Big Data

Business Analytics: Moving From Descriptive To Predictive Analytics

Bill Schmarzo By Bill Schmarzo CTO, Dell EMC Services (aka “Dean of Big Data”) January 9, 2014

I recently met with two organizations that had very impressive business intelligence (BI) implementations. However, they struggled to understand how predictive analytics could augment their existing BI capabilities. The purpose of this blog is to clarify the differences between business intelligence (a.k.a. descriptive analytics) and predictive analytics (a subset of data science) and highlight their complementary natures. In particular, you will learn that predictive analytics:

  1. Can provide more insightful and actionable answers to the organization’s most common questions, (“Who are my most valuable customers?” “What are my most important products?” “What are my most successful campaigns?”) than those generated by business intelligence alone.[1]
  2. Can provide more future-looking answers and recommendations to questions that cannot be addressed at all by business intelligence.

Business Intelligence (Descriptive Analytics) Versus Predictive Analytics

Figure 1 is a pretty common way to view the worlds of business intelligence and predictive analytics. Business intelligence is the world of descriptive analytics:  retrospective analysis that provides a rearview mirror view on the business—reporting on what happened and what is currently happening. Predictive analytics is forward-looking analysis:  providing future-looking insights on the business—predicting what is likely to happen (usually associated with a probability) and why it’s likely to happen.

Figure 1: Business Intelligence versus Predictive Analytics

Figure 1: Business Intelligence versus Predictive Analytics

Business intelligence looks for trends at the macro or aggregated levels of the business, and then drills up, down, or across the data to identify areas of under- and over-performance. Areas may include:  geography, time, products, customers, stores, partners, campaigns, or other business dimensions.

Business Intelligence is about descriptive analytics (or looking at what happened), slicing-and-dicing across dimensional models with massive dissemination to all business users.

Predictive analytics, on the other hand, builds analytic models at the lowest levels of the business—at the individual customer, product, campaign, store, and device levels—and looks for predictable behaviors, propensities, and business rules (as can be expressed by an analytic or mathematical formula) that can be used to predict the likelihood of certain behaviors and actions[2].

Predictive analytics is about finding and quantifying hidden patterns in the data using complex mathematical models that can be used to predict future outcomes.

What’s Different About BI and Predictive Analytics? The Answers…

Maybe the easiest way to understand the differences between business intelligence and predictive analytics is to look at the answers they can generate. For example, business intelligence allows you to answer questions about the demographics or characteristics of your customers, products, stores, etc., and answer questions about the performance of your business across a number of different dimensions (see Figure 2).

Figure 2: Descriptive Analytics (Business Intelligence)

Figure 2: Descriptive Analytics (Business Intelligence)

On the other hand, predictive analytics allow organizations to go beyond the answers generated by BI by providing more predictive answers and recommendations to many of the same questions (see Figure 3).

Figure 3: Predictive Analytics Answers And Recommendations

Figure 3: Predictive Analytics Answers And Recommendations

Sample BI versus Predictive Analytics Answers

Below is a more detailed list of the types of answers that predictive analytics can answer that not only enriches the answers provided by business intelligence, but also provides answers to questions that business intelligence could not previously address.

1 9 14 Bill Table 1

Summary

Predictive analytics takes the questions that business intelligence is answering to the next level, moving from a retrospective set of answers to a set of answers focused on predicting performance and prescribing specific actions or recommendations.

For example, if we change the three key business questions that we asked earlier (most valuable customers, most important products, most successful campaigns) to a future tense, then you can see that we need a predictive analytics approach that is completely different from the conventional BI approach (see table below).

1 9 14 Bill Table 2

No matter what, you still need business intelligence to know what really happened in the past, but you also need predictive analytics to optimize your resources as you look to make decisions and take actions for the future.


[1] These three questions can be answered with more granularity, relevance, and actionability through the integration of currently non-accessible detailed transactional “dark” data (now available in much larger volume and detail) and new sources of internal (consumer comments, work orders, technician notes) and external (social media, blogs, newsfeeds) unstructured data.

[2] Per Dr. Pedro Desouza, the concepts of Big Data and Predictive Analytics get used interchangeably. They shouldn’t. “Big Data” provides data at a very low level of granularity with enough points that can still be statistically meaningful. Before the era of Big Data, business analysts needed to aggregate data in order to have enough points to predict with reasonable confidence. Now we can develop predictions at the individual customer level, without the need to aggregate the data, for example, to the store level.

Bill Schmarzo

About Bill Schmarzo


CTO, Dell EMC Services (aka “Dean of Big Data”)

Bill Schmarzo, author of “Big Data: Understanding How Data Powers Big Business” and “Big Data MBA: Driving Business Strategies with Data Science”, is responsible for setting the strategy and defining the Big Data service offerings and capabilities for Dell EMC Services Big Data Practice. As the CTO for the Big Data Practice, he is responsible for working with organizations to help them identify where and how to start their big data journeys. He’s written several white papers, is an avid blogger and is a frequent speaker on the use of Big Data and data science to power the organization’s key business initiatives. He is a University of San Francisco School of Management (SOM) Executive Fellow where he teaches the “Big Data MBA” course. Bill was ranked as #15 Big Data Influencer by Onalytica.

