Printer Friendly

Data Analytics--Army Financial Management Imperative.

Data analytics and data science are buzzwords du jour in business and government management, these days. We live in the era of "big data." But when you ask people about data analytics, they either shrug their shoulders, or everyone gives a different answer.

There is no common understanding of what it means. Is it number crunching? Is it machine learning and real time cloud computing? Is it math and statistics? Is it slick, interactive visualizations?

The answer is yes. The sheer volume of data being produced is staggering, requiring new tools and fresh ways of thinking to turn the raw data into useful information for better decisions. There is, however, a key ingredient that has always been a requirement for analysis of any kind. That is the functional expertise leaders have for their business processes. This is true whether the endeavor is supply chain, financial management, personnel management, or the art of war. Stephen Few, one of the early gurus of data visualization, says it well when he notes, "No analytical technologies or technical skills will overcome a scarcity of sound reason."

Data scientist Drew Conway was one of the first practitioners to define the multiple skill sets that encompass data science. He constructed a three-part Venn diagram, linking math and statistics with hacking skills (computer programming) and substantive expertise (functional knowledge). As Diagram 1 on the left shows, overlap of any two domains can give at most, partial solutions. At worst, it creates a danger zone where people with functional knowledge and some programming abilities produce solutions that are unrealistic because they don't have the math skills to verify or validate the data. Only by connecting all three areas could someone fully explain the complex interactions that make up data science. Stephan Kolassa is a software creator, author, and forecaster who expanded on the three-part diagram by adding communication (the visualization and presentation of data) as a key skill set. As his diagram (Diagram 2) shows, adding communication presents a broader range of partial solutions. Both diagrams attempt to demonstrate and clarify what data science actually is while also pointing to the realization that very few people can be experts in all areas. Data science Is a team effort to provide decision support to organizations.

Decision support In the Army is centered around a unit commander. The staff gathers information and conducts analysis to present recommendations to the commander for decision. The staff then distributes orders and guidance to subordinate units. If the information takes too long to get to the commander, the decision point Is passed, and even the best analysis becomes irrelevant. If the data is delivered fast enough to be considered real time, but can't be communicated In a meaningful way, it does no good. And, if the data is random, full of errors, or Improperly assessed and points to wrong, or meaningless conclusions, the time and effort to analyze it becomes a waste of manpower and resources.

Recently, the Army Financial Management School hosted a two-day working group, comprised of officers, non-commissioned officers, and civilians who are on the front lines of data analytics. The group represented organizations that included brigade, division, theater army, combatant command headquarters, and the special operations community. On the first day, group members shared current tools they had developed, explaining the problems the tools were designed to solve, what questions were being answered, and how raw data was transformed through the analytical process into useful information.

The second day, the group talked about the purpose and use of data analytics for the Army, and how the Army could address the growing mountain of data, to cut through the noise and provide actionable information to commanders. This decision support is the main goal of data analytics. The group also talked about the common questions that commanders ask and common problems that financial managers deal with, to try to identify areas for improvement across the Doctrine, Organization, Training, Materiel, Leadership, Personnel, Facilities, and Policy (DOTMLPF-p) domains.

The Data Analytics Process

The group identified a six-step data analytics process that can facilitate turning raw data into useful knowledge. When a question or problem is presented, the following steps are necessary:

(1) Access to Relevant Data. There are multiple sources of data which must be combined to draw conclusions and support decisions. Financial managers need access to those sources. Depending on the organizational level and duty position, some financial managers are currently not able to access the data necessary for analysis, so the process never even starts.

(2) Formatting. Data sets are not always formatted in a way that can be linked to other data, processed through analytical tools, or put Into useful visualizations. The data structure sometimes has to be rearranged or reformatted before analysis can begin.

(3) Data Connections. Analysts must have a good understanding of data relationships. This is the foundation for how databases work. One piece of information from a certain group of data is somehow related to another piece of information from a different group. Crucially relevant knowledge can be developed by linking the data through relationships. Financial managers may not need to be experts in a specific database programming language, but they must be able to understand how database relationships work to link the information.

(4) Validation and Cleaning. After relationships have been identified and the data sets joined together, the resulting output frequently has some unexpected complications. The larger and more diverse the data being joined, the greater the potential for problems. There might be Items that are double-counted, while others have been uncounted. Validation and cleaning Is a critical quality control step that has to happen after joining data, but prior to analysis, to make sure the newly joined data is accurate and relevant.

(5) Analysis. This is where the value Is added, as raw data turns into meaningful knowledge. Statistical methods Identify significant correlation. Tests are done to find causality, trends are identified, comparisons are made, and conclusions are drawn. This is also the step where numbers are aggregated and computations are performed to discover and highlight relevant Information. Analysis is the stage where information is transformed from raw data into actionable Insights.

(6) Communicating the Analysis. Based on the volume and type of Information, and the functional questions being asked, financial managers need to use communication techniques that are helpful in making the pertinent information understandable. The old saying goes, "A picture is worth a thousand words," but there is significant cognitive science actually backing it up. The human brain can find meaning much more easily with a graph or chart, than with a table filled with text. However, detailed decisions often require careful scrutiny. Fortunately, with modern software and the right training, financial managers have the potential to develop data visualizations for the commander, while also allowing detailed, drill-down capability for the analyst.

