Seeing in All Directions Requires Business Analytics and Intelligence

Businesses that are taking advantage of the ability to access massive amounts of data in real-time are discovering that they don't need to choose between understanding the past and predicting the future. Today, business analytics and intelligence are helping companies both understand the past, and make predictions about the future.

Business intelligence tools and practices have evolved to the point where they are readily accessible to enterprises of all sizes. There is no shortage of technology to choose from, including tools for data collection, reporting, and visualization. What these solutions do not offer, however, is predictive analytics related to issues often identified by business intelligence activities.

Some solutions, says  Beverly Wright, executive director of the Business Analytics Center at Georgia Tech's Scheller College of Business, "describe BI as offering insight into what has happened with the broader field of analytics — and particularly advanced analytics — anticipating what will happen under various future scenarios."

What Happened vs. What Will Happen

Business intelligence is like a rearview mirror, it's a good tool for understanding what happened in the past. Enterprises need business intelligence to understand the details of key business measures about sales, inventory, and operational costs. For example, a retail analyst may want to know about sales of various products in different regions and across stores. In some cases, the analyst may want information about the number of products sold in each store over some period of time. This kind of information is useful for planning inventory or when considering financing to replenish stock.

In other cases, an analyst may be more interested in comparative information, such as the percent increase in sales of each product by store. Comparative information helps identify problem areas, such as stores that are underperforming or products with decreasing market demand. It is important to identify these problem areas as early as possible so the business can make changes and mitigate the risk of losses. Course correction requires more than just spotting monthly variances.

The Past Is Prologue

While business intelligence reporting can help identify problems after they arise, business analytics can predict future key measures using historical data. Eric Siegel, author of Predictive Analytics, describes this process as "the regimented discipline of using what we do know—in the form of data—to place increasingly accurate odds on what's coming next."

Predictive analytics is also useful for macro-level decision making, such as predicting volumes of sales of various products in the next quarter. These forward-looking techniques are helpful for micro-level predictions, such as predicting products a customer may be interested in. In fact, one of the values of predictive analytics is that it forms the foundation for offering personalized recommendations to customers.

Prediction is the New Expert Opinion

Before the advent of predictive analytics, an analyst would use information from business intelligence reports to predict the future. An inexperienced analyst may formulate projections based on oversimplified assumptions, like next month's sales should be the same as this month's sales, assuming no significant changes have occurred in the market. A more experienced analyst would know that simplified assumptions like that can lead to poor quality decision making.

For example, seasonality is an important factor in retail. For some businesses, the end of the year holiday season is a time of peak sales. Other businesses, may have different seasonal patterns. Some patterns may change on a day-to-day basis. Restaurants, for example, may have higher sales on weekends than on weekdays. Agriculture and construction may have periods that span months. While an experienced analyst may be able to reason about these and other factors, not all analysts may have this knowledge.

Predictive analytics democratizes the benefits of knowledge derived from historical data. According to analytics expert Chris Brahm from Bain's global advanced analytics practice, there "are a number of players who are entering the market who are providing analytics for managers and front-line workers that go beyond what traditional BI did."

For example, forward-looking techniques include multiple types of regression that are used to make predictions about one variable based on the values of other variables.

One especially important type of predictive model is RFM analysis, which stands for recency, frequency, and monetary analysis. These models use a combination of how recently a customer made a purchase (recency), how often they purchase (frequency) and how much they spend (monetary). These three variables are used as inputs to various kinds of regression and other predictive models to estimate when that customer will purchase again as well as the value of their purchases.

Business intelligence tools and techniques have proven their value and will continue to be useful in the future. Predictive analytics now complement historical reporting and provide the means to look forward as well as backward through the rear view mirror of business intelligence tools. The combination of these two practices are the foundation of new business value.

Tim O'Reilly, CEO and founder of O'Reilly Media and open source advocate, notes companies that have adapted these are “uncomfortable bringing so much attention to this because it is at the heart of their competitive advantage. Data are the coin of the realm. They have a big lead over other companies that do not 'get' this.'"

In our over 40 years of existence, Keyence has seen staggering success due to our data-guided business practices. Join us on this journey, if you're interested in becoming a data master, stay tuned, we have lots more coming. Download the 3 Key Features digital brochure and learn more about how you can use A.I. to drastically change your data-driven strategy.

LATEST POST : How Decentralizing Data Makes Analytics a Team Sport