Targeted Data, Not Big Data
Many pundits and organizations are declaring Big Data as the route to future success for banks. I disagree. Big data may be a route to success if you have the scale and capacity to use it, but it’s not the only or even a necessary route. The far more effective route is “targeted data,” which provides the answer to a specific management question or means to achieve a particular goal.
In marketing, targeted data is generally defined as data that identifies the reason for a customer taking a specific financial action. For example, if your goal is to open more personal checking accounts at a lower cost, there are two groups you can target. The first group is those who are intensely dissatisfied with their current provider, so intensely dissatisfied that they will take action to leave them. The second target is people moving into your trading areas.
The list of movers needs to capture the prospects’ names before they move since movers open checking accounts no later than one week after establishing themselves in their new home. The list of dissatisfied consists of those who are actively searching for a new checking account within your trading area. Both lists constitute targeted data because they include people who are seeking to obtain a new checking account. How effective is such data? You can generate new checking customers at half the normal acquisition cost with them. I know this because I have done it. At Maryland National Bank, I developed a new movers checking sales program after talking a major mortgage processor into the list business, generating accounts at half the next best cost. At MBNA America Bank, I grew the personal loan portfolio at a rate in excess of 35% for years using targeted data based on life events.
Targeting List Sources
There are two requirements in using targeted data. The first is understanding your product customers well enough to pinpoint who to target when. The second is finding a list source that will provide valid names on a timely basis. Target names can be purchased from traditional list sources to meet many of your objectives, especially when the financial need develops over time. For example, to sell high-dollar personal loans, you will target those who are experiencing life events that frequently involve spending a large amount, such as marriage, birth of a child (especially the first one) and renovating a home after buying it. Organizations that sell goods related to these events compile and publish lists. And, because the actual event is known well in advance, the lists are available on a timely basis.
When the timing required is as precise as selling checking accounts to people moving into your trade area, creativity is required because published list sources are out of date before you get them. Will mortgage brokers or real estate agents provide their customer names to you after those customers have selected the house to buy? Is there an online website that enables you to advertise on a geographically-targeted basis?
Compare this to big data, where you collect everything there is to know about someone, mine it for knowledge about the financial needs of customers and then use it. Big data is a more expensive approach because of the specialized systems and staff needed. Big data is inefficient since it collects every piece of data on every person in order to find those few who are within the time frame of making the financial decision or action you have targeted with your product. Further, the resulting lists are composed of those likely to be in the market, not those who are definitely in the market. This means some recipients of your offer will be highly inappropriate, creating an adverse reaction at receiving your offer, which can lead to reputation risk problems longer term, or employee dissatisfaction and disbelief in your targeting.
One of the strengths of targeted data is its focus on the cause for someone needing to take a financial action. Big data, by contrast, almost always uses correlations of behaviors to target offers, and correlations can change with environmental changes. Therefore, big data models are subject to decay over fairly short times and must be continually tested and adjusted to stay useful. Worse, some big data models will remain valid for a long time and then change suddenly, leaving users at a loss for their ineffectiveness and suffering the consequences from using an outdated model too long.
A second strength of targeted data is you can generate at least part of the list you need internally at low cost. You just need a good customer relationship management (CRM) or similar system to move the data from where it is captured in your organization to where it has the greatest information value. In our checking account example above, you need a way to move the names of people moving into your trading area from your mortgage approval process to your retail area. When branch staff learn about an impending marriage are they trained to recognize this as a possible lending opportunity and preserve that information for use by the loan area?
Targeted data can be used in all areas of your bank, not just in marketing. For example, changes in checking account deposit behaviors, such as a decline in value deposited, can be used to warn your personal loan area of possible collection issues with particular customers. Similarly, analyzing deposit and payment behavior in the operational accounts of commercial borrowers brings their cash flow to view for monitoring loans automatically, or possibly offering line increases.
Mr. Merkle is the founder and CEO of Unionville, Penn.-based CashFlow Insights, LLC, a firm specializing in improving the profitability of financial services firms. He can be reached at [email protected].