Data Discovery Process

Data Discovery Process for Marketing Strategies

When developing effective marketing strategies whether it’s for financial institutions or other industries, there are many factors that can help you narrow in on your best opportunities.  To help you get there data discovery should be part of your journey.

Data Discovery Process

Why should FI marketers complete a data discovery process?

  • Helps prioritize your marketing initiatives for developing your market plans
  • It leads to a final stage of recommendations and outcomes that places value on each recommended marketing initiative
  • Sets you up to automate your next best action in real-time-based key attributes such as high value or change in value among your customers
  • Can be a turnkey type of solution when developing marketing strategies based on segments and their overall value

In this article, we will highlight what some of the activities might look like when going through the data discovery steps within the financial services sector. In fact, we will look at the process of defining and analyzing customer value which has been a common need based on our experiences with FIs.

Deep Dive into Data Discovery for Financial Institutions

The data discovery exercise has now been recognized as a core business discipline for any organization that is commencing its journey toward a more data-driven business strategy. At a high level, the 4 stage concept can be considered to be ubiquitous for all industries. But within these 4 stages, the specific tasks and activities can be quite different for specific industries and in fact for specific companies. 

The concept of customer value for a bank or FI can be quite unique due to its complexity. For example, customer value can be quite simple for a retailer as it can be determined based on customer purchases. In other industries such as insurance, the premiums of the policyholder are what determines value. However, the risk element is more significant in how these premiums are priced which adds another level of complexity. For banks, though, there are many layers of complexity which will be discussed as we go through the 4 stage process.

First Stage:  Preparation

In the first stage of preparation, a  series of meetings are established with key stakeholders of the bank. Given that the exercise is to define value, a number of meetings would be set up across a variety of different areas within the bank:

  • Personal Loans
  • Mortgages
  • Retail banking(Deposits/Withdrawals)
  • Wealth/Investments
  • Credit Card

It should be noted that in many cases, each area might have its own algorithms for calculating value.      For example, we did work for an FI where the entire exercise was determining how to define value for a credit card customer. We had to understand the revenue components of the credit card such as interest revenue, fee revenue, and the actual revenue associated with each credit card spending transaction. At the same time, we had to determine the cost component which was the risk of default as defined by not making a payment within 60 days.  The end deliverable, though, was an algorithm that defined credit card customer value.

This same kind of process would then be conducted across all the other areas of the bank where each area might have its own algorithm.

Second Stage: The Data Audit

Once we understand the mechanics of how we calculate customer value, we then need to conduct theData Discovery Process: Customer Value Chart second phase of the project which is the data audit. In this phase, we would be looking at every field from each file to understand its usefulness and reliability in calculating customer value. The data audit would also reveal how we can link all this data so as to obtain that one view of the customer which can be used to calculate customer value. The end deliverable here would be to link all the required information from these disparate areas of the bank into one overall algorithm. This sounds easier said than done as much of the hard work in this exercise occurs within this data-intensive stage.

Third Stage:  Preliminary Analysis

At this point, we enter stage 3 which is the preliminary analysis. It is at this point where can identify our high-value customers i.e. how much do high-value customers contribute in value relative to the entire base? See the below chart for example:

From the above chart, our top 20% of high-value customers contribute 69% of all the value. Now that we have identified them, can we further explore what they look like in terms of what is contributing to customer value.

DecileValue Segment% Purchase Amount% Interest% Credit Loss% Card Fees
1High44%48%27%39%
2High19%21%14%19%
3Medium13%13%9%13%
4Medium9%7%6%16%
5Medium7%5%5%4%
6Medium5%2%3%3%
7Low2%1%2%4%
8Low0%0%1%0%
9Low0%0%0%0%
10Low1%3%34%1%

This chart just looks at some of the metrics which drive credit card profitability. But what is this telling us? Not surprisingly, our best customers are those who generate revenue by revolving their balances. Yet, this group also encompasses the largest amount of credit card losses thereby suggesting the need for a risk strategy that focuses solely on high-value customers. This is not surprising as well as our best customers will be those that are most engaged and who will have a high degree of credit losses. Nonetheless, one could argue that nothing needs to be done to the bottom 80% from a risk strategy standpoint.

But how do we create a strategy for obtaining more of these types of high-value customers? Further segmentation besides value would focus on how this behaviour is changing over time. For example, can we identify medium-value customers that are growing or more importantly high-value customers that are at risk of defection? Both these outcomes would warrant the need for a  defined marketing strategy that encompasses both digital and offline channels. Even customers that are considered no longer active should be given some attention by marketers particularly if we can identify high-potential value customers or those inactive customers which most resemble high-value customers based on demographics such as age, income, etc.

Final Stage:  Recommendations

This leads us to the final stage or recommendation stage where with our analytical outcomes from stage3, we can now develop a roadmap that actually places a value on each recommended initiative. It is this roadmap that allows us to prioritize initiatives and the development of a marketing plan. With these stages now completed,  we can automate these processes such that high value and change in value can be determined in real-time. This can be a  turnkey type of solution with marketing developing strategies based on these segments and their overall value. Although the end game here is extremely attractive for marketers, no one should underestimate the amount of upfront work in the development of these analytical outcomes. But to use that tried and true old phase which is very appropriate here: “No pain, no gain”.

Over 40 Years of FI Marketing Experience

360 Integral Marketing helps financial institutions and fintech organizations overcome today’s challenges with a team that has the marketing expertise and industry experience to deliver customized solutions. 
Richard Boire About the author

Richard is an industry leader with over 25 years of expertise in data-driven strategies, data analytics and predictive modelling, with a focus on financial services. He successfully launched the Boire Filler Group, which was sold to the leading data analytics company, Environics Canada, in 2016. He is a prolific guest speaker and has also authored “Data Mining for Managers: How to use data (big and small) to solve business problems”.