Menu
Banking Exchange logo215mar2015
Menu

12 key definitions in advanced analytics

Part 2 of series: Get to know the lingo of this critical new science

  • |
  • Written by  Steven Simpson, Saggezza
  • |
  • Comments:   DISQUS_COMMENTS
12 key definitions in advanced analytics

Advanced analytics programs are moving the meter on how banks attract and retain customers, prevent fraud, and generate new revenue streams.

As a result, adoption of third-party analytic services continues to increase. If you are grappling with how advanced analytics techniques can significantly impact your bank’s business, knowing the lingo—and understanding why those concepts matter to your bank—is vital.

Reviewing the following quick list of key terms—along with the explanations of their significance—will help you understand what’s being discussed.

1. Advanced analytics: Examination of data or content using sophisticated techniques and tools, typically beyond those of traditional business intelligence, that uncover deeper insights, make predictions, or generate recommendations.

Why it matters: Traditional analytical tools examine the bank’s past trends and historical data. By contrast, tools for advanced analytics focus on forecasting future events and customer purchasing or behavior patterns. This allows the bank to conduct “what-if” analyses to predict the effects of potential changes by the bank. Applied well, this can help by improving customer profit and retention; creating new products or services, further segmenting customer bases; spotting potential fraud; and more.

2. Prescriptive Analytics synthesizes big data, mathematical sciences, business rules, and machine learning to make predictions and then suggests choices to take advantage of the predictions.

Prescriptive Analytics goes beyond predicting future outcomes. It does so by suggesting actions to benefit from the predictions and showing the decision maker the implications of each decision option. Prescriptive analytics not only anticipates what will happen and when it will happen, but also why it will happen.

Why it matters: Predictive analytics uses statistical models based on past events to predict what is likely to happen given certain conditions. Prescriptive analytics suggests actions based on criteria. For example, this technique could compare the Zillow Zestimate—or estimated value from Realtor.com—to outstanding mortgage balances to calculate LTVs. Based on those LTVs and other characteristics, a bank could learn that a specific action (such as a change in a campaign pitch or message) may result in a given rate of success when an offer is made for a home equity line of credit.

3. Prescriptive Action: The statistically proven best action.

Why it matters:  Imagine having a set of recurring best practices that have been statistically proven to move the needles that matter most: profit, retention, cross-sell, number of products per customer, average product profitability, etc.

Want more banking news and analysis?

Get banking news, insights and solutions delivered to your inbox each week.

From prescriptive analytics an action will be suggested. But what action is best? It will be important to try multiple actions, compared to a control group, refine the action based on what has been learned, and repeat what works. In a short time, the best action will be revealed for specific customer segments.

4. Average Product Profit: The average product profit is a calculation of the profit divided by the number of product units.

Why it matters: For years, bank executives thought it was always imperative to know the average profit of their products so they would know which product to sell.

However, there is a better approach. Naturally, you want to know which products are unprofitable, but more importantly, within each of those average profit curves—for each product—there are customers with both unprofitable and profitable habits or usage characteristics.

You want to increase the Loyalty Measure [see later definition] of the profitable segment and gain the ability to change the habits of the unprofitable segment. You also want to target customers with characteristics and trends of the profitable customer-base.

A simplistic, general sales campaign, based on Average Product Profit, may only result in “successful” selling to more customers who will use the product unprofitably and lower the average profit. Advanced analytics will help target a different action with better ultimate payoffs.

5. Data collection and cleansing: The first step in the process of advanced analytics is to simply collect and capture bank data arising from multiple sources—online, branch, social media, phone calls, for example.

Why it matters: The process of collecting, reviewing, and cleaning the data involves removing duplicates, inaccuracies, inconsistencies, and other errors that can skew findings and produce flawed results.

6. Algorithm: A mathematical formula placed in software that performs an analysis of a set of data to help solve a problem or find a common divisor.

Why it matters: Once data is collected and cleansed, algorithms compute the information to add meaning for more data-driven decisions. Using advanced analytics to answer “what if” questions combines historical performance data with algorithms, and other data, to determine possible outcomes. Examples of algorithms may include likely profits or contributions by customer, product, officer, and branch, a loyalty measure that can reveal a relative retention or attrition statistic, number of products per customer, a next product model, and more.

