Proactive Supply Chain Performance Management with Predictive Analytics
Read this article. A predictive performance management model is introduced to manage complex business network collaborations and minimize uncertainty. Pay attention to the innovative performance management systems characteristics. What other attributes would you add to the list?
Background Research
Performance management represents a
cohesive element that unites different business improvement initiatives,
directs strategy formulation, and plays the key role in strategy
implementation and monitoring. PM is of great importance for managing
supply chains. Timely and accurate evaluation of the entire supply chain
and individual companies is prerequisite for successful functioning.
Performance management should be a part of any proper supply chain
strategy, planning, and reporting process.
There has been lot of
research efforts during the last decade that deal with various aspects
of supply chain performance management. Shepherd and Gunter provide
taxonomy of supply chain performance measurement systems and metrics
with critical review of the contemporary literature. The study shows
that despite substantial advances in the literature in the last decade,
there are still important topics related to supply chain performance
management that did not receive adequate attention, such as process
modeling, data integration, software support, and forecasting. Estampe
et al. provide analysis of various supply chain performance models by
stressing their specific characteristics and applicability in different
contexts, so that decision-makers can evaluate and apply models and
metrics that best suit their needs. Arzu Akyuz and Erman Erkan also
provide a critical literature review on supply chain performance
measurement. The results show that performance measurement in the new
era is still an open area of research. This particularly refers to
framework development, collaboration, agility, flexibility, and IT
support systems. Gopal and Thakkar give a comprehensive review of supply
chain performance measurement systems. The analysis shows that, in
spite of considerable evidence from the literature, there is a large
scope for research to address critical issues in supply chain
performance measurement, including metrics, benchmarking, integration,
business intelligence, and collaborative decision making.
The
goal of supply chain performance management is business process
optimization through monitoring and analysis of key performance
indicators. By measuring and monitoring metrics against predefined goals
companies can provide added value to large volumes of data generated
over time. This type of analysis allows companies to track various
metrics at different organization levels and to take timely actions.
This
has elevated the importance of key performance indicators and their
ability to measure, predict, and manage the business health of a supply
chain in real (or near real) time. KPIs have morphed from static siloed
measures to dynamic real-time enterprise metrics. Key performance
indicators are normally backward-looking because they are based on
historical information and often do not help in forecasting future
events or performance. Predictive metrics would make it possible to
predict future problems in supply chain operations and to enable
proactive evaluation and improvement actions in advance. Statistical
modelling and data mining under the guise of predictive analytics have
become critical building blocks in setting the new KPI standard for
leading indicators.
However, most of the existing
performance measurement systems are mostly based on financial indicators
(i.e., costs). They are also internally focused, incompatible, and
historical. In today's business environment this performance measurement
approach is no longer adequate. New business climate requires novel and
innovative performance management systems which have the following
characteristics:
(ii) comprehensive (takes into account relevant variables),
(iii) use also non-financial KPIs,
(iv) adaptable to specific supply chain configuration and member companies,
(v) optimal number of KPIs,
(vi) simple and easy to use,
(vii) timely and actual,
(viii) accurate and consistent,
(ix) foster improvement, not just monitoring,
(x) represent information at different supply chain levels,
(xi) provide analytical tools that offer multidimensional reporting,
(xii) enable a proactive management, rather than reactive,
(xiii) provide insights into newly emerging trends, opportunities and problems,
(xiv) deliver information to right people at any time and on any device.
In response to these
challenges for measuring supply chain performance, a variety of
measurement approaches have been developed, including the following: the
balanced scorecard, the supply chain council's SCOR model, the
logistics scoreboard, activity-based costing (ABC), and economic value
analysis (EVA). Lambert and Pohlen introduced a framework for
developing metrics that measure the performance of key supply chain
processes, identify how each firm affects overall supply chain
performance, and can be translated into shareholder value. Cai et
al. propose a framework for analyzing supply chain KPIs accomplishment,
so that management strategy can be attuned by interpreting the analysis
results. This framework aligns performance at each link
(supplier-customer pair) within the supply chain. Gunasekaran et al.
developed a framework for supply chain performance measures and metrics
considering the four major supply chain processes: plan, source, make,
and deliver. Metrics were classified at strategic, tactical, and
operational levels in order to clarify management authority and
responsibility for performance.
