Using Big Data to Avoid Value Chain Planning Mistakes
1. Using Big Data to Avoid Value Chain
Planning Mistakes
Author
Salil Amonkar
Director Consulting Services, Bodhtree
samonkar@bodhtree.com
Co-Author
Manju Devadas
VP Consulting Services, Bodhtree
mdevadas@bodhtree.com
2. Take a well-known brand and check the relevant social mining data: Facebook Likes, 5-star vs.
1-start ratings, blog comments with the words ‘hate’ or ‘awesome’ in the same sentence as
‘brand X,’ or the words ‘consistent’ and ‘flaw’ as part of one phrase in a user review. The
chances are that the data are telling you something about perceptions of product quality,
comparisons to competitors or the reach of a social marketing campaign. But how does all of
this relate to Value Chain Planning?
Consider the case of after-sales support. Most interactions between support and the customer
are unstructured – telephone conversations, video tutorials, user forums, etc. – and therefore
not easily handled by traditional data capture and management solutions. Consequently,
invaluable insights to customer experience never make it up the value chain. If captured and
analyzed effectively, these interactions could trigger improved reliability, product ease-of-use
and effective prioritization of R&D investments.
Business Performance Management illustrates another angle of the potential value
unstructured data offers to Value Chain Planning. Signals from social and internal data mining
reflect Demand, Supply and Pricing behaviors embedded in all aspects of the value chain. Most
companies today rely on manual human analysis to factor in these elements, predict demand
cycles and decide product pricing, often resulting in extreme misalignment between supply and
demand.
But can enterprises leverage machines for this planning rather than requiring humans to sort,
search and analyze the unstructured data? Automating strategy is a game changer, and it has
become a reality with Big Data. With today’s technology, it is not only possible to capture and
manage unstructured data but also run language analysis processes that can distill the essence
of the customer’s meaning (i.e. emotive component).
This aspect is missed by traditional structured data analysis.
Using Big Data in Value Chain Planning
3. In the support example above, a simple usability or quality issue may never be identified and
reported to Value Chain Planning, leaving vulnerability for competitors to exploit. The silos
inherent to the traditional data approaches fail to “connect the dots” that convey the full story
of the issue or opportunity being faced.
Figure 1 Traditional Data Analysis Challenge
The two key factors in Value Chain Planning, demand and supply, rely on price, or perceived
value, to maintain their delicate balance. Human and technological advancements are
constantly in a race to optimally balance Demand, Supply and Price, a premise referred to as
the DSP problem. Today’s enterprise resource planning solutions approximate demand using
structured data and historical information. By augmenting this data with customer preferences
from unstructured sources via Big Data Analytics, a far more accurate predictive model emerges
for Demand Planning. This is illustrated in the figure 2.
Using Big Data in Value Chain Planning
4. Figure 2: The Big Data Effect on Demand Analysis and Enterprise Supply Chain
Why is Big Data Needed in Value Chain Planning?
Automotive Recalls – Leveraging Big Data to Avoid or Mitigate a Costly Business Event
The recent past offers multiple examples of automotive recalls that cost manufacturers billions
of dollars and required several years to resolve. A consistent element among these impactful
events is the reactive nature of the recalls, creating a multiplier effect on cost.
Every automobile has a prescribed service schedule with regular maintenance visits. During
these service interactions, maintenance centers download large amounts of data from the car’s
sensors. However, in today’s structured data management environments, the total information
captured is so enormous and unwieldy that it generally remains at the local level and is never
utilized by manufacturers to gain insight to the fleet’s overall status, including any emerging
safety and reliability issues.
Using Big Data in Value Chain Planning
5. But this sensor data is not the only feedback channel available to automobile manufacturers.
Customers, service workers and industry experts create millions of posts on social media,
product reviews, and blogs expressing their perspectives and preferences surrounding the
product. Traditional data analysis approaches are not equipped to analyze these massive
quantities of unstructured data.
