Data Analytics: Transforming Data into Insights, Decisions, and Actions

By Dr. Pooya Tabesh—Substantial changes are underway in the business data analytics’ landscape. Digital data in the universe doubles every two years and the size of business-related data continues to grow exponentially. In line with these trends, the adoption rate of popular big data analytics tools, such as data mining, machine learning, and other artificial intelligence techniques, is rising in every industry[1]. In this regard, and according to a report by IBM, the demand for data scientists and analysts is projected to grow 30% over the next three years[2].

Many organizations have started to make significant capital investments in data analytics to obtain useful insights from data and transform the insights into decisions and actions that enhance business processes and outcomes [3]. In a broader sense, these actions generate new data points that can be re-analyzed for gaining further insights. This way, a self-perpetuating cycle of business data analytics (see Figure 1) can significantly benefit organizational decision-making processes and outcomes.

Figure 1: Business Data Analytics Cycle

A wide range of data analytics tools is available to organizations. However, organizational resources are limited and a hasty choice in acquiring costly tools, applications, or human know-how may not always guarantee success. Like other traditional business analysis tools, data analytics tools and systems should be deployed properly in order to create desired outcomes. That is, the effectiveness of these tools depends on how well they fit the problem domain at hand and how much they empower the organizations in addressing their most pressing challenges. In this respect, executives and managers who are familiar with internal and external organizational challenges are responsible for determining the data analytics strategy of organizations.  To this end, managers should be able to understand the basics of data analytics and their applications in order to effectively integrate them into existing business processes. A basic understanding of advanced data analytics tools enables executives and managers to ask the right questions and identify the right data analytics techniques that will provide valuable solutions or insights. According to an article published in McKinsey Quarterly, full exploitation of data analytics is impossible without senior and middle management’s involvement in the process[4].

In this blog post, the intention is not to discuss technical details of data analytics tools. Instead, the focus is on providing easy-to-understand explanations of two general approaches to transforming data into meaningful insights and actions. In addition, a brief non-exhaustive list of several data mining techniques relevant to each approach is provided. For readers interested in technical details, hyperlinks to additional online sources are integrated.

Applications of Data Analytics

Many different data analytics tools are now available which can be grouped into the following two categories:

  1. Descriptive tools: Understanding the past and describing the present

The first category of analytics tools help managers understand the current state of their business based on data from what has happened in the past. These methods address the ‘What happened?’ question by uncovering existing states or patterns at an aggregate level. In their simplest form, descriptive analytics tools produce reports and summaries based on historical data. These approaches could also include simple description of statistical realities in a dataset including but not limited to mean, variance, or correlation coefficients. For example, correlation analysis of data from a grocery store retail chain might reveal that total sales at grocery stores is positively related to average household income in the specific zip code in which they operate.

Furthermore, data mining descriptive analysis can help uncover hidden and potentially useful information related to business processes. For example, consider a large bank that needs to better understand its current customers in order to provide more specialized services to them. By using clustering data mining tools, the bank might be able to uncover distinct categories or groups of customers and their attributes that are otherwise unknown. Additionally, the patterns of events or associations between two specific variables could be discovered using data mining techniques. In this regard, association rule discovery techniques can help supermarkets estimate what products are more likely to be purchased together so that they could be placed near each other. For example, association rule discovery might reveal that, in a hypothetical supermarket, if customers buy onions there is an 80% probability that they will also buy tomatoes. Such insight enables store managers to determine the best location for each of the items. One final example of a descriptive tool is Sequential patterns discovery, which is a technique that reveals the sequence of events in the past. For instance, this technique could be used to identify the sequence of purchases made by customers and unveil unknown customer behaviors.

The descriptive insights are based on the past and provide a realistic snapshot of the current business situations. Data visualization techniques can be used to effectively summarize and present the descriptive insights to internal and external stakeholders.

The descriptive tools form a basis for more advanced analysis of data in order to transform insights into specific organizational decisions and actions. Descriptive tools are often the first to be used in data analytics projects and are complemented with predictive tools.

  1. Predictive tools: Predicting the future and providing advice

The predictive analytics tools can help managers predict probable future states, patterns or outcomes based on analysis of existing data. The predictive methods address the questions such as ‘What is likely to happen?’ or ‘What should we do next?’ among others. Stated differently, predictive models enable decision makers to predict the estimated value of a variable of interest using existing data. A variety of techniques are used to predict events or outcomes that are otherwise unknown to decision makers. For instance, a large bank can rely on a classification model to predict if an existing customer is likely to open a new savings account. Such a predictive model, that is developed based on an analysis of enormous amounts of data related to similar customers, may enable the bank to identify easy-to-convert customers and engage in targeted advertising activities to attract them. Regression analysis tools form another type of predictive model used for predicting the value of continuous numeric variables. For example, regression models can help predict future sales based on market conditions and other relevant historical data of customers and competitors. Another class of popular predictive tools is related to anomaly detection. These tools enable decision makers to identify outliers and unusual patterns in events or behaviors. As an example, anomaly detection tools are widely used in fraud detection by identifying fraudulent activity in credit card transactions.

Data analytics techniques could be combined and used for both description and prediction purposes. Therefore, descriptive and predictive tools are often developed and implemented together. Additionally, these tools can be evolved into decision support systems that provide advice for decision makers regarding different courses of action and their possible outcomes. It is worth pointing out that existing data analytics tools provide approximate predictions and are prone to estimation errors. Thus, these quantitative techniques should be supplemented with human judgment to form reliable decision support tools.

Concluding Remarks

Recent developments in data analytics have ignited a transformation in the practice of marketing, human resource management, supply chain management, operations, and finance, among other business functions[5],[6],[7]. In addressing these changes, the major challenge for managers remains to be the selection of appropriate analytics tools to apply to the relevant business problems and the commitment to support the business data analytics cycle to transform data into actions. These goals cannot be reached without first understanding the tools in the toolbox and their specific applications.

Pooya Tabesh, Ph.D.

Assistant Professor of Management

 

[1] https://hbr.org/cover-story/2017/07/the-business-of-artificial-intelligence

[2] https://www.ibm.com/analytics/us/en/technology/data-science/quant-crunch.html

[3] https://www.sciencedirect.com/science/article/pii/S014829631630488X

[4] https://www.mckinsey.com/business-functions/digital-mckinsey/our-insights/making-data-analytics-work-for-you-instead-of-the-other-way-around

[5] https://www.shrm.org/hr-today/news/hr-magazine/pages/1114-hr-finance-analytics.aspx

[6] https://www.forbes.com/sites/bernardmarr/2016/04/22/how-big-data-and-analytics-are-transforming-supply-chain-management/#417be4d039ad

[7] https://www.mckinsey.de/files/the-age-of-analytics-full-report.pdf

 

Share this Post
Translate »