In an attempt to overcome some
limitations of traditional PM systems, many companies have initiated
balanced scorecard (BSC) projects. Based on the methodology of Robert
Kaplan and David Norton, these companies created a balanced set of
metrics representing financials, customers, internal business processes,
and innovation. The goal was to enable better decision making by
providing managers with a broader perspective of both tangible and
intangible assets. However, the supply chain has specific processes and
metrics which require special types of balanced scorecards. Bhagwat and
Sharma proposed a balanced scorecard for supply chain management that
measures and evaluates day-to-day business operations from the four BSC
perspectives and with specific SCM metrics.
Only a few
leading-edge companies are currently using true extended PM systems
(either in-house developed or implemented PM software applications) that
not only measure the performance of their enterprise but also measure
that of their supply chain wide activities. Most companies are still in
the internal or integrated stage of the maturity model (as shown in
Figure 1) where they focus on the performance of their own enterprise
and measure their supply chain performance with financially oriented
metrics.
Figure 1 Supply chain performance measurement maturity model.
![](https://learn.saylor.org/pluginfile.php/5619163/mod_book/chapter/50114/528917.fig.001.png?time=1716315548544)
One
of the prerequisites for effective supply chain performance measurement
is the initiative geared toward standardization of supply chain
processes and metrics. Standardized models facilitate application
integration and collaboration, enable benchmarking for performance
comparison, and provide best practices for process improvement and
gaining a competitive advantage. By standardizing supply chain processes
and metrics for managing such processes, companies are able to not only
compare their results against others, but to also gain visibility of
operations over a supply chain that may cross corporate borders.
Partners in a supply chain can communicate more unambiguously and can
collaboratively measure, manage, and control their processes.
SCOR
(supply chain operation reference) model represents a universal
approach to supply chain management that can be applied in diverse
business domains. SCOR combines business process engineering,
benchmarking, and best practices into a single framework. This
standardization movement facilitates the deliverance of business content
for the supply chain KPI and predictive analysis model. Kocaoğlu et al.
propose a SCOR based approach for measuring supply chain performance. It uses the analytic hierarchy process (AHP) and technique for
order preference by similarity to ideal solution (TOPSIS) together for
the linking of strategic objectives to operations.
Although the
SCOR model provides a layered metric system and thereby a business
context for the KPI system, it is only the first step toward the
pervasive supply chain performance measurement. In order to implement
such a complex PM system as found in supply chains, member companies
need to extract, transform, and load all the relevant data into a unique
and integrated data source.
Supply chain integration is critical
to both operational and business performance. In order to achieve the
full effect of supply chain integration, companies need to use
collaborative and intelligent web information systems to actively manage
process uncertainty. Most of the existing information systems that
support supply chain management are repositories for large amounts of
transactional data. These systems are said to be data rich but
information poor. The tremendous amount of data that is collected and
stored in large, distributed database systems has far exceeded the human
ability of comprehension without analytic tools. Analysts estimate that
80% of the data in a transactional database that supports supply chain
management is irrelevant to decision making and that data aggregations
and other analyses are needed to transform the other 20% into useful
information.
As a result of the huge amount of data
generated within supply chains, new tools and methods should be
developed which are capable of storing, managing, and analyzing the
data, as well as of monitoring global supply chain performance. Chae and
Olson propose an analytical framework for supply chain management,
which is composed of three IT-enabled capabilities: data management,
analytics, and performance management.
Increasingly,
transactional supply chain data is processed and stored in enterprise
resource planning systems, and complementary data warehouses are
developed to support performance management and decision-making
processes. Some organizations have developed data warehouses and
integrated data-mining methods that complement their supply chain
management operations.
Investment in business
intelligence software is necessary if the organizations want to manage
their supply chain more effectively. The studies show that main drivers
for BI initiatives include complexity of operations, huge amounts of
data, requirements for cost reduction, increased revenue, and efficiency
improvement, as well as high exposure to risks. Data warehousing
and online analytical processing technologies combined with tools for
data mining and knowledge discovery have allowed the creation of systems
to support organizational decision making.
