Extreme Uncertainty Drives the Need for Advanced Business Analytics

Global pandemics, regional conflicts, economic turmoil, internal and external crises: all drive extreme uncertainty in the business world.  Extreme uncertainty can drive a business to a sure death. 

According to McKinsey, “in a serious crisis, uncertainty can reach extreme levels, and the normal way of working becomes overstrained. At such times traditional management operating models rarely prove adequate, and organizations with inadequate processes can quickly find themselves facing existential threats” (McKinsey: When Nothing is Normal: Managing in Extreme Uncertainty).

Extreme uncertainty is usually accompanied by its evil twin, too many choices, which can be as equally challenging.  Having too many choices, especially during times of extreme uncertainty, makes decision-making at all levels (strategic, tactical, and operational) particularly challenging.  Businesses simply cannot evaluate all possible scenarios in the time needed to make and act on a decision.

To the rescue: Business Analytics.  Business Analytics “is the scientific process of transforming data into insight for making better decisions” (INFORMS:  Institute for Operations Research and the Management Sciences).  Companies leveraging data and advanced business analytics are 23 times more likely to acquire customers and 19 more times likely to remain profitable versus competitors (McKinsey: Five Facts: How Customer Analytics Boosts Corporate Performance).  Business Analytics cover a broad spectrum of decision-enabling capabilities with techniques that are classified in three wide-ranging categories.

Three Categories of Business Analytics

1) Descriptive Analytics

Descriptive Analytics drive decisions based upon describing what has happened in the past.  Techniques for performing Descriptive Analytics include:

Traditional Reporting – static reports, financial statements, enterprise systems output.

Descriptive Statistics – data summarization and characteristics, collective properties of a data set (frequency distribution, mean/median/mode, measures of variability).

Data Visualization – visual representation of data showing patterns, trends, and outliers using charts, maps, and graphs.

Dashboards – interactive business intelligence utilizing visualization, ad hoc queries, pivot tables.

Data Querying – a question or query from a database, often via real-time data access allowing combinations and calculations.

2) Predictive Analytics

Predictive Analytics leverage models constructed from past data to predict the future, often used to determine the impact of changes to one variable on another.  Predictive Analytics models utilize probability to provide forecasts or predictions to address uncertainty.  Techniques for performing Predictive Analytics include:

Linear Regression – used to predict the value of a variable based on the value of another variable (e.g., factors that drive the price of a home).

Classification – versus Linear Regression, is a process of categorizing a set of data into classes (e.g., customer purchasing behavior, customer churn prediction, credit card fraud detection).

Logistic Regression – used for binary classification of outcomes of categorical variables (e.g., is an email spam or not, whether or not a political candidate will win an election).

Time-Series Analysis and Forecasting – allow predictions for calculations such as marketing plans, cash flows, and purchasing.

Game Theory – modeling of interactions among rational agents (e.g., salary negotiations, collective bargaining, product pricing decisions).

Simulation – leverages probability and statistics to model the impact of uncertainty on a decision (e.g., Monte Carlo Simulation for portfolio optimization, epidemiology, threat modeling).

3) Prescriptive Analytics

Prescriptive Analytics extend Predictive Analytics by providing recommendations to achieve goals.  Like Predictive Analytics, Prescriptive Analytics provide predictions but additionally provide a prescriptive decision (course of action to be taken) instead of simply predicting options.  Techniques for performing Prescriptive Analytics include:

Linear Programming – minimizing or maximizing a function within the bounds of known constraints (e.g., portfolio return on investment, production line scheduling, distribution routing).

Optimization – uses mathematical modeling to optimize supply, production, and distribution costs within a complex network (e.g., supply chain network, telecommunication network, transportation network).

Decision Analysis – uses payoff tables, decision trees and sensitivity analysis to identify monetary value outcomes and drive recommendations.

Conclusion

While contemporary businesses face continued extreme uncertainty, Business Analytics offer a unique toolset to reduce both uncertainty and the corresponding challenge of condensing too many choices.  Descriptive, Predictive, and Prescriptive Analytics provide the techniques for businesses to reduce uncertainty while simultaneously improving competitiveness.