The Potential of Big Data/Data Mining and Supply Chain Management
Gartner and other research firms have stated the following with regard to data mining and
social media:
• Using data mining and social media for returns, customer service, feedback and
collaboration holds the key to improving value chain ROI.
• Using social media to analyze procurement community reviews and reactions holds
substantial promise for effective procurement within the supply chain
• Companies using Big Data analytics for Supply Chain Management have shown
consistent improvement in Supply Chain performance.
Big Data offers multiple routes for Business and IT to navigate the vast quantity of information
involved in capturing these opportunities and provides actionable insights for quick wins that
validate long-term strategies. Information that has traditionally remained in silos (either
corporate or across the value chain) can be aggregated to provide an end-to-end real-time view
of value chain processes, unlocking latent potential for revenue increase and cost-savings
leading to higher profits.
The Big Data Difference
Extending the automobile recall case study, the following scenario illustrates how Big Data can
be pivotal in altering the costly progression of events.
The enormous amount of structured information downloaded at each service center could be
combined into a Big Data management solution and processed with Big Data analytics.
Additionally, the unstructured social media information from customer reviews, service blogs,
social media posts, etc. can be captured into the same Big Data system revealing customer
preferences, perceptions and other expectations important to predicting demand cycles.
Using Big Data in Value Chain Planning
6. The structured portion of the data can be used to uncover quality trends, especially during the
period following a new model or design. These insights enable the manufacturer to proactively
prevent or mitigate issues that could otherwise result in massive recalls several years down the
road.
Another recent report describes how a well-known worldwide fast food chain franchise applied
a similar solution to proactively maintain product quality among it’s more than 34,000
restaurants worldwide (See the article cited in the reference section for the full case study,
including success metrics for the Big Data Strategy).
This is illustrated in Figure 2 shown earlier and reproduced below:
Using Big Data in Value Chain Planning
7. Other Relevant Use Cases for the Big Data Insights
A similar solution can be used to determine customer preferences – invaluable feedback, which
can be incorporated into product design and engineering to improve desirability and expand
the target market. The same customer preferences can be fed into Big Data analytics to more
accurately predict demand cycles and determine the impact on the DSP problem.
Usage of Big Data in the Solution
What is Big Data?
In very simple terms, Big Data is the enabler that helps you find the proverbial needle in the
haystack of your universe of structured and unstructured information. Big data refers to very
large data sets that are challenging to store, search, share, visualize and analyze and need
methods different than traditional databases to capture and manage.
The three key components of big data are as follows:
• Big Data Storage and Management system, which is typically based on distributed file
management system such as one provided by Hadoop (also referred to as HDFS), the
associated components for data services management (such as Cloudera Manager) and
retrieval (such as NoSQL which stands for Not Only SQL)
• Big Data integration mechanisms such as Big Data Connectors, Big Data Loaders and Big
Data integrators
• Big Data Analytics including Statistical Analytical engines (such as Project R for Open
Source based analytics and Oracle R which is provided by Oracle)
To illustrate how Big Data would make a difference, we will compare the traditional Supply
Chain Management approach to the Big Data-enabled solution.
Traditional Approach
Traditional solutions use structured data with rigid algorithms to roll up information and
produce guidance. These solutions are generally incapable of processing massive information
quickly. These usually adhere to the following solution strategy:
• Map structured sources of information
• Gathers data only provided by partners in supply chain
• Analyze
• Communicate
Using Big Data in Value Chain Planning
8. Big Data Approach
Big Data enables the mash-up of both structured and unstructured information with the ability
to run patterns and statistical models to derive actionable intelligence quickly. Big data
solutions apply the following general approach:
• Get structured information as in traditional approach
• Add unstructured data feeds from value chain
• Mash-up this data
• Structure mashed-up data using new analytical methods
• Mine data for subtle insights far-reaching business impact
• Feed information to drive business actions via dashboards and reports
In the automotive example above, we would use the mash-up of the information from
unstructured and structured channels to determine that a specific sensor pattern was
associated with customers reporting engine issues and that both could be linked to gaskets
produced at a specific plant in Asia.
