A core task of management accountants is to provide decision-makers with the right information, at the right time, in a suitable format for at-a-glance key insights. New technologies, such as predictive and prescriptive analytics, artificial intelligence, and real-time data processing, allow data quality to improve and trigger analytical reasoning and data-driven decision-making.

However, these new technologies can also potentially be highly complex and overload the decision-maker, even deteriorating decision quality. For the full potential of digitalisation to be unlocked, management accountants need to pay particular attention to five steps (for an overview, see the graphic, "5 Steps to Successfully Digitalise Reporting").

Step 1: Use BI software to support data analysis and reporting

Especially for small and medium-size companies, Microsoft Excel is the tool of choice when it comes to data preparation and reporting. However, Excel as the sole management information tool can be seen as a barrier when it comes to the inclusion of big data into the reporting routine, due to its limited integration and evaluation options as well as its low data storage capacity. Additionally, Excel offers limited options for data visualisation (especially when it comes to newer forms of visualisation designed to present big data as described in this article) and only a limited number of interaction techniques (eg, filtering, sorting).

Fortunately, the barriers for data analysis and reporting are decreasing, and the proportion of companies relying on business intelligence (BI) software for managerial decision-making is rising. When choosing the right BI tool, the company's needs and its existing IT infrastructure should be taken into account. Two tools — Microsoft Power BI and Tableau — have become market leaders (identified by analysts Gartner and BARC for many consecutive years). Based on our research, we provide indications of when to choose which software product (see the table, "Benefits of Leading BI Tools").

Step 2: Use new data sources to increase understanding

The inclusion and combination of various data sources such as the internet of things or social media for a new and better understanding of business models, products, and processes is the most important and the most challenging task when it comes to big data use. Data that has been stored from various sources and occasions needs to be put to use, and "value" needs to be generated. Especially in accounting, the identification of new and better growth or sales predictors could potentially change the way decisions are made today.

Nonetheless, it is not recommended to include every data source available into a pool and see if a trend or correlation can be detected to derive better predictors. Data sources should be carefully chosen with respect to their potential to contribute to the company's strategic goals. For example, while for stores in a shopping mall low customer flow simply follows good weather, financial service providers need to draw up customer profiles allowing them to assess and predict a customer's potential for default by combining personal information (eg, credit records and information on their income and fixed costs) and general market information (eg, information on industries, interest rates).

Step 3: Identify new insights and correlations from big data

Earlier identification of indicators gives management more decision-making leeway. When algorithms are highly complex, or for an initial analysis of the data with a limited number of variables, visual analytics can be used, as they present more intuitive ways to understand large amounts of data. Their inherent goal is to present the whole dataset within one frame to uncover insights that might otherwise be missed.

Step 4: Integrate new insights into reporting to enhance decision-making

Steps 2 and 3 identify early-warning indicators or new insights in the company's business logic. Visuals such as a parallel coordinates plot that can present multiple attributes in one chart or the heat map using example data from the wine industry presented in the sidebar "Interactive Visualisation for Assessment of Distribution and Correlation" are necessary for these tasks but should not be used in end-user reporting. High analytical skills are needed to derive information from a parallel coordinates plot, placing an unnecessary burden on the decision-maker. This task should be performed by trained management accountants or data scientists, and the understanding gained should be transformed to a simpler format so that the decision-maker can act on it (eg, if Y [order intake] predicts X [revenue], then reporting should focus on Y, to increase the manager's ability to take action by indicating problems very early on).

Step 5: Use user-friendly design and appropriate visualisation types for interactive dashboards

The optimal tool for managerial decision-making is a dashboard. One definition of a dashboard — by authors Ogan M. Yigitbasioglu and Oana Velcu-Laitinen — is a "visual and interactive performance management tool that displays on a single screen the most important information needed to achieve one or several individual and/or organisational goals, allowing the user to identify, explore, and communicate problem areas that need corrective action". Thus, a dashboard should only feature standard visuals known to management such as bar, column, waterfall, or line charts, which will clearly emphasise deviations from the plan that indicate a need for course-correction. Management can only take appropriate action if deviations stand out early on.

Visual choices should follow general design recommendations, such as in Hichert and Faisst's International Business Communication Standards, but also represent the data structure and allow for adjustments by users (such as filtering, panning, and drill-through). In addition to bar, column, line, waterfall, and scatterplot charts, visualisation types such as the sunburst, which is a development of a pie chart, and the choropleth presented in the sidebar "Useful Interactive Visualisation in a Dashboard" are helpful in a dashboard display. They transmit information in an understandable format while condensing a lot of data either in a hierarchy or by location.

