PDF version of this report
You must have Adobe Acrobat reader to view, save, or print PDF files. The
reader is available for free
download.
Predictive Analytics
Copyright 2022, Faulkner Information Services. All
Rights Reserved.
Docid: 00018031
Publication Date: 2207
Publication Type: TUTORIAL
Preview
Predictive analytics is the art and science of using past data to predict
future trends or events. Predictive analytics belongs to a larger field
called data analytics, which is divided into four logical disciplines: (1)
descriptive analytics, which details what happened; (2) diagnostic
analytics, which details why it happened; (3) prescriptive analytics, which
details what should happen next; and, finally, (4) predictive analytics,
which forecasts what might happen in the future.
Report Contents:
Executive Summary
[return to top of this
report]
At its core, data processing is an exercise in time travel. For most of the
modern, i.e., electronic, era (from the 1950s to the early 1990s), data
processing was about traveling into the past. Private sector companies and
public sector agencies would collect large volumes of “transaction data”
like sales or tax records and employ algorithms to process that data into a
meaningful account of what happened yesterday or last week or last year
or any arbitrary past period.
Related Faulkner Reports |
Big Data Analytics Tutorial |
Data Analysis and Data Mining Tutorial |
Digital Twin Tutorial |
While knowing “what happened” in the past is important, knowing “what
might happen” in the future can be even more valuable, enabling managers
to:
- Avoid potential problems
- Create new opportunities
- Improve critical decision-making
- Achieve competitive advantages
This future-focused approach to enterprise operations is called
“predictive analytics,” or the art and science of using past data to
predict future trends or events.
As depicted in Figure 1, predictive analytics belongs to a larger field
called data analytics. As catalogued by analyst Catherine Cote, data
analytics is divided into four logical disciplines:
- Descriptive analytics, which answers the question, "What
happened?" - Diagnostic analytics, which answers the question, "Why
did this happen?" - Prescriptive analytics, which answers the question, "What
should we do next?" - Predictive analytics, which answers the question, “What might
happen in the future?”1
Figure 1. Types of Data Analytics
Source: Wikimedia Commons
How Does Predictive Analytics Work?
Predictive analytics (PA) capabilities are manifested through software
applications, in which, as analyst Linda Tucci describes, multiple
variables are “measured and analyzed to predict the likely behavior of
individuals, machinery, or other entities. The software relies heavily on
advanced algorithms and methodologies, such as logistic regression models,
time series analysis, and decision trees.”2
A common example familiar to all computer users is a spam filter, which
is utilized to predict the likely value – and safety – of incoming
e-mails.
As observed by analyst Natallia Sakovich, practicing PA on an enterprise
level involves four basic steps:
- “Identify what problems need to be solved. Ask yourself all
the questions you want to get answers to. For example, ‘What groups of
our customers are most likely to buy our new product?’ - “Check if you have the necessary data. For example, to answer
the question above, you need comprehensive information on your customers
(personal information, buying history, interactions with your brand,
etc.). - “Build and teach predictive analytics models. You should
create a system based on [deep learning (DL), machine learning (ML), or
artificial intelligence (AI)] algorithms and your requirements. - “Put insights into action. Predictive analytics will only bring
benefits if you act according to the insights the system provides.”3
PA Market
[return to top of this
report]
As forecast by Technavio, the predictive analytics market is expected to
increase by $17 billion from 2021 to 2026, realizing a remarkable compound
annual growth rate (CAGR) of 21.36 percent.4
One of the key market drivers is the need for operational efficiency.
Critical applications include:
- Process optimization
- Failure point identification
- Scam detection
- Fraud mitigation
- Customer retention
- Strategic decision-making
Facilitating the rise of predictive analytics will be contributions from
other, complementary technologies. “Increased acceptance of innovations
and technological advances, such as AI (artificial intelligence), ML (machine
learning), blockchain, cloud computing,
advanced analytics, big data, IoT (the Internet of Things), virtual assistants, automated vehicles,
as well as augmented and virtual reality, is expected to pave the way for
the adoption of predictive analytics throughout enterprises and
industries.”
