Subscribe
Share
Share
Embed
BlastPoint Deep Dive Podcast explores the intersection of data, AI, and customer insights. Each episode dives into how AI-driven solutions help businesses understand and engage their customers, improve satisfaction, and reduce churn. Join us for expert insights, real-world examples, and practical tips on leveraging AI to transform the customer experience.
Whether you're a business leader or a data enthusiast, this podcast will give you the tools to stay ahead in today’s competitive landscape.
If you want to learn more about what we do and offer, contact us today to schedule a demo at https://blastpoint.com/about/contact/.
How Predictive Intelligence is Transforming Deposit Growth in Banking
===
[00:00:00] Welcome back to the deep dive. Today we are cracking open a file that is, uh, keeping just about every bank executive up at night. And that's the challenge of deposit growth. Exactly. The challenge of growing deposits in modern finance, if you're watching the banking sector, you know this is the ultimate strategic lever.
It's everything, profitability, liquidity, lending power. But right now, that lever is, well, it feels stuck. You've got rising interest rates, making the competition for cash, just intense. And digital banking means customers can switch so easily. It's frictionless. So our mission today is to really analyze the solution banks are pivoting toward.
They're moving away from that. Traditional marketing, the spray and pray approach, right? And they're embracing what's called predictive intelligence. This tech is all about precision. Answering three key questions. Mm, who will grow their deposits, when is the right moment to reach them, and, uh, how to tailor that outreach.[00:01:00]
And we really have to start with the urgency here. The environment is just, it's brutal. We're looking at FDIC data showing that total domestic deposits rose a mere 0.4%, just 0.4% between June, 2023 and June, 2024. I mean, put that in perspective. That's the slowest annual growth rate the industry has seen in well over a decade that is basically stagnation.
And when banks face that, what do they usually do? They open the marketing floodgates. They just spend, they do, but the returns are just not there anymore. Our sources show retail banks are shelling out an average of get this about $561 to acquire a single new customer, $561. And that's supposed to buy loyalty.
It's supposed to. But it's not [00:02:00] working, is it? Clearly not. No. Despite these huge costs, churn is a massive headwind. We're seeing data pointing to a pretty frightening trend, which is between 20 and 28% of Gen Z and millennial customers are likely to switch their primary bank within just two years. Two years.
So you spend over $500 and they're gone before you even see a return on that investment. Exactly. Wow. So almost one in three young customers are just bailing out. Why? I mean, they signed up for the bonus. What's behind the quick exit when you ask them? The reasons are really consistent. Yeah. It centers on feeling, you know, undervalued and getting poor personalization.
They're tired of the generic offers, tired of the generic offers, the irrelevant emails. The message from them is loud and clear. You can't just spend more money. You have to spend smarter. You need intelligence that tells you which customers are actually worth that $561 and how to keep them, and crucially, how to keep them satisfied once they're in the door.
Okay, so let's unpack that, this shift from just spending more to being [00:03:00] predictive. What is predictive intelligence or PI at its core? Is it more than just a buzzword? It is at its core, PI is really the application of machine learning and, uh. Deep behavioral analytics to forecast future financial behavior.
So predicting the future in a way. Yeah. It uses algorithms to determine which customers are statistically most likely to increase their balances or maybe transfer funds from a competitor. The fundamental shift is from looking backward at old data to looking ahead at probability. And this is where it gets really interesting.
The sources we looked at lay out these four kind of powerful contrasts between the old way and the new way. The first is just the philosophical difference. Traditional analytics, it reports on what happened last quarter. It gives you a dashboard. It's history. It's history. PI on the other hand. Uses these models that anticipate what will happen, you know, which households are likely to, uh, increase their savings over the next 90 days.
And the second difference is segmentation. Traditional [00:04:00] systems are so rigid, they use static segmentation. Right. Those broad buckets like all customers between 40 and 55, or everyone who lives in the zip code. Exactly. Whereas PI uses dynamic behavior-based micro segments. A customer isn't defined by their age.