Standardized Tools vs. Standardized Knowledge Base

The data working group at the Financial Management School asked the question, "What is the best method for the Army to conduct data analytics?" Would it be better to have a common dashboard that every unit used? If that were to happen, the institutional training courses would teach the common dashboard, and no matter what unit the financial managers were assigned to, they would all know how to use those tools. Conversely, perhaps it would be better to teach a more basic data analytics framework, so all financial managers would have a baseline skill set to build their own tools. This seems appealing, because every commander is different, and every situation is unique. The questions asked across the Army are not all the same. The problems are not all the same. Perhaps, instead of teaching common tools, we work to ensure a common knowledge base.

The Fort Jackson working group settled on a middle ground. There are, indeed, efficiencies to be gained by using the same tool across the Army. After all, there are great similarities between infantry brigades across the Army and divisions are structured in common ways. Everyone uses some of the same systems, such as the Defense Travel System, General Fund Enterprise Business System (GFEBS), and Global Combat Service Support-Army (GCSS-A). Some business problems are the same, even If the answers to the questions vary. Everyone has cost drivers, even though the specific cost drivers may be different across the force. The process of finding and displaying cost drivers, however, could be standardized.

The working group agreed that we have a fundamental problem. The raw data itself is not standardized. Organizations have the freedom to design their own cost structures rather than mandated hierarchies and naming conventions. Because the underlying structure of the data is different across the force, it won't fit neatly into a common tool for all units. This led to the shared realization that we need to restructure the data in the system. Otherwise, a certain amount of customization of tools and solutions will still be necessary, whether we like it or not.

The Army has implemented Enterprise Resource Planning systems to handle financial transactions (GFEBS) and logistics (GCSS-A), and will soon implement a new human resource system (the Integrated Personnel Planning System-Army, or IPPS-A). Each of these systems is said to be an enterprise system. Ironically, they are three separate systems. While this might be understandable from the realities of the Army acquisition process, from a data perspective, it creates problems, rather than solving them.

Diagram 3 below shows how the Army systems compare to one truly "enterprise" system. Separate systems negate the benefits to be reaped from one, overarching system. As it currently stands, financial managers must develop a database outside all of the disparate systems to combine the information to do their jobs.

Units end up creating work-arounds in an effort to join information from multiple systems that have similar data. This is time-consuming and inadequate for the pace of operations in today's Army. We need to come up with processes and systems that can be linked together, resulting in data that Is not only helpful to leaders, but also timely and accurate.

Compounding the difficulty of standardization, not every system has all the required fields to provide answers to the questions being asked. For instance, GFEBS does not have a geographic data field for where a transaction took place. Therefore, It is sometimes tough to discern where and how a unit is spending the money. Consider this: it could be as simple as pulling a report and running it through a mapping visualization.

When GFEBS was fielded, there were several processes that were left up to each command to choose for themselves how to resolve. For example, the hierarchy structure for how financial transactions are tracked In GFEBS is not standardized. Each command has a different philosophy and implementation plan for the Work Breakdown Structure to track costs and manage funds.

Unless the data architecture becomes standardized, the Army will have to continue creating customized tools and reports. This means the Army must continue to provide the education, training, and policies that enable financial managers to access the data, reformat, make connections, validate and clean, analyze, and visualize and communicate the knowledge to leaders, so we can provide meaningful informational support for our commanders. Only by doing those actions will financial managers be able to provide timely and relevant decision support to commanders.

Conclusions

Data analytics is the convergence of math, statistics, and computer coding, all laid over a foundation of business expertise. It enables a real world understanding of the data relationships, which, when coupled with visualization techniques, makes the information easy to understand. The resulting insight Is helpful to a decision maker's ability to weigh risks and make tough choices. Success can be measured in the timeliness of the data to support decisions, clarity of communication, and accuracy of conclusions.

Timely and accurate data analysis has the potential to create a high amount of value added to commanders and staffs across the Army. There are some areas within business processes and system structures that could be standardized to facilitate easier analysis of Information. The data analytics process is a team effort. Some of the roles and responsibilities for taking raw data and turning it into actionable knowledge belong to the financial management (FM) community; some are outside FM. Financial managers must partner with both the non-FM data engineers, whose focus is more on the raw data and system architecture, as well as the operational leaders and commanders, who provide training and mission assessments. FM leaders must interject their functional expertise into the flow of data to help interpret the financial effect of operations. Doing this would enable Army leaders to make more timely and accurate, resource informed decisions.

MAJ(P) Kevin Linzey, CDFM-A

MAJ(P) Kevin Linzey, CDFM-A, is an instructor at the US Army Financial Management School, Ft. Jackson, SC. He has been a financial manager at the tactical level as well as 2, 3, and 4 star commands from Korea to Germany, to Afghanistan. As part Of ASMC, he served previously as the Vice President of the Greater Stuttgart Chapter in Stuttgart, Germany.

Caption: Diagram 1: Drew Conway Venn Diagram

Caption: Diagram 2: Stephan Kolassa Venn Diagram

Caption: Diagram 3: ERP System
COPYRIGHT 2019 American Society of Military Comptrollers
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2019 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Linzey, Kevin
Publication:Armed Forces Comptroller
Date:Mar 22, 2019
Words:2202
Previous Article:Translating Between The Auditor and the Field Perspective.
Next Article:Fiscal Year 2018 Audit: United States Coast Guard Continues to Improve and Solidify Results.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |