Data has always played a key role in Supply Chain Management. Order forms, supplier data, stock movements, production schedules, shipment information and warehouse movements are all examples of data generated by day-to-day supply chain operations. Typically this data is used to support decision making both strategically and on a day to day basis to manage and respond to fluctuating customer demand.
For most companies today the problem is not collecting supply chain data but rather making sense out of it and using it for driving improvement in supply chain operations.
However, information generated by a complex network of activities within the supply chain often needs to be analyzed and immediate decisions made in order to mitigate potential problems or maximize opportunities. Data analytics can support decision making of supply chain managers by analyzing the large amounts of data, help to understand patterns, and measure supply chain activities.
Analytics can enhance each of the activities described above, but the greatest value is provided when it is used to interlink the various processes and functions within the supply chain to improve overall supply chain performance.
What Is the Role of Data Analytics in Supply Chain Management?
The fundamental role of data analytics in supply chain management is to analyze the data of operational processes and to generate relevant information for decisions in supply chain management.
A supply chain by its nature is a data-rich area of a business. A purchasing department will produce purchase orders and supplier contact details, a warehouse will hold details of stock movements and fulfillment of orders, a logistics department will hold information regarding the shipping, transportation and delivery of products to customers. A manufacturing department will hold information on production including quality information, a sales department will hold information on customer demand and a finance department will hold historical cost information for the supply chain.
When datasets from different datasets are not linked together it can be very difficult to understand the interrelation.
For example, an organization might identify that its inventory levels are increasing. Understanding where this is happening in the supply chain, and why, requires data on demand, purchasing, supplier lead time, and product-specific stock movement. Increasing logistics costs may need to be analyzed in conjunction with metrics measuring shipment volume, travel routes, carrier performance, and fluctuating customer demand.
Supply chain analytics links together the various datasets and views them from a supply chain perspective to generate supply chain analytics.
Understanding why something has happened and where it could happen next is key to truly improving a supply chain, and will help the organization in question to look for areas that could be changed for improvement.
Why Supply Chain Analytics Has Become Important
The modern supply chain is complex and spans globally. It includes multiple suppliers for all raw materials required to manufacture a product (no single supplier for any one material). Products are manufactured in multiple locations around the world and are stored in multiple warehouses before being distributed to end customers in a multitude of geographies around the world.
Complex supply chains can often cause problems in the optimizing of cost, of stock, of service, of supplier performance and of process.
If a report on inventory, suppliers or transportation is only produced on a monthly basis then, by the time it has been analyzed and actions have been implemented, the underlying situation will have changed.
By analyzing Supply Chain Data, supply chain analytics creates a continuous and updated analytical view of your business. It can also identify dependencies and help analyze processes in more detail than analyzing single metrics.
By examining inventory analytics in conjunction with demand and purchasing data, supplier analytics connected with procurement spend and production requirements for suppliers, and logistics analytics connected with order volume and delivery performance, organizations have a multitude of connections within the activities of their supply chain.
The value comes from these connections.
Turning Supply Chain Data Into Actionable Insights
Having data available does not automatically result in insights.
Different definitions can often be found within a company’s various business units to determine similar metrics. Information is stored in various Supply Chain Systems such as the ERP system and other procurement, warehouse management, transportation management systems and even spreadsheets.
Data preparation is important to gain value from data.
After the data has been collected and structured, organizations can apply data analysis methods. They identify patterns, exceptions and even relationships between different supply chain data.
An example of such an analysis could be: The current high levels of inventory are not just bad in general. It depends on several factors such as products, locations and events. Also, the reasons for the high levels of inventory might differ from an increase in demand over the company’s normal safety stock to purchasing behavior and longer lead times of suppliers.
Tie analysis to decision. Ensure the supply chain data that you collect and for which you build models has value beyond the report that is produced.
The Role of Data Analytics in Inventory Management
One of the key areas in which value can be created in supply chain management through the use of supply chain analytics is in managing inventory.
Inventory levels that are too high can tie up large amounts of capital in excess stock while underutilizing warehouse space. Conversely, maintaining low levels of inventory may ensure adequate supply, but does pose the risk of stockouts impacting customer service.
Data analytics creates value in analyzing historical and current stock data together with demand data, purchasing data and supplier data of a company. It can highlight slow-moving products and situations of excess stock, as well as exceptions from normal stock developments.
Since there is more to inventory than just its value (also see the article on inventory analysis), also analyzing the different products in the stock helps in making better inventory decisions.
Furthermore, predictive analytics can be used to predict potential future requirements for Inventory in terms of quantity and delivery time. Here historical demand information as well as other relevant variables are analyzed for patterns to create future scenarios for planning.
However, the quality of the resulting analysis greatly depends on the underlying data which has to be adequate and relevant.
A Practical Example: Using Data Analytics to Reduce Excess Inventory — A Manufacturing Use Case
The organization in question is a manufacturer of goods, who maintain a stock of components at a number of warehouses across the country supplied by a variety of different suppliers. It has noticed that the total value of stock held has increased over the last year but cannot determine from the monthly reports currently produced why this is. The initial assumption is that the increase is due to higher sales and thus an increase in required safety stock.
At first glance the company’s monthly reports give a good indication of the total value of its inventory but provide no further insight into the reason for a rise in inventory value over the last year. After applying supply chain data analytics to the existing data (inventory, purchasing, demand, supplier lead-time by product) the team discovered that an increase in the overall inventory value was concentrated on only a handful of products, most of which were experiencing lower than expected demand while being purchased in larger quantities than required by current demand.
Initially the company had thought that an increase in the total inventory value over the past year was mostly due to higher sales and increased safety stock. A data analytics supply chain analysis for the company however shows that most of the increase in inventory value is caused by a small group of products. Most of these products even have lower than expected sales, but purchasing quantities are not being reduced. There are also several products which are purchased in bulk because of the minimum order quantities that suppliers have.
Analyzing the components of excess inventory can lead to a completely different type of conversation than simply asking for its reduction. For instance, supply chain teams could ask the following questions to increase productivity:
- Which products are accumulating in the inventory?
- Are purchasing quantities adequate in relation to current demand?
- Are there supplier constraints that could be alleviated?
- Could it be more efficient to keep inventory in other locations?
- Are safety stocks too high for some product categories?
The insights gathered from the analysis can then be evaluated to consider a range of different action scenarios, including potential changes to order quantities, negotiations with suppliers and changes to the inventory holding policies. Additionally, stock may need to be redistributed between different locations.
This information can then be used by management to determine what specific action needs to be taken. The value of the data is not in the creation of additional reports but rather to examine multiple data sets to gain new insights to a business problem and support a team in making better supply chain decisions.
These are examples of how data analytics in supply chain management can be applied in other processes and departments, for example when logistics costs are increasing, a detailed analysis of all relevant data (e.g. all shipments by all carriers on all routes for all orders) could reveal the causes of the rising costs, and opportunities for cost savings; in the procurement department, all supplier spend data combined with their delivery and quality performance could highlight better suppliers, and in the planning department, historical demand data combined with current inventory levels could highlight the impact of variable demand forecasting on stock levels.
This pattern repeats in each of the different scenarios for application of data analytics in supply chain management described above in steps to achieve supply chain decision-making.
Business problem → Relevant data → Analysis → Root-cause insight → Business decision → Continuous measurement
Data analytics in supply chain management can simply serve to create a bridge between existing data and decisions of the Supply Chain Department.
Data Analytics and Demand Forecasting
While historical demand data provides valuable information on past demand (i.e. past patterns and developments), demand is one of the most difficult variables to manage for any supply chain as it directly affects purchasing, production, storage and distribution.
Understanding of demand variability can be developed by combining historical demand data with other data relevant for the analysis.
This is where predictive analytics in supply chain can become relevant.
This information helps plan the supply chain for future demand. This does not mean all uncertainty in the supply chain demand can be removed. However, there is enough information available to create forecasts that help planners make informed decisions.
All these aspects have to be considered when planning purchase orders, scheduling production, calculating optimal inventory levels and designing distribution processes. Demand analysis has an impact far down the supply chain.
The Role of Data Analytics in Procurement and Supplier Management
The data generated by the procurement function of a supply chain includes information relating to the suppliers of a company, details of purchasing transactions conducted by the organization, prices paid for goods purchased, quantities bought, categories of purchasing (i.e. raw materials, indirect goods etc.) and payment related information.
In examining purchasing data from past suppliers, for example, a company can make a formal analysis of buying habits (for example, quantity purchased, average price paid per unit) over time for particular purchasing categories.
For instance, a supplier’s purchasing pattern can be analyzed based on various factors such as different categories, business units and locations. His supplier delivery performance, quality, purchasing activity, etc. can also be analyzed in conjunction with the above data.
This information can be analyzed to find supplier risks, for example supplier concentration, supplier’s declining performance, new dependencies in the supply base etc.
Once you have identified all supplier data (delivery performance, quality etc.) to complement purchasing data, Supplier risk analytics will help you to investigate supplier Concentration, Deterioration and Dependencies issues. Analytics does not remove risk but provides visibility of problems.
All in all, with the various analyses that can be run using supplier data, it is clear that there is considerable potential for supply chain and procurement analytics to overlap. Consider for instance the impact that supplier decisions have on inventory, production, cost and even service levels to customers.
The Role of Data Analytics in Logistics
Logistics and Transportation: Logistics data related to the movement of goods around the world, including information on shipments, routes, modes of transport, carriers, delivery times and costs, can also be analyzed.
Information derived from data from transportation operations, like information on shipments, routes, carriers, times of delivery, cost of freight and volume of freight, can help identify patterns of performance and help organizations to identify major causes of high logistics costs.
For Transportation Operations and Management a Logistics analytics solution enables users to analyze their Transportation Operations and Management data and answers questions like:
- Why do the Transportation Costs for a certain location differ from location to location?
- Why do certain routes experience delays on a regular basis?
- How does the Carrier rate, and the resulting Transportation Cost perform for my organization?
- Are there changes in the Shipments that are sent by my organization?
The search interest in the keyword data also shows a strong interest in data analytics in logistics, data analytics for logistics and in general analytics for logistics and supply chain management.
Connecting this type of data to the relevant orders, stock levels and demand allows companies to investigate the root cause of transportation problems that occur in the logistics function and which are influenced by orders placed with other parts of the organization, the physical distribution warehouse(s) and inventory management policies and customer service commitments.
Warehouse and Operations Analytics
Another key location within a supply chain to implement business analytics within warehouse management to measure and improve performance, is the warehouse.
Business analytics can be applied to warehouse data. This enables analysis of data regarding the inventory movements, such as the number of orders, the picking process, the throughput and the storage of the warehouse and the shipment to customers. A trend can be identified in the change of the various warehouse processes over time.
Combining warehouse analytics with relevant Inventory and Logistics data will also create even better insights on the impact of Warehouse Operations to the rest of the Supply Chain.
The same principle applies to operations.
Business analytics in supply chain management enables Supply Chain Decision-Makers to analyze operational performance and to connect this to Procurement, Inventory Holding, Logistics and even to Demand Information in order to analyze the effects of Decisions in any part of the Supply Chain.
Using Analytics to Improve Supply Chain Visibility
It is challenging to achieve Supply Chain Visibility as most companies operate with a highly fragmented set of information systems.
A Procurement department may have supplier performance scores while a Warehouse department has data on inventory. The Logistics department will be working with data from their respective transportation systems and even management is typically left to interpret summarized data in periodic reports.
This inability to understand the supply chain fully in the information age has created additional challenges.
Supply chain analytics leads to more Supply Chain Visibility if relevant information is combined and depicted with consistent metrics, on dashboards, and by means of suitable analytical models.
A simple dashboard with KPIs to monitor the supply chain is not sufficient. The underlying data has to be correct and sufficient, the metrics used have to be well defined and the information generated by the analytics has to be relevant for the decisions taken by the individual user.
The data used must be accurate and sufficient. Metrics and data must be clearly defined and relevant to the respective decisions.
Descriptive, Predictive and Prescriptive Supply Chain Analytics
When it comes to data analytics in supply chain management, three different levels can be differentiated.
A starting point in the utilization of data analytics in supply chain management is given by the descriptive analytics approach. It mainly aims at the analysis of historical data (past events) and thus enables companies to get an accurate understanding of the performance of their Inventory, Suppliers and Operations.
With predictive supply chain analytics, companies are able to use the data at hand to determine the future by recognizing patterns in the historical information that has been compiled. In many cases this type of supply chain risk analysis can be used for demand forecasting, inventory forecasting, and even trying to prevent possible future disruptions.
Prescriptive analytics examines the possible actions (or scenarios) that may occur, helps compare alternative solutions and forecasts the possible consequences each of these solutions may incur. It allows the organization to determine the best possible solution to a problem and thus is very useful when the organization needs to compare alternatives and their resulting consequences.
This does not necessarily have to be the case. First a reliable descriptive analytics should be established to be able to work with accurate data to then gradually develop more advanced analytical applications such as predictive and prescriptive analytics.
Big Data and Supply Chain Management
Big Data in Supply Chain Management, which is relevant for various applications in supply chain management and Big Data Analytics, is increasingly discussed due to the excess of data that is generated in today’s supply chains.
Information can be collected from enterprise systems, as well as from suppliers and transportation companies. Sensors and monitoring systems installed at warehouses, in vehicles and on production equipment can generate vast amounts of data.
Once large amounts of data have been gathered, it can be processed using Big Data Analytics in order to look for patterns within the information.
However, more data is not necessarily better data.
So the information collected has to be judged for relevance, reliability and whether it can be put to good use in a specific supply chain decision.
Supply Chain Risk and Resilience Analytics
Supply chain disruptions can originate from a multitude of different factors, including supplier dependencies, logistics service failures, changes in demand as well as other internal supply chain operational failures.
As risks emerge within the supply chain, analytics tools help manage risk by identifying patterns and other factors in performance issues.