Bill has over three decades of experience in data warehousing, BI and analytics. Bill authored Dell EMC’s Vision Workshop methodology that links an organization’s strategic business initiatives with their supporting data and analytic requirements, and co-authored with Ralph Kimball a series of articles on analytic applications. Bill has served on The Data Warehouse Institute’s faculty as the head of the analytic applications curriculum.

Previously, Bill was the vice president of Analytics at Yahoo where he was responsible for the development of Yahoo’s Advertiser and Website analytics products, including the delivery of “actionable insights” through a holistic user experience. Before that, Bill oversaw the Analytic Applications business unit at Business Objects, including the development, marketing and sales of their industry-defining analytic applications.

Bill holds a masters degree in Business Administration from the University of Iowa and a Bachelor of Science degree in Mathematics, Computer Science and Business Administration from Coe College.

Read More

Join the Conversation

Our Team becomes stronger with every person who adds to the conversation. So please join the conversation. Comment on our posts and share!

Leave a Reply

Your email address will not be published. Required fields are marked *

12 thoughts on “Business Analytics: Moving From Descriptive To Predictive Analytics

  1. Bill, I think this is a great framework for looking at BI and Predictive Analytics. I think that you make an important point that it’s not a matter of moving from one to the other but employing both really well for competitive advantage.

  2. MIke, totally agree! It’s not an either/or discussion and I didn’t mean to position it as such. It’s a matter of using both at the appropriate times to achieve, as you say so well, competitive advantage. Thanks for reading the blogs!!

  3. As stated above, as more organizations strive to implement advanced analytical techniques, the need for solid Business Intelligence remains. Adding to the discussion here, the first graphic in the blog post above is a modified version of content developed and released as part of EMC’s Data Science & Big Data Analytics curriculum, first introduced in late 2011. The original graphic displays the topmost group of techniques as Data Science, rather than Predictive Analytics, which the blog here points out can differ from Data Science. It is included in the introductory module of the five-day course that EMC offers for people who would like to become Data Scientists https://education.emc.com/guest/campaign/data_science.aspx, for those who may be interested in additional materials.

    In addition, the original version of this graphic is also part of the free, 90-minute overview module developed by EMC Education Services, for those who want to learn what Big Data is, and how to begin taking advantage of it at their companies. The free module may be viewed here: https://education.emc.com/bigdata

    The topmost graphic above shows a way to distinguish Business Intelligence and Data Science. As the post mentions, Predictive Analytics and Data Science are not necessarily the same thing, as Data Science is generally broader and encompasses a larger group of techniques spanning multiple disciplines. Earlier iterations of this graphic, which correctly portray this distinction, may be seen as part of this presentation, if people wish to see this point as part of a larger context: http://www.slideshare.net/emcacademics/iccdba-conference-8-feb-2013-david-dietrich .

  4. Allen, prescriptive analytics (that is, analytics that tell the user specifically what to do, like a recommendation engine) fall within the Predictive Analytics category. If you can imagine a continuum of analytics, I’d have descriptive analytics on the far left, predictive analytics in the middle, and prescriptive analytics to the far right. Part of the power of analytics is to get to the point where we can make specific, actionable recommendations to our users … and that’s prescriptive analytics.

    Thanks for asking and for reading the blog!

  5. To take the analytics discussion one step further, BI provides descriptive insight into what happened in the past. Predictive analytics predicts trends and behaviors in the future. I would argue that in order to deliver true and immediate value to most businesses, analytics must be “prescriptive.” Prescriptive analytics involve delivering information to frontline business decision-makers in a way that is insightful and actionable and enables them to improve decisions being made, by making them more informed. While this may seem obvious, many analytics solutions today do not embrace this important capability. For example, they might present refined data reports, but still place a significant burden of analysis on the recipient.

    Prescriptive analytics deliver digestible information in a timely manner, making it absolutely clear to the decision-maker that some particular action needs to be taken. An example would be the difference between delivering:
    – A list of all purchases made over the corporate maximum (BI)
    – A list of employees who are likely, based on past behavior, to exhibit out-of-policy purchase behavior (Predictive)
    – Actionable information about a specific employee whose spending behavior the last two weeks of employment was out-of-character – though not out-of-policy – when compared with a history of spending (Prescriptive)

    Each of these deliverables has value to the business, but for everyday business decision makers on the front lines, prescriptive analytics allows them to be smarter and drive immediate value from analytics on a day-to-day basis.

  6. Patrick, spot on about how the world needs to migrate to prescriptive analytics. It’ll dramatically change the way we deliver information and insights to our users – both internal users and external customers. And I love the example!

    Thanks for sharing!

  7. Bill,

    Great article! One other element of the descriptive vs. predictive discussion worth noting is what we discussed last week; the value of context. Descriptive analytic are very focused upon questions of “what” rather that of why. Answers to “what” are readily found in structured data, and that’s what most organizations are comfortable with.

    However, richness and context are found mostly in unstructured data. This context helps to answer questions of “why” which leads directly to predictive analysis. If I understand “why” I have real power. Hence, a huge step in moving from descriptive analytics and towards predictive is to start analyzing unstructured data in conjunction with structured, and to start asking “why” things happened.

  8. Very interesting. Do you have any published journal or conference papers on this big data on hospital readmission? Or do you have contact info for one of authors: Dr. Pedro Desouza that I can inquire some information?

  9. The way I see it, BI tells you how things ARE while predictive analytics tells you how things WILL BE. BI shows you past trends while predictive analytics takes those trends and shows how that will effect you in the future.