With more and more non-bank companies now offering banking products, data that was previously accessible only to financial institutions is now in the hands of fintechs and other companies launching embedded banking products. This data can provide a wealth of insights into the financial health of customers, along with information about how they spend their money and where.
The only question is, what to do with all of it?
This is a question banks have grappled with for decades, but for newcomers — and especially untraditional finance companies like X, Apple and Walmart that are joining the game — it may be tough to know where to even start.
As a former chief architect for a large financial institution, a previous CTO at an innovative challenger bank, and an adviser for a fintech investment firm, I’ve seen this question from all sides.
I’ve seen large banking institutions miss out on valuable revenue opportunities because their legacy systems and poor data architecture prohibit them from getting the right data at the right time. I’ve also seen fintech newcomers struggle to fully understand how their customers are using their product due to not knowing how to interpret the banking data in front of them.
The companies that really succeed are the ones that do both well. They structure their banking data in a way that allows their team to easily understand how customers are using their product. Then they’re able to turn this understanding into actionable insights that allow them to improve their customer experience and mitigate fraud.
Proper data structure
The deeper you understand your customer, the better you’ll be able to make product improvements and refine your marketing and sales strategy.
As a recovering banking chief architect, I still keep in touch with some of my friends who remain in that world. And I used to have a running joke with the chief architect of one of the largest banks in the world, based in the U.S. and who will remain nameless. Whenever I would see this friend, I’d ask him a simple question: “How many data systems do you have now?” It was a joke because he would never know the answer.
Forget learning how to use data. If you don’t have a single source of truth for your customer data that is easily readable, you’ll never be in a position to properly leverage it to scale. Unfortunately, my friend’s dilemma is all too common in the banking world, both in legacy institutions and emerging fintechs.
Fixing such a dilemma can take years — I should know, because I had to do it. But getting your data structure right can make a world of difference in the long run.
So what’s the key? First and foremost, it’s important to have all of your data housed in one centralized system. Having data scattered across multiple systems, all with different data points, will only make things harder at scale.
Second, it’s important to understand what sort of data you will be collecting as a bank so that you can ensure it is structured correctly in your system.
Banking data can be broken into three pillars:
- Customers: This is all of the data related to who your customers are, including their demographic data, contact information and behavioral history.
- Accounts: These are your different products and the attributes associated with each. For example, if your savings account has a minimum account balance requirement of $100, or if your credit card product charges a specific monthly interest rate, this data would be housed in the account pillar.
- Transactions: These are the records of all the money movement happening using your product, whether that is money moving into, out of, or within your product.
Now, once all of the data pillars are structured correctly, the next step is to make sure that the relationships between them are clear and well-defined.
This is key to giving your team immediate insights into who is using your product (customers), which products they are using (accounts), and how they are using those products (transactions). At scale, having this clear view will allow you to leverage the data to grow.
Delivering a better customer experience with data
The deeper you understand your customer, the better you’ll be able to make product improvements and refine your marketing and sales strategy, with the ultimate goal of delivering an exceptional customer experience.
In my previous life as the CTO of a challenger bank, we embarked on a mission to truly understand how our customers were using our banking product. We knew building savings was a key use case (as it is for many banking products), but we didn’t know exactly what those customers were saving for.
So we dug into the data. We found that many customers had their savings accounts labeled as “Mexico” or “Europe.” Now, we knew they weren’t saving up to buy all of Europe, but we did uncover that many of our customers used the savings product to fund their vacations.
Since all of this data was centralized in one location, we could easily see this as a common pattern among our customers, helping us understand the cluster size of travel savers versus other types of savers.
This learning prompted us to build new savings features tailored specifically to vacations, build out cross-selling campaigns for products that specifically benefit international travelers and influence our marketing strategy to appeal to the adventurer. All of this led to happier customers and increased customer LTV.
When you can use data to truly understand your customers, that is where the gold is. When you can clear up these mysteries, that is where the opportunity lies for your company to deliver compelling offers just at the right time and solve problems for them that other organizations can’t. If you don’t know your data, you don’t know your customers.
Mitigating Fraud With Data
With regulatory scrutiny of fintechs and their sponsor banks increasing (and I am happy to see the push for transparency in partnership banking), building a comprehensive strategy to mitigate fraud should be top of mind for anyone operating a banking product. From my experience, fighting bad actors and thwarting illegal financial schemes becomes much simpler when your data clearly explains who is moving money to whom.
This is where the relationship between the three pillars of data becomes crucial. For example, seeing transaction data on its own could be more helpful. But when you can pair that transaction data with the customer data, you can see how people are moving money using your product. And if you don’t know who’s who, it’s hard to know who is moving what.
The vast majority of the time, this money movement will be expected behavior, but now and then, you’ll see something that catches your eye. Understanding what is normal and why money moves is a great way to understand what is abnormal – when something doesn’t make sense, that is a good time to dig in and understand the “why” behind the data. Often, the “why” is fraud.
For example, why would a customer only send a transfer to another customer to have that customer return it? Why would someone load funds and then try and take the funds back out of the platform?
Knowing who your customers are and understanding the why behind money movement is how you can effectively detect anomalies and irrational money movement – dig into this, and you can often find bad actors.
Aside from knowing who is moving money to different locations, it’s essential to structure your data to provide visibility into how money flows in and out of the system. Transaction data must be structured utilizing double-entry bookkeeping.
With such a configuration, each transaction has two corresponding sides — a debit and a credit. Then, through a process known as reconciliation, you match both sides of the transaction to ensure they are always equal.
From time to time, you may encounter a situation where there is an imbalance. Your credits do not equal your debits. This is your team’s cue to start chasing down the source of the imbalance to understand where the mismatch occurred and bring it to resolution. Now, much of the time, this imbalance is benign and easily addressable, but sometimes it can be a sign of fraud.
If this imbalance is unaddressed, issues can compound and become more serious. Money could leak out of the product, funds could be accidentally (or purposefully) co-mingled, or someone could use the platform to print money. All of these instances could result in losses for your business and potentially even a knock on the door from the regulators.
The banking data’s job is to provide a clear view of how money moves in and out of the product. With this understanding, your entire team becomes exponentially more effective at keeping your product safe and secure for your customers.
I have a mantra around this: funds in motion are funds at risk. My advice? Track the micro-movement of money and collect every scrap of information along the way.
There’s nothing more exciting than launching a fintech or embedded banking product and watching the customer and usage growth. It’s one of the most thrilling parts of being in the industry. On the flip side, there’s nothing worse than dealing with a deluge of data without a plan to structure or utilize it.
Finding avenues for growth and mitigating fraud becomes more straightforward when you find the magic combination of having usable data and knowing how to leverage it.
And this efficiency is one of the keys to building a long-term, scalable business.