity: Transaction history (deposits, withdrawals, trades), account balances, and portfolio performance metrics.
   • Client Demographics: Age, income, net worth, location, and relationship tenure.
   • Communication History: Email exchanges, phone calls, and meeting notes.
   • Risk Tolerance: As documented in the client's investment policy statement (IPS)

2. Data Processing & Feature Engineering: The raw data extracted from Salesforce was processed and transformed into a format suitable for AI analysis. This involved cleaning the data, handling missing values, and creating new features that could potentially be predictive of churn. Examples of newly engineered features included:

   • Volatility Score: A measure of the portfolio's historical volatility based on standard deviation of returns.
   • Withdrawal Rate: The percentage of assets withdrawn by the client over a given period (e.g., annually).
   • Communication Frequency: The number of interactions (emails, calls, meetings) with the client over a given period.

3. AI Algorithm Development: Golden Door Asset's data scientists developed custom algorithms to identify churn patterns. These algorithms used a combination of machine learning techniques, including:

   • Logistic Regression: To predict the probability of churn based on the input features.
   • Random Forest: To identify the most important features influencing churn.
   • Support Vector Machines (SVM): For non-linear classification of clients at risk.

   The algorithms were trained on historical data from Precision Financial Group's client base, using a training/testing split of 80/20. The model's performance was evaluated based on metrics such as accuracy, precision, recall, and F1-score.

4. Automated Alert System: An automated alert system was implemented to notify advisors when a client was identified as being at high risk of churn. The alerts were triggered based on a pre-defined churn probability threshold. Advisors received the alerts via email and through a dashboard within the CRM system. The alerts included:

   • Client Name: The name of the at-risk client.
   • Churn Probability: The estimated probability of churn within the next 90 days.
   • Key Risk Factors: The top factors contributing to the churn risk (e.g., high withdrawal rate, increased portfolio volatility, decreased communication frequency).
   • Recommended Actions: Suggested actions for the advisor to take, such as scheduling a meeting with the client or adjusting their investment strategy.

Results & ROI

Within the first six months of implementing the AI-powered churn prediction tool, Precision Financial Group achieved significant improvements in client retention:

• AUM Retention: The firm retained $150,000 in AUM that would have otherwise been lost due to client attrition. This was a direct result of proactively reaching out to at-risk clients and addressing their concerns.
• Churn Rate Reduction: The quarterly churn rate decreased from 1.5% to 1.0%, representing a 33% reduction in client attrition.
• Improved Client Satisfaction: Client satisfaction scores, as measured by Net Promoter Score (NPS), increased by 15%, indicating that clients felt more valued and supported.
• Increased Advisor Efficiency: Advisors spent less time reacting to client attrition and more time proactively engaging with clients, leading to increased efficiency and productivity. Time spent onboarding new clients also decreased by 10%.
• Reduced Acquisition Costs: With improved client retention, Precision Financial Group reduced its marketing and acquisition costs by $10,000 per quarter.
• Average predicted churn probability accuracy: achieved an average prediction accuracy of 82% over the testing period.

Specifically, the AI tool identified a client with a $200,000 portfolio who was considering transferring their assets to a competitor due to concerns about rising interest rates. The advisor, alerted by the AI system, proactively reached out to the client, explained the firm's strategies for managing interest rate risk, and reassured the client about the long-term prospects of their portfolio. As a result, the client decided to stay with Precision Financial Group, saving the firm $200,000 in AUM and associated revenue.

Key Takeaways

1. Proactive Retention is Crucial: Shifting from a reactive to a proactive client retention strategy is essential for RIAs in today's competitive landscape.
2. Data is Your Best Asset: Leveraging client data to identify churn patterns can provide valuable insights and enable advisors to take timely action.
3. AI Can Enhance Human Connection: AI-powered tools can augment, not replace, the role of advisors by providing them with the information they need to build stronger relationships with clients.
4. Integration is Key: Seamless integration with existing CRM systems is crucial for maximizing the value of churn prediction tools.
5. Continuous Monitoring & Optimization: Regularly monitor the performance of your churn prediction tools and optimize their algorithms to ensure ongoing accuracy and effectiveness.

About Golden Door Asset

Golden Door Asset builds AI-powered intelligence tools for RIAs. Our platform helps advisors proactively identify and address client needs, optimize portfolio performance, and enhance client engagement. Visit our tools to see how we can help your practice.