Teach students to address accounting questions using data analytics

Many employers now want accounting hires to develop an analytics mindset. An analytics mindset starts with understanding the importance of analytics and data to the accounting profession, but also includes asking the right questions, accessing appropriate data, performing analytics techniques and statistical analysis, and interpreting the results for stakeholders.

Though data analytics and the analytics mindset may be new, instructors should keep in mind that the students will still be using it to accomplish the primary goal of accounting — providing high-quality and timely information to decision-makers. To reach that goal, accountants have always been asking questions. Answering those questions has not always been easy. However, with the new data analytics techniques and tools available, it is easier for students and practitioners to answer questions that accountants have asked for years, as well as address new questions.

Traditionally, accounting curricula focused on compiling, formatting, and auditing data, but not on data analytics. Data analytics allows us to enrich the curriculum by answering traditional accounting questions with data. Then, based on those answers, users can refine the questions and utilize new and better data and techniques to provide answers to those new questions. This process reflects the iterative nature of data analytics.

Accounting questions fall into five broad categories with the newest category, adaptive and autonomous analytics, providing opportunities to answer questions that could not be answered before. Below are examples of data analytics techniques that students can use to answer each type of question:

• Questions of “what happened?” or “what is happening?” are addressed with descriptive analytics. Descriptive analytics summarizes the past, helping decision-makers to understand what happened in an easy-to-understand format. Techniques include simple descriptive statistics (e.g., averages, counts, horizontal analysis, maximums, minimums, pivot tables, ratios, standard deviations, and vertical analysis) as well as simple visualizations (e.g., histograms, bar charts, and pie charts).

• Questions of “why did something happen?” or “why it is different from something we expected?” are addressed with diagnostic analytics. Diagnostic analytics identifies the root cause of an outlier/anomaly or a phenomenon needing additional analysis. Techniques include using Benford’s Law, correlations, duplicate analysis, drill-downs, fuzzy matching, hypothesis tests, internal controls tests, reconciliations, and variance analysis. Diagnostic analytics often combines multiple datasets to help decision-makers evaluate the relations, patterns, and linkages between variables that might be related to an anomaly.

• The question of “will it happen in the future?” is addressed with predictive analytics. Predictive analytics predicts future performance using historical data and a user-specified algorithm, which is a set of instructions on how to combine data. It describes things that possibly might happen. Techniques include classifications, regression analysis, decision trees, and time-series/forecast analysis.

• The question of “what shall we do based on what we expect will happen?” is addressed by prescriptive analytics. Prescriptive analytics predicts the future and recommends a course of action. Techniques include cash-flow analysis, goal-seek analysis, marginal/incremental analysis, optimization analysis, scenario analysis, and sensitivity analysis.

• Questions of “how does the system adapt to changes?” and “how can we continuously run analytic solutions?” are addressed with adaptive and autonomous analytics. Adaptive and autonomous analytics mimics human capabilities, such as judgment and reasoning, creating insights from structured and unstructured data. It uses self-learning algorithms that continuously learn and rapidly modify themselves to optimize performance as new data and outcomes are created. Techniques include artificial intelligence applications such as machine learning, artificial neural networks, and natural language processing.

The table, “Matching Analytics Techniques to Accounting Questions,” presents sample questions by branch of accounting in each of the five categories of analytics. As students categorize their questions, they can determine the appropriate analytics tools and techniques to address those questions.

To move the profession forward, the accounting academy needs to embrace data analytics throughout the accounting curriculum. Incorporating all five categories of analytics into the curriculum should help develop students who not only can ask questions, but also answer those questions with data using appropriate tools and techniques. Data analytics helps students ask and answer deeper questions, providing actionable intelligence for organizations.

— Vernon J. Richardson is distinguished professor of accounting and the W. Glezen Chair in the Sam M. Walton College of Business at the University of Arkansas in Fayetteville; Marcia Weidenmier Watson, Ph.D., is the Jesse H. Jones Professor of Accounting in the Michael Neidorff School of Business at Trinity University in San Antonio, Texas. To comment on this article or to suggest an idea for another article, contact Courtney Vien at Courtney.Vien@aicpa-cima.com.