The Four Disciplines of Analytics
Business Analytics (BA) has come a long way since its first recorded application in the late 19th century. While technologies and techniques have evolved over time, the general purpose of BA has stayed the same. In broad terms the goal is to address the question, “How can we improve our business using our own data”. In the early 1900s Henry Ford hired Frederick Taylor—who is credited as one of the earliest implementers in this field—to consult on Ford’s manufacturing process (Team ASM IBMR, 2020). Now, more than 110 years after Taylor and Ford joined forces (Saylor, 2013), the business intelligence market is predicted to be a $40.5 billion industry by 2025 (Whiting, 2020). Let’s review the four basic disciplines of Business Analytics to get a better sense of how organizations are capitalizing on their business data.
The Four Disciplines
Descriptive Analytics
How has revenue grown over the past five years? What was our employee retention rate last year? What percentage of new clients are opting for paperless statements? These are rather simple questions, sure, but we need Descriptive Analytics to answer them. In order to know how our business has been performing both historically and in near real-time, we need to collect, summarize, and report on as much data as we can. Techniques generally include data aggregation and data mining (Mehta, 2017). In exchange for rather mundane work we get insights into our strengths and areas of opportunity. This knowledge enables businesses to identify key questions to attempt to answer using diagnostic techniques.
Diagnostic Analytics
Why did our revenue drop three years ago when it increased in each of the other last five years? What internal and external factors might explain why our retention rate fell below our stated goal last year? Are there correlations between say, basic demographics and the likelihood of a customer opting for paperless statements? Diagnostic Analytics help answer these types of questions. Generally, we want to know why something happened or didn’t happen. Pinpointing correlations or better yet cause-and-effect relationships helps us strategize ways to improve in the future.
Predictive Analytics
What do we expect company revenue to be over the next three years? What do we believe our employee retention rate will be over the next 12 months? Do we think the likelihood of a customer opting for paperless statements will steadily increase over time? We employ Predictive Analytics techniques to address these kinds of “based on our historical data, what should we anticipate happening next” types of questions. Techniques typically include statistical modeling such as forecasting and time series analysis. Now that we have a sense of where for example our Key Performance Indicators (KPIs) are headed going forward, what can we do to influence the expected outcomes?
Prescriptive Analytics
Let’s focus on just one of our example questions: what percentage of new clients opt in to receiving paperless statements? By sending electronic documents instead of paper versions we cut costs, reduce transit time, and help save the environment. Sounds like a win-win all the way around. So how can we increase the adoption rate from 50% to 55% next year? Thanks to the results of our descriptive, diagnostic, and predictive analyses we might know that the rate has gone up steadily by an additional 2% over the past five years, so that gets us to 52%. We also know that customers under the age of 40 are twice as likely to elect the paperless option than clients aged 41 and older. Based on this information, we can brainstorm ideas to bring the rate up. We run millions of scenarios that include our ideas to see what the new outcomes could be. Ultimately our optimizations and models show us that three out of our four possible interventions can help us hit 55% while the fourth is superfluous. Not only that, but our simulations constantly receive feedback and may tell us to change course over time to help us reach our goal. This is the power of Prescriptive Analytics.
Each of the four disciplines contributes to our goal of improving our business using our own data. Many of our KPIs are timebound and therefore encourage us to iterate through each discipline, adjusting and improving as we go. Every day more companies are adopting Business Analytics practices. At Northwestern Analytics, we are thrilled to be helping businesses get the most out of their data, no matter how mature their data operations are.
References
- Mehta, A. (2017, October 13). Four types of business analytics to know. Analytics Insight. https://www.analyticsinsight.net/four-types-of-business-analytics-to-know/
- Saylor.org (2013, August). Scientific management theory and the Ford motor company. Saylor.org. https://resources.saylor.org/wwwresources/archived/site/wp-content/uploads/2013/08/Saylor.orgs-Scientific-Management-Theory-and-the-Ford-Motor-Company.pdf
- Team ASM IBMR (2020, May 6). The history of the evolution of business analytics. Institute of Business Management & Research. https://www.asmibmr.edu.in/blog/the-history-of-the-evolution-of-business-analytics/
- Whiting, R. (2020, April 27). The coolest business analytics companies of the 2020 big data 100. CRN. https://www.crn.com/slide-shows/cloud/the-coolest-business-analytics-companies-of-the-2020-big-data-100