Anticipating Customer Needs In a Digital World
Whether it’s a weather forecast for mountaineers or a pregnancy prediction by Target’s coupon marketing machine, good predictions matter. Good predictions are particularly important for cross-selling and retention when customers expect the enterprise to know them – and the best way to demonstrate that you know your customers is to accurately anticipate their needs and intentions.
Anticipating customers’ needs and intentions was easier when customers came into branches and talked to their bankers about their lives. Today, bankers have millions of digital breadcrumbs to use to piece together a life story to try to develop the same insight. This level of analysis and prediction just isn’t possible with a spreadsheet, a pot of coffee, and a long night; machines have to do the work.
How it Works
Micronotes Predict consumes financial institutions’ marketing customer information file (MCIF) via a secure file transfer protocol (SFTP) each month. The last six months of MCIFs are joined with a Micronotes interview response file, where available, in order to create a time series that accurately represents a six-month snapshot of financial institution’s customers and can be processed using predictive algorithms. These predictive algorithms output propensity values between 0 and 1 indicating the probability that the customer will acquire a particular product in the next month. A file containing propensity scores across all customers, for each product, is automatically consumed by Micronotes to target interviews, and made available to the financial institution for targeting email, direct mail, and CRM based marketing. Scores are used to target high propensity customers for products and services since the higher the propensity score, the more likely a user is to acquire the product. Predict is able to consistently learn by using six months of user and interview response data to systematically reduce targeting errors.
Micronotes Predict™ cuts mortgage audiences to size to maximize revenue per customer interaction.
Micronotes Retain™ predicts which attractive customers are at risk of attrition and takes action to retain them.
System of Record for Cross-Selling
Micronotes is the only predictive marketing platform to output scores reflecting explicit customer input and is therefore the system of record for cross-selling products to the financial institution’s customers via all the marketing channels.
The value of a marketing campaign is equal to the value of good leads produced minus the cost of mistargeting. All else being equal, reducing mistargeting and missed opportunities improves campaign value and customer experience, simultaneously. Machine learning is the only way to systematically reduce mistargeting, commonly referred to as improving the f-measure, in data science vernacular. In other words, making fewer mistakes – both in calling-out who is in the market for a loan, and who is not in the market for a loan.
Moreover, we probabilistically score response and user data to project more potential customers to target, thereby solving the sparse data problem common in many data sets.
On the retention side, we use machine learning to predict which profitable customers are at risk of leaving to initiate account reviews and customer outreach to retain more customers.
On the risk side, we’re predicting delinquency to encourage the use of skip-a-pay products and to schedule meetings with a banker to work through financial rough spots.
Since Micronotes Predict is always learning directly from customer responses, and scores are downloadable for consumption by other marketing systems—email, direct mail, multichannel advertising systems, CRM marketing— Micronotes Predict is truly the system of record for cross-selling financial services.
Products and Services With Scores
With Micronotes Predict, we will provide scores for all of the services below, free of charge:
Certificate of Deposit
Online Bill Pay
Home Equity Loan/Line of Credit
Mobile Check Deposit
Money Market account
More sales, less marketing
Fewer loan losses
Downloadable scores to target all marketing