Supply chain risk analytics can incorporate a variety of data points to assess and monitor risk, including supplier information, purchasing history, inventory levels, logistics and operational data.
The data science results of a supplier risk analysis may help, for example, a company’s management team understand the concentration of supply at a single location and thus better plan additional processes or even set up a second supplier.
From Analytics to Supply Chain Optimization
While analytics can be used to track and measure a company’s supply chain, the ultimate goal of analytics is to support decision making to improve supply chain operations.
One way organizations utilize supply chain analytics is to evaluate and optimize their current supply chain processes, including their inventory levels, their supplier base, their logistics strategy and how their warehouses are running.
Optimization data science and advanced analytical techniques to solve supply chain optimization problems and perform what-if analyses.
An organization, for example, might like to compare different distribution networks, analyze the cost of different sourcing options or measure the impact that changes in demand have on inventory and logistics.
Note that while analytics can be used to underpin key supply chain decisions, the quality of the resulting decisions depends on the quality of the underlying data and assumptions. In this respect, supply chain analytics is best used within a decision-making framework rather than in isolation from the relevant supply chain expertise.
Why Data Quality Matters
One of the most common pitfalls when developing analytics for supply chain management is suboptimal data quality.
Incorrect supplier data, inconsistent product codes, missing transactions and many other errors cannot be corrected by an advanced analytical model.
When using data from various sources in the supply chain (e.g., from different IT systems), it has to be ensured that data with differing characteristics (e.g., differing naming conventions, measurement units, definitions) is prepared correctly for the analytics tasks. The number of supplier entries as well as product entries for the different locations can differ and need to be aligned (e.g., by eliminating duplicates).
Data quality, data preparation, data cleansing, data mapping and data integration all need to be addressed when creating a supply chain analytics environment to support decision-making.
Finally, as a repeatable data foundation is established, the maintenance of the created supply chain dashboards as the supply chain changes can be made so much easier by analytics engineering.
Why Organizations May Use Supply Chain Analytics Outsourcing
The necessary capabilities to support and run these systems can be too expensive for many companies to build in-house.
A company with a limited analytics staff can outsource supply chain analytics to obtain specialized services that the company’s analytic staff is not capable of handling or would take too long to develop within the company.
However, for organizations with limited resources to build up necessary knowledge in analytics an external partner could be very helpful in supporting with certain projects like inventory analytics, supplier analytics, logistics analysis, demand forecasting etc. or also just supporting on an ongoing basis with necessary extra capacity.
It is important to define the business objectives that an outsourcing contract should pursue. Instead of just transferring reporting tasks to external experts, the goal should be to improve data collection, analysis and usage in the supply chain.
The Broader Role of Data Analytics in Supply Chain Management
The role of data analytics in supply chain management is continually evolving as organizations gather more data and face increasingly complex decisions with regards to their supply chains.
Basic usage of analytics in supply chain management is to provide visibility into supply chain operations to assess and measure performance. More sophisticated usage of analytics in supply chain management is to identify patterns, forecast and simulate ‘what if’ scenarios to make decisions and choose alternatives.
Here is a list of the main Supply Chain Analytics applications, organized in several categories:
- Inventory: understanding stock levels, movements and excess inventory
- Demand: identifying patterns and supporting forecasting
- Procurement: analyzing purchasing and supplier information
- Supplier management: monitoring performance and risk
- Logistics: understanding transportation costs and delivery performance
- Warehousing: analyzing inventory movement and operational performance
- Risk: identifying dependencies and potential vulnerabilities
- Planning: connecting demand, supply and inventory information
- Optimization: evaluating potential scenarios and alternatives
- Management reporting: consistent visibility of the supply chain!
Connecting all these applications together is the idea of transforming raw supply chain data into meaningful information that informs a wide array of decisions.
Building a Data-Driven Supply Chain
Just as with any technology initiative, data analytics in support of the supply chain needs to include data, technology and supply chain expertise.
Combining data, technology and supply chain expertise is where the strongest results are seen.
However, improving the data foundation and increasing the organization’s analytics maturity first are most important. For the above-mentioned analysis scopes, and their corresponding applications, starting with simple reporting first is recommended. Once the organization’s data foundation has improved and sufficient analytics have been developed, more advanced types of analytics, such as predictive analytics, risk analytics, scenario analytics and even supply chain optimization, can be developed and added where meaningful.
Kepler Advanced Analytics supports organizations on their analytics journey by bringing together data analytics, analytics engineering, AI and industry specific supply chain and procurement knowledge to help them generate meaningful business insights from often disjointed and fragmented operational data.
The value of data analytics to manage a supply chain lies not in more dashboards or in more data.
It is about helping supply chain leaders understand their business better and make decisions based on better information.