In a highly-fragmented market, prominent PA participants include:
- Amazon
- HPE
- IBM
- Microsoft
- Oracle
- Salesforce
- SAP
- SAS Institute
- Teradata
Amazon Web Services, for example, offers Amazon Monitron, “an
end-to-end system that uses machine learning to detect abnormal conditions
in industrial equipment and enable predictive maintenance.”
PA Use Cases
[return to top of this
report]
Predictive analytics are employed across a wide variety of industries.
For example:
Retail
According to analyst Natallia Sakovich, “Retailers are probably the
leading users of predictive analytics applications.” Common use cases
include:
- “Predictive marketing – Algorithms analyze market trends,
buying habits and personal details of customers to further identify
buying patterns and perform customer segmentation. - “Predictive inventory – Algorithms analyze various factors
(region, season, buying habits) to forecast the demand for various
products. - “Predictive supply chain – Algorithms help companies optimize
several aspects of supply chains. [First], they make logistics more
efficient by determining the fastest and most cost-efficient routes
considering toll roads, traffic, weather conditions, etc. [Second],
trackers monitor fuel consumption and driving behavior, thus reducing
transport costs. And [third], sensors monitor the conditions of machines
and their components, anticipating technical maintenance and avoiding
downtime.”5
Manufacturing
Assembly line mistakes can be quite expensive and difficult to reverse.
To help reduce the risk, predictive analytics are applied throughout the
manufacturing process. Common use cases include:
- Fault detection and failure prediction
- Forecasting product demand
- Cost modeling for product pricing
- Analytics for predicting warranty and product maintenance6
Analyst Catherine Cote reports that: “Some algorithms even recommend
fixes and optimizations to avoid future malfunctions and improve
efficiency, saving time, money, and effort. This is an example of
prescriptive analytics; more often than not, one or more types of
analytics are used in tandem to solve a problem.”7
Healthcare
As itemized by analyst Mary K. Pratt, predictive analytics serves the
interests of physicians, patients, and healthcare administrators. Prominent use cases include:
- Clinical predictions and disease progression – PA helps inform
patient prognoses – short- and long-term. - Hospital overstays – PA helps identify which patients are
likely to exceed the “average length of stay” for their particular
conditions. - Hospital re-admissions – PA helps identify which patients are
likely to be readmitted once discharged. - Resource acquisitions and allocations – PA helps identify which
resources are likely to be in high demand, and where these resources
will be needed. These were critical issues during the height of the
COVID-19 pandemic as hospitals and other healthcare facilities quickly
exhausted basic supplies like masks and gowns, and essential equipment
like respirators. - Patient engagement – PA helps healthcare organizations to
better understand and engage their patients. - Insurance reimbursements – PA helps identify which claims are
likely to be declined and which claims could yield higher payments. - Optimal treatments – “Treatments for certain conditions, such
as some cancers, should be tailored to the patients and their diseases
to achieve the best patient outcomes. But no individual can analyze all
the data required to make those treatment decisions, which is where
predictive analytics technology comes in.”8
Cybersecurity
The US National Institute of Standards and Technology (NIST) is working
with stakeholders from across government, industry, and academia to
research and prototype methods and tools to enable predictive risk
analytics and identify cyber risk trends.
NIST’s goal is to enable information sharing among risk owners about
historical, current, and future cyber risk conditions and is intended to
help not only enhance existing cyber risk mitigation strategies but also
improve and expand upon existing cybersecurity risk metrology efforts.9
Sports
Contributing to the consternation of traditional baseball fans, many of
whom are still trying to adapt, decades later, to the designated hitter
and other unnatural developments, predictive analytics – or just plain
“analytics” – has taken control of our national pastime, dictating, for
example:
- How many pitches a starting pitcher should be permitted to throw.
- How position players should be positioned to defend against a
particular hitter. - Whether a player should be allowed to bunt or steal a base in a
specific situation.
While analytics often yield results that vary with old-school baseball
strategy, a number of clubs, notably the Oakland Athletics, have used predictive
analytics to optimize the potential of their players.
PA Future
[return to top of this
report]
Predictive Analytics Is Still Emerging
As a data analytics discipline, predictive analytics is still evolving
and maturing, which has had the impact of retarding wholesale enterprise
adoption.
As reported by analyst Cath Everett, David Semach, partner and head of
artificial intelligence (AI) and automation for Infosys Consulting in
Europe, the Middle East and Africa, believes there are three relevant
factors affecting PA growth:
- First, the absence of a single “silver bullet” PA tool or toolset.