They're defined by what they're doing now. So if they suddenly start paying down debt or researching CDs online, the system instantly shifts them from, say, a low intent saver to a high propensity CD purchaser. It's completely fluid, which changes the output, right? The old way is that dreaded blanket offer the get $200 for a new checking account mailer that everyone gets and that PI delivers.
It delivers personalized, timely engagement because the system knows that customer's intent is high right at this moment. And that leads to the last big difference moving from lagging indicators like campaign ROI from three months ago to real time opportunity signals. So the system is combining all this data to trigger outreach before the customer even starts shopping around.
You're getting ahead of it. You're [00:05:00] preempting the competitive shopping process entirely. That's the goal, that precision is really the key to everything. Efficiency, retention, growth. It is. So let's drill down into how this works in practice. How are banks using PI to actually build smarter deposit campaigns?
We found three really specific ways. The first one is just fundamental, targeting the right segments. It's all about efficiency. I mean, if a bank has 10,000 customers, only a tiny fraction are actually ready to move money today. So PI finds that fraction. Exactly. The models isolate, say the top 500 customers who show the highest intent to add funds in the next six weeks or so, and focusing on like just that small group.
I mean, the results are pretty staggering. We saw a mid-size institution that got 54% deposit growth in just a few months with this strategy. Yeah, but hold on, 54% that sounds. Almost too good to be true. Did the source material give us any kind of control data? Like how do we know that wasn't just a good economic environment where everyone was [00:06:00] saving more?
That's a critical question. And in this case, yeah, they did. The institution said the outperformance came from two things. One, they cut their general marketing budget to almost zero. For that quarter, so they were all in on this targeted group all in and two, they didn't just sell any product. They targeted high balance existing checking customers with a premium high yield savings product.
Felt exclusive and timely. And the control group, the control group, got the standard generic offer and they saw almost no lift at all. So the system really proved its value by focusing the budget only where the behavior was already trending towards saving. That makes a lot more sense. It's about converting people who are already leaning that way.
So what's the second application? The second is timing offers for maximum impact. Because even a perfect offer can fail if you deliver it at the wrong time. Timing is everything in finance. It is. PI uses signals to detect these life events or financial behaviors that mean a customer is suddenly flushed with cash or ready for something [00:07:00] new.
So give us some examples. What are those signals they're looking for? Well, some are seasonal, like tax refund season, obviously. Mm-hmm. But it gets more subtle. It could be a change on their credit report showing a big loan was just paid off. Like a car loan? An auto loan, yeah. A student loan. Yeah. That frees up a lot of monthly income.
Or maybe their payroll deposits are consistently higher, which signals a new job. PI flags these moments as green lights. That's incredibly specific. And the third one, personalizing messaging at scale. I'm guessing this is where that dynamic segmentation really shines. It really is. The goal might be the same grow deposits, but the language, the channel, the product.
It has to shift for each segment. So a 28-year-old professional whose model shows high savings potential, they might get a message about maximize your first $10,000 with an investment starter account, right? While a 65-year-old retiree who the model says values safety and fixed income, they might see something about structuring your [00:08:00] retirement cash flow with a custom CD laddering strategy.
So they're both deposit campaigns, but they look and sound completely different. Exactly, yeah. And the result is much higher engagement, a real deposit lift and lower costs, because you're only talking to people who are already thinking about making that kind of move. And speaking of inputs, let's pivot to the data itself.
Banks have their own raw materials, you know, checking balances, loan histories, but our sources say the big missing pieces. Context. Context is everything. Internal data only gives you half the picture. If a customer has $10,000 in their checking PI can't really predict their next move unless it knows why that money's there and what the rest of their financial life looks like.
Exactly. So how does PI fill in that missing context? What are the external signals? It's layering on top. We're looking at about three main buckets of data enrichment. First is expanded demographics. This goes way beyond age and address, so things like estimated household income, life stage. Yeah, they're a first [00:09:00] time home buyer, a recent empty nester, things like that.