7. Data Lake/Data Warehouse: It is important not to confuse these terms.

A Data Lake is a large-scale repository and processing engine that provides massive storage for any kind of data, enormous processing power, and the ability to handle virtually limitless concurrent tasks. A data lake or big data application is created from unstructured data. It couples existing data with external data and is meant to answer questions not yet conceived.

A Data Warehouse is comprised of structured data and designed to answer specific questions.

Why it matters: Analytics can only be applied to well-organized data. By itself, a data warehouse is actually outdated technology.

As an example, a bank may begin by looking to determine why a group of branches had a decline of $10 million in deposits on the prior night. The first place management looks reveals that two branches actually increased deposits by $5 million each. Now they are not looking for $10 million, but rather $20 million.

By the time the bank is finished moving through a Dashboard [see later definition] empowered by a data lake, the to-do list may become highly focused—such as specifically working with two clients with large commercial loan renewals, below the Weighted Average Rate for loans of that type, that are coming up in the next four months and that are owner-occupied medical offices.

8. Dashboard: A graphical reporting of static or real-time data on a desktop or mobile device.

Why it matters: After data is collected, cleansed, and computed, the dashboard provides the consumption layer. With a customized dashboard the bank’s key players gain insight into the real-time performance of centers of influence such as: customer relationships, loans, deposits, and profits.

9. Loyalty Measure: Loyalty and retention measures deliver a quantified metric that can be used to focus cross-selling efforts to either improve retention on your best customers or offer the right product at the right time to other customer segments.

Why it matters: Loyalty metrics can help in the big picture as well as the very small. They provide valuable insight on how different customer segments are trending. They can also provide early warning signs of an issue with an individual customer.

The Loyalty Measure provides important information when comparing new customers (in which the bank views which branches, or which officers, have the best success at cross-selling). They can help a bank consider what action could increase Loyalty Measure on a profitable customer segment, which could result in extending the average life of the relationship.

The goal is to understand how to best measure customer loyalty to combat attrition, better manage customers’ experience, and find ways to cross-sell and improve relationships.

10. Value Finder or Opportunity Segment: This refers to zeroing in on a grouping of customers that are statistically and quantitatively derived, with indicators that suggest a specific action on a specific list of customers who would be best served by a specific product or service the bank offers.

Why it matters: Uncovering untapped opportunities within specific customer segments leads to the potential to cross-sell and up-sell customers. Over time, through the use of advanced analytics, the bank will be able to identify other segments of opportunity by applying statistical models on historical behavioral data.

11. Institution Specific Equation: Your customers, services, and products result in a unique equation derived specifically from your institution’s circumstances. Your next product model, loyalty measure, profit algorithms, and more will self-adjust, adapt, and become unique to your bank over time.

Why it matters: No two banks are alike. In fact, even within a single institution there may be the need for different equations that deliver prescriptive actions, for different markets and types of products. For example, a recommended next product model will likely vary for a customer in a large metro market versus a rural market, for a retirement-aged customer versus a millennial, and a college student versus a corporate executive.

12. Data Scientist: Data scientists usually have a PhD in statistics or mathematics. They are a new breed of analytical data expert with the technical skills to solve complex problems and recommend outcomes by analyzing data.

Why it matters: According to the job recruitment website Glassdoor, Data Scientists are at the top of best jobs in 2016 and for good reason. These high-demand jobs require a high quantity of specialized knowledge, making them a rarity at this time. The annual base salary averages $117K.

The costs of that expertise, and the need for expertise in data science to be married to banking expertise, has caused some institutions to outsource rather than hiring data scientists.

Don't let fancy terms around advanced analytics become an obstacle to your bank discovering the profitable insights this disruptive and transformative technology can provide. Keep this list handy when meeting with third-party providers … or even during your next management planning meeting.

Coming up, the conclusion to the series: Benefits of adopting advanced analytics

About the author

Steven Simpson is senior vice-president—financial institutions at Saggezza—a global solutions provider that develops and implements leading analytics products. He can be reached at Steven.Simpson@Saggezza.com or at 786-859-4100.

Part 1 of this series: 5 key questions on advanced analytics—answered

Also see “Dawn of the data-driven bank,” which discusses the experiences of two banks with advanced analytics

back to top

Sections

About Us

Connect With Us

Resources