- Second, the time and effort to build predictive models and to
aggregate the data necessary to feed them. - Third, the resistance of some business leaders who do not trust
machine predictions.
Supporting these findings, Semach reveals that “A survey we did in early
2020 found that 91 percent of business decisions are made with a lack of
supporting data but are based on human experience and gut feeling.”10
AI Will Complement or Replace
As with many enterprise practices, artificial intelligence is
revolutionizing predictive analytics. As analyst Linda Tucci reports,
“Advanced machine learning techniques are reducing the need to deeply
understand how different variables affect each other, automatically
choosing the best combination of algorithms for a given task. There is
also a growing market of industry-specific analytics tools with prebuilt
models and templates that embody best practices and dramatically simplify
the predictive analytics process.”11
Still unresolved is whether AI will continue to complement predictive
analytics, or affect a complete replacement. In other words, will PA
functionality be reduced to a line item on your AI system’s menu.
IoT Is the Force Multiplier
If predictive analytics is the engine that powers enterprise
decision-making, the Internet of Things (IoT) is the future fuel source,
producing large volumes of real-time data about real-world operations.
According to analyst Natallia Sakovich, “The primary use case [today] is
predictive maintenance in smart manufacturing. IoT sensors installed on
machines continuously collect data on their performance and send it to the
processing platform where predictive models perform the analysis, identify
abnormalities, and suggest maintenance of specific spare parts. By
applying such analytics, plants and factories eliminate equipment
breakdowns and avoid downtime.”12
As the IoT expands to encompass all mechanisms, including the human
variety, predictive analytics will feature in the maintenance and
management of all:
- Manufacturing systems
- Production systems
- Transportation systems
- Energy production systems
- Healthcare systems, particularly as biological monitoring "things" are
worn or implanted
Web Links
[return to top of this
report]
-
Amazon: http://www.amazon.com/
ASIS International: http://www.asisonline.org/
US National Institute of Standards and Technology: http://www.nist.gov/
References
1 Catherine Cote. “What Is Predictive Analytics? 5 Examples.”
Harvard Business School | President & Fellows of Harvard College.
October 26, 2021.
2 Linda Tucci. “What Is Predictive Analytics? An Enterprise
Guide.” TechTarget. December 2021.
3 Natallia Sakovich. “10 Examples of Predictive Analytics.”
SaM Solutions. January 21, 2022.
4 “Predictive Analytics Market by Deployment and Geography –
Forecast and Analysis 2022-2026.” Technavio | Infiniti Research Limited.
April 2022.
5 Natallia Sakovich. “10 Examples of Predictive Analytics.”
SaM Solutions. January 21, 2022.
6 David Lechevalier, Anantha Narayanan, and Sudarsan Rachuri.
“Towards a Domain-Specific Framework for Predictive Analytics in
Manufacturing.” US National Institute of Standards and Technology.
7 Catherine Cote. “What Is Predictive Analytics? 5 Examples.”
Harvard Business School | President & Fellows of Harvard College.
October 26, 2021.
8 Mary K. Pratt. “Predictive Analytics in Healthcare: 12
Valuable Use Cases.” TechTarget. December 13, 2021.
9 “Cybersecurity Risk Analytics.” US National Institute of
Standards and Technology. September 10, 2020.
10 Cath Everett. “AI More Likely to Complement than Replace
Predictive Analytics.” TechTarget. August 26, 2021.
11 Linda Tucci. “What Is Predictive Analytics? An Enterprise
Guide.” TechTarget. December 2021.
12 Natallia Sakovich. “10 Examples of Predictive Analytics.”
SaM Solutions. January 21, 2022.
About the Author
[return to top of this
report]
James G. Barr is a leading business continuity analyst
and business writer with more than 40 years’ IT experience. A member of
“Who’s Who in Finance and Industry,” Mr. Barr has designed, developed, and
deployed business continuity plans for a number of Fortune 500 firms. He
is the author of several books, including How to Succeed in Business
BY Really Trying, a member of Faulkner’s Advisory Panel, and a
senior editor for Faulkner’s Security Management Practices.
Mr. Barr can be reached via e-mail at jgbarr@faulkner.com.
[return to top of this
report]