What Machine Learning Taught Us About Deposit Marketing: Part II

Deposit Marketing

Earlier this week we highlighted the lessons that machine learning taught us about the Unified Deposit Formula (HERE). Embodied in the Unified Deposit Formula is a marketing and amplification equation. In this article, we expand on the lessons we learned from artificial intelligence when it comes to deposit gathering and explore how we can better use the Formula to optimize deposit gathering.


The Unified Deposit Formula


To recap, the Unified Deposit Formula brings various disciplines of the bank together to help explain and then optimize deposit costs. It takes a bank’s asset growth to determine the deposit needs and then looks at the value sensitivities of the service base. It is an acknowledgment that a bank’s ability to raise lower cost deposits is also based on the bank’s marketing reach and prowess. The stronger the marketing effort, the lower the deposit cost.


Also inherent in the model is the concept that deposit marketing is more than just price and promotions and that a banks customer base, products, and employees all contribute to determining the value sensitivities of any given deposit raising effort. The Formula can be seen below:


Unified Deposit Pricing Formula


Marketing & Amplification


For the sake of this article, we want to focus on the details of the Marketing & Amplification calculation. This part of the Unified Deposit Formula equation is a measure of the “Total Effective Addressable Market.” This is the dollar size of the market that you can effectively draw from. It includes the portion of the deposits that are not kept at your bank by your current customers plus the service areas net of your existing customers.


This portion of the Formula looks like this:


The Addressable Market Formula


Available Wealth Dollars


For this calculation, you take the number of potential customers that you are trying to market to and then multiply that number by their “Wealth $s” which is your assumption on the investable assets that are available for deposits. This number varies by geography and by customer set, but this data is usually readily available either through third-parties or by utilizing your bank’s in-house data set. Graphical detail can be found for the counties in your state HERE, and as a rule of thumb, we generally take 10% to 20% of the household net worth number.  


Finally, since you already have a portion of your customer’s wealth already, you want to deduct that from the equation. This is if you have 20% of your customer’s deposits, then you can further obtain 1-20% or 80% of the available deposits that are likely stored at other banks or investment firms.


Coverage Percentage


Once the total size of the market is known, then it is a question of your marketing effort. You have already captured the effectiveness of your message in the Elasticity Equation, so now you are trying to figure out if you can show your value proposition to that market.


For example, if you are sending a message to your commercial loan customers about a new $500 bonus offer for a new checking and savings account and 16% open the email and click to your website to learn more, then your message coverage is 16%.


Customer Discernable Wealth


To put this equation into action, we figure out the “Customer Discernable Wealth” first. If you have 3,000 commercial loan customers that do not have a deposit relationship with your bank (a 0% share of wallet), each with an $85,000 of investable funds and a 16% coverage area, then the size of your existing market would be approximately $41 million. 


Service Area Discernable Wealth


We then do the same calculation for your prospects with one additional factor. If your service area has a population of 150,000 target commercial prospects that are not your customers, each with investable funds of $45,000 and you can assume a marketing coverage of 5%, then your addressable new market is approximately $338 million.


However, another lesson from machine learning is to build in an adjustment factor for your brand. Because the Elasticity Equation data is from a test of known customers and focus groups, that number overestimates the response. Some banks call this a conversion factor, but we like to think of it of knowledge about our bank, or a “Brand Recognition Factor.”


While a machine learning algorithm can adjust for this Brand Recognition Factor, banks that want to do this calculation in Excel can use a simplified version. Since your customers already know your bank, they get a “1.” Prospect has a spectrum of knowledge and get a 0.15 factor if they never heard about your bank before your marketing message. The marketing message itself serves as an introduction which is the reason the factor is usually a positive non-integer.


If you have some brand recognition and you have marketed to the customer before so that the household can recall your name, then the factor is 0.50%. If your local bank has the brand recognition of what a national bank might have, then the Brand Recognition Factor is usually between 0.75 and 0.85.


The result of this equation is the effective size of the market scaled for your bank.  In our example above, our existing customer base could provide us with $41mm of deposits while our new market could provide us with $338mm, so we know that our total addressable market is $379mm.


The Value of Marketing


One material takeaway from this Formula is the power and quantification of marketing. By building your brand, you increase your Brand Factor making your bank more effective with its message. Increasing the coverage factor also increases the size of the market. This is why it pays to constantly be building your prospects while experimenting and quantifying your marketing efforts. If you can add another 10,000 prospects to your email database, increase your Brand Recognition Factor from 0.15 to 0.35 and increase your open rate by 10%, then your bank can pick up another $9mm in addressable deposits from just that one effort. Do that repeatedly and scale it, and a bank can essentially double its addressable deposit base in which to draw from.


Up Next


Next week, look for Part III, which will delve more into the concept of the Predicted Market Penetration at Optimized Offering where we will quantify why rate should be your deposit attribute of last resort. We will look at the effectiveness of various promotions and then optimize value to produce the lowest cost and the best performing deposit base that your bank can achieve.


Until then, use the Market & Amplification calculation, or at least the concepts above, to start expanding your bank’s brand, market, and marketing effectiveness rate that will result in lower cost deposits.