Putting the data to work

Digitalisation and big data will remain mere buzzwords if management accountants cannot transform data from various sources into understandable information. Our five-step process shows how to successfully integrate new data sources into the reporting routine and satisfy the demand for self-service information and dashboarding. It also emphasises the need for the right tools for easy integration of various sources and for newer forms of visualisation and interaction.

Interactive visualisation for assessment of distribution and correlation

A heat map represents data through a colour-coded system, and the format resembles a table, as it consists of rows and columns. Although no numerical values are presented, they can be inferred by colour: Very dark and very light colours are an indication of extremities.

Useful interactive visualisation in a dashboard

This choropleth visualisation is one of the most used geographic visualisation types. It applies the concept of the heat map and transforms a colour code onto a map, conveying different regions with different granularity (countries, states, or districts). Colours in different shades are used to distinguish between regions with high and low values. If a region attracts interest, a deeper analysis should be made possible by jumping into the next hierarchy level, and if more choropleths are used to visualise different variables, interactions should be linked to view the changes induced simultaneously in other variables.

More information on how to use this visualisation using wine industry example data is available here.

5 steps to successfully digitalise reporting

1. Use BI software to support data analysis and reporting

• Use BI tools (beyond Excel) to expand analysis and reporting capabilities, especially when you want to include data-intense visualisations and advanced interaction.

• Choose a software suited to your needs; market leaders are Microsoft Power BI and Tableau.

2. Use new data sources to increase understanding

• Include data sources that add value to achieve strategic goals.

• Do not shy away from unstructured data sources (eg, social media, internet of things) that require high effort.

3. Identify new insights and correlations from big data

• Use algorithms and intelligent software to sort data.

• Use visual analytics (big data visualisations) wherever algorithms are highly complex.

4. Integrate new insights into reporting to enhance decision-making

• Create new and relevant KPIs from the obtained insight, related to the company's strategic goals.

• Use BI support to enable self-service.

5. Use user-friendly design and appropriate visualisation types for interactive dashboards

• Rely on accepted design guidelines.

• Choose easy-to-understand and standard visuals to enhance recognition, acceptance, and understanding.

• Choose new visualisation types where necessary.

Benefits of leading BI tools

When to favour Power BI

If you are a novice user, Power BI is easy to understand and use compared to Tableau, as it works in a similar fashion to all other Microsoft products.

If your dataset is changing a lot (eg, adding new dimensions for analysis), including new tables using primary and secondary keys is easier than doing so in Tableau, as it works with only one data table.

Creating a dashboard and arranging visuals on the page is more efficient and intuitive than with Tableau, which asks you to create the visuals first (on separate pages) before being able to integrate them onto one dashboard.

If you have a limited budget for BI support, then Power BI is the better choice, as it is integrated into microsoft 365. Access is easy, and low costs are guaranteed.

When to favour Tableau

Tableau has a different process — the user first selects the data and then gets a selection of ready-to-use visualisations based on the dataset. The options are supported by research, as the program was designed by visualisation researchers.

Customising design characteristics according to the decision-maker's needs and adjusting to corporate identity requirements is easier than in Power BI

Using interaction (cross-visual filters, drill-down options, tool-tip information, and including charts for further analysis) is easily and more intuitively implemented in Tableau compared to power BI.

Heimo Losbichler, Ph.D., is professor of controlling, and Lisa Perkhofer, Ph.D., is research project manager, both at the University of Applied Sciences Upper Austria. To comment on this article or to suggest an idea for another article, contact Oliver Rowe at Oliver.Rowe@aicpa-cima.com.

LEARNING RESOURCES

Data Visualization Certificate

Learn fundamental types of visualisation techniques and explore a sampling of the most common reporting tools for producing meaningful visualisations. Find this course in the AICPA store and in the CGMA store.

COURSE

Microsoft Power BI: Power BI Series

This nine-part self-study online series will help you develop the skills necessary to use Microsoft Power BI tools, which will allow you to perform enhanced data analysis.

COURSE

Power BI — Use PBI Desktop to Create Reports and Dashboards

Power BI Desktop is a new standalone tool that adds visualisation to your data. This webcast focuses on learning the Power BI Desktop interface, connecting to data sources, and creating the visualisations.

VIDEO

Power BI — Introduction to MS Power BI Tools

This webcast provides an overview of the various Microsoft Power BI tools and should be your first step in learning the complete set of Power BI tools. By the end of this webcast you will have a thorough understanding of what each tool can accomplish.

VIDEO