Types of supply chain analytics
By Siena Analytics
Supply chain analytics is reshaping modern supply chains, and it’s easy to see why: It provides useful, actionable data that can help improve efficiency for individual companies and for the ecosystems in which they operate. It is a necessary precursor for effective automation, and for meeting the rising demands of customers in an increasingly complex, omni-channel market.
But supply chain analytics is not one catch-all thing. Rather, it includes five main categories, all of which can help supply chain organizations adapt and evolve in a changing marketplace. As early as the 1990s, supply chain analysts were using technology-driven analytics to optimize supply chain functions. The 2000s brought the advent of predictive analytics, and by 2017 supply chains had the ability to access roughly 50 times more data than was available just five years earlier. That’s driving continued advances in applications such as machine vision, natural language processing and machine learning.
What is supply chain analytics?
Supply chain analytics is the application of high-level intelligence derived from an organization’s data at various points in its supply chain, from procurement and processing to inventory management, distribution and beyond. This method of analysis is critical to telling a detailed story about operational processes at every level, and provides a foundation for efforts to automate and continually improve logistics operations. It can help companies enhance and optimize the effectiveness and efficiency of their supply chains in various dimensions:
- Individuals: Measure productivity and other people-related data
- Resources: Evaluate the flow of raw materials, software and systems
- Activities: Optimize physical and digital processes at each stage of the supply chain
- Technologies: Streamline software and hardware to support each of the above
Supply chain analytics relies on data collected at all touchpoints of the supply chain, from sourcing and manufacturing to logistics and customer service. This data then informs key operational decisions on purchasing, scheduling, holding or carrying capacity, staffing and more.
There are five common types of supply chain analytics that create a hierarchy of capabilities for distribution leaders, supply chain managers and other decision makers:
- Descriptive analytics
- Predictive analytics
- Prescriptive analytics
- Diagnostic analytics
- Cognitive analytics
Each of these types of supply chain analytics can be adapted, combined and customized to provide targeted solutions for businesses, based on their most pressing operations needs.
Descriptive analytics is a method of aggregating existing sources of data and applying statistics and modeling to understand an organization’s past and current data, and displaying it via dashboards in the form of graphs, charts and reports to answer questions about the current health of a business. Most major supply chain participants, from manufacturers to retailers, have used descriptive analytics for many years to help forecast demand, plan for impending weather events and set prices. But although descriptive analytics is generally thought of as the baseline of analytics types, many organizations are just beginning to apply it to logistics.
Examples of descriptive analytics
Without supply chain analytics, supply chain operations are much more dependent on the skills and vigilance of key personnel to spot worrisome trends. By capturing and looking for patterns within operational data, descriptive analytics can automate the process of understanding what has and is happening. For example, distribution centers capture bar codes on products for tracking purposes. If “no-reads” on a particular tunnel suddenly drop from 98% to 60%, descriptive analytics could be set up to check current data inputs for possible causes, such as a broken bar code scanner or malfunctioning cameras used to identify the location of the bar code on packages as they move through the distribution center.
Descriptive analytics uses historical data to understand:
- Fill rates
- Supplier lead times
- Inventory dollars
- Sales and operations data
- Social media engagement and conversion
- Survey results and other forms of user generated content
Why is descriptive analytics important?
Descriptive analytics can be thought of as the baseline of supply chain analytics. Making historical and “now” data easily accessible lets team members use their own knowledge and experience to make more insightful decisions in the course of day-to-day operations. They can also add new data points to be measured, adapted and evaluated on the fly.
The value of descriptive analytics empowers supply managers to quickly make sound, data-driven decisions.
While descriptive analytics focuses on the past and present, predictive analytics focuses on the future.
This applies complex mathematical models to large amounts of historical data gathered to perform descriptive analytics, to help supply chain managers predict what will happen in the near future. Data algorithms and machine learning play an important role in creating these models, which can be used for a variety of forward-looking applications:
- Goal setting
- Planning, scheduling and forecasting
- Performance and risk management
- Fraud reduction
Examples of predictive analytics
Complex machine learning models, like those used in predictive analytics, can be modified and reapplied to understand and respond to emerging trends. For example, predictive analytics can give supply chain managers a view on how a weakening economy will impact sales, so they can reset their staffing and inventory plans. Predictive analytics can also help retailers make accurate recommendations for customers.
Why is predictive analytics important?
Predictive analytics can give businesses a measured, behind-the-scenes glimpse into the future and synthesizes information to provide insights to the right people at the right time, saving time and money across all touchpoints of the supply chain. While the predictions are not always precisely correct, using predictive analytics often points the way to increased efficiency and lower costs.
Prescriptive analytics builds on predictive analytics by taking a deeper look into the future, comparing various scenarios to provide insight to a supply chain manager as to what his next step should be. This approach uses more sophisticated machine learning techniques and typically requires more data to effectively anticipate various outcomes. For example, by gathering and analyzing enough data, companies can essentially create a digital twin of their warehouse ecosystem, enabling them to run simulations to test different approaches. This way, a distributor could test whether shutting down a tunnel during slow periods would