Second is geospatial trends, meaning the models look at things like, has a competitor just opened a new branch down the street, or is this neighborhood seeing rapid growth and an influx of higher income residents? It helps predict market opportunity. And the third one, the behavioral indicators. This is the juicy stuff, right?
This is the juicy stuff. Spending patterns that happen outside their own bank, changes in credit utilization, their digital engagement habits. And when you enrich the data with all of that. You could surface these incredibly powerful insights. For example, PI can identify a household that's at risk of leaving, not because their balance dropped, but because, but because they increased their credit card spending at a competitor's bank and started looking at competitor rate sheets online.
A static dashboard would never show you that. So that's the insight. But moving from insight to action is what makes it a system. Let's walk through that continuous cycle, what the sources called the pi [00:10:00] flywheel, right? It have to be a continuous loop. So step one, enrich. You're constantly aggregating and.
Cleaning the internal and external data. Step two model. The machine learning generates updated scores for every customer predicting their intent to deposit or their risk of leaving. And step three is target. You're segmenting based on those scores. So instead of a group of 10,000, you have a microsegment of 500 people with say, a 75% chance of opening a CD this month.
Step four is where you act on it and engage. You deploy that specific personalized message based on what the model predicted they need. The loop closes with step five measure. You track what actually happened, compare it to the prediction, and use that result to retrain the models so it gets smarter with every single campaign.
It's a self-optimizing system, and the results of this flywheel approach, they really speak for themselves. They do, let's look at the numbers. We saw one regional bank that secured an estimated $230,000 in new deposits from a single [00:11:00] AI driven campaign. And what's even more impressive than the dollar amount, I think, is the engagement, is that the campaign drove customer engagement 80% higher than their standard marketing.
That 80% lift is the real proof. It means the system is actually working, it's getting the right message to the right person at the right time, and it works for retention too. Another institution reported reducing its deposit attrition rate by 7% in a single quarter just by using pi. Just by targeting proactive campaigns only at those households, the system had flagged as high risk for leaving.
What's also fascinating is how PI uncovers these segments that were just overlooked before. You know, the quiet steady savers who don't respond to a $200 bonus. Are very responsive to a real conversation about long-term wealth strategy. Exactly. So it directs the marketing budget to these high value segments, meaning you can grow deposits without actually increasing your budget, which ultimately connects this all back to the highest level of strategy.
This isn't just a marketing tool anymore. It's a core growth capability. [00:12:00] It's a competitive imperative. I mean, McKinsey projects that this kind of data-driven personalization can drive up to a 15% revenue lift across the entire bank, and the benefits are felt everywhere. Huh? Institutions using AI across both deposits and lending are reporting 20 to 30% lower acquisition costs and double the retention rates, two times higher retention compared to banks using the old methods.
Yeah. These are just. They're non-negotiable metrics in today's environment. So the future of banking isn't reacting to what a customer just did. No, it's not even reacting to what they just searched for online. The next generation of successful banks will anticipate those actions. PI empowers them to understand the whole journey and respond with precision.
It's a complete flip of the script. No more guessing. It's about anticipating needs. And that leads to, I think, a necessary and fascinating final thought for you to consider. If this technology lets institutions understand their customer's journeys with such precision, I mean [00:13:00] identifying households at risk of attrition before they even consciously know they want to switch.
What does that mean for privacy? What does that level of anticipation and data collection mean for customer privacy and critically for customer trust? That tension, you know, between convenience and surveillance is the line we always walk with these powerful data tools. As you move forward, maybe reflect on your own bank, how often does an interaction feel truly personalized versus when you get yet another generic blanket offer?
That's the dividing line. That distinction is the bright line separating the old guard from the new generation of predictive banking. Thank you for joining us for the deep dive. We'll see you next time.