Inventory
management is probably the key supply chain process, and inventory costs
represent a large portion of the total supply chain costs.
Incorporating predictive analytics in an inventory management process
can lead to many benefits such as cost reduction, higher customer
service level, optimal reordering policy, enhanced productivity, shorter
cash-to-cash cycle time, and ultimately increased profitability.
Different data mining techniques were used for solving supply chain and
inventory management related problems. Dhond et al. used neural-network
based techniques for inventory optimization in a medical distribution
network which resulted in 50% lower stock levels. Symeonidis et al.
applied data mining technology in combination with an autonomous agent
to forecast the price of the winning bid in a given order. When it
comes to inventory management, Stefanovic et al. used different data
mining models clustering retail stores based on sales patterns and
one/two weeks out-of-stock forecasting.
When data mining
predictive analysis capabilities are closely integrated into every stage
of the data life cycle, they incorporate intelligence into reporting,
data integration, OLAP analysis, and business performance monitoring.
This helps supply chains to increase business agility and creates a
tangible competitive advantage.
As supply chain performance
management systems mature, it is almost certain that predictive
analytics will be an integral part of future PM and BI solutions.
Predictive analytics leverage historical and current performance data in
order to make predictions on future performance. For example, there
could be a predictive analytical model that can make predictions about
supply chain management costs, perfect order fulfillment, return on
working capital, and so forth. The main difference between predictive
analytics and classic BI is that predictive models contain an additional
stage which utilizes certain data mining algorithms to predict future
performance outcomes. This information can be used as a valuable
resource for the decision-making process and refinement of strategic,
tactical, and operational supply chain plans. One recent survey showed
that 87 percent of respondents said that they think predictive analytics
is important to the budgeting and planning process, but only 17 percent
are employing technologies that include the capability. Other
studies show that the ROI of predictive mining applications is almost
five times greater than that of nonpredictive applications using
standard query, reporting, and analysis tools.
Although with
great practical potential, there have not been many research projects
related to predictive performance management. Seifert and Eschenbaecher
introduced a predictive performance measurement approach as a planning
tool for virtual organizations to anticipate the performance of a
planned virtual team. Derrouiche et al. proposed an integrated
framework for supply chain performance evaluation based on data mining
techniques. It enables the development of a predictive
collaborative performance evolution model and decision making which has
forward-looking collaborative capabilities. Maleki and Cruz-Machado
proposed a framework for developing data mining models based on Bayesian
networks which account uncertainty and mutual dependency among supply
chain performance measures.
Predictive KPI analytical models
can identify critical KPIs and their likely impact on other indicators,
both at the same level or higher-level KPIs. Besides typical predictive
data mining models, other models can be created in order to further
clarify prediction results, analyze key influencers, and perform
scenario analysis such as goal seeking and what-if analysis. Thus,
predictive PM systems will not only notify users about KPI risks, but
will also provide additional information necessary for taking corrective
actions.
On the other hand, most of the existing PM systems are
not business user-friendly which makes them less attractive for wider
implementations. Modern PM systems have to be web-based, interoperable,
modular, and customizable.
Based on the analysis of the existing
research results and also the methods and tools used in real-world
supply chain operations, two main conclusions can be
derived.
(i) Organizations show a strong interest in proactive supply chain performance management, but very few PM projects deal with predictive analytics.
(ii) Most of existing research efforts, methods, and
tools are targeted to specific supply chain PM aspects, without a
comprehensive approach which would provide organizations with concrete
models, methods, and tools for achieving business objectives.
What
is needed is a unified supply chain performance management system to
collect, integrate, and consolidate all relevant data and to use
business intelligence tools like data warehousing and data mining, to
discover hidden trends and patterns in large amounts of data, and
finally to deliver derived knowledge to business users via web portals.
In this context, a novel supply chain PM model and software solution
that fulfills most of the previously stated requirements are proposed.
In the subsequent sections, more information related to supply chain
OLAP modelling, KPI design, data mining KPI prediction model, and PM web
portal is presented.