In its simplest form the solution will use Social Mining and Big Data to determine patterns of
consumer behavior across geographies. These patterns can be fed as input to scenarios for
demand-supply balancing solutions to determine the most profitable strategies.
This is illustrated in the figure 3 below
Figure 3 Big Data Implementation Architecture
Using Big Data in Value Chain Planning
9. Big Data-Driven Value Chain Planning Solution Approach and Benefits:
• Determine value chain unstructured information feed and inventory all unstructured
information to be collected from the value chain
• Architect Mash-up with traditional structured information: determine the Big Data
architecture to bring in the unstructured information and mash it up with the structured
information
• Statistical patterns determination and data mining – set the patterns required to alert
stakeholders to key value chain disruptors (quality-related such as contamination and cost-
related such as a holdup in supply chain node)
• Proactive crisis management – continuously monitoring the possibility of crisis and taking
steps to preempt it, quickly capturing the magnitude of an unfolding crisis to rapidly
respond, diffuse and assist.
• Keeping tabs on the competition – understanding gray market tactics used by competition
to get access to sensitive information
Big Data-Driven Approach Addresses These Key Steps:
• Simplify and scale your mountain of content. The Big Data platform will help you gather and
manage huge quantities of structured and unstructured value chain information
• Identify the needle in the haystack of content. Track trends over time and use them to
identify important business-impacting patterns.
• Lighten your load when it comes to reporting. This platform can aggregate your data so you
can easily extract the information you need.
Recommended Approach for Success
By combining the experience of multiple Big Data implementations with a deep understanding
of Value Chain Planning solutions we enable the following:
• Determine the right approach for the problem at hand
• Architect the Big Data solution in the most seamless and incremental way
• Proof of concept for feasibility and establishing confidence in the solution
• Incremental roadmap approach that seeks to minimize risk and helps achieve the
business benefits incrementally over time
The author and his peers bring thought leadership and expertise in both Value Chain Planning
solutions and Big Data solutions. Typically they can review the unique situation at your
enterprise and put together individual ROI and business value cases, as well as assess how well
Using Big Data in Value Chain Planning
10. the value chains are aligned with the current business needs. This approach will consider how
to modularize it for your business model and to feed the final result into your tactical demand
planning, supply planning or pricing applications.
Conclusion
The core of Big Data analysis is the ability to find the most accurate answer, faster and with
minimal cost irrespective of data volumes. Value chain optimization involves finding answers to
specific problems that affect your supply and demand well before the actual impact occurs,
allowing corrective action with minimal delay.
With volumes of data growing for social mining and the machine memory and processing
power becoming cheaper, getting inputs to feed into your demand planning (e.g. Oracle
Demantra, i2, SAP DP) and supply planning (Oracle ASCP, SAP APO) applications is bound to
become critical in the coming years across industry verticals. Executing this process
systematically to drive real-time optimization will differentiate the leaders.
While the automobile industry served as the topic for the current discussion, the same
concepts can be widely adopted across industries and verticals. Unlike traditional approaches
that use only structured supply chain management solutions, leveraging Big Data in
combination with Value Chain Planning solutions provides the ability to avoid costly mistakes
that can directly contribute to deficiencies and/or limitations in the Value Chain Planning
process. The same solution can be augmented with Social Mining to make profitable revenue
decisions by validating market insights with scenarios for Demand/Supply balancing.
References and acknowledgements
Gartner’s predictions for Big Data
Gartner’s articles on Value Chain Planning and Management
From the mail to McDonald’s: Big Data is all around us
http://fcw.com/articles/2013/03/12/big-data-all-around.aspx?m=2
In addition the authors would like to thank Krishna Guda (CEO of Bodhtree) for his
tremendous insights, support and guidance around usage of Big Data in improving Value
Chain Planning.
Using Big Data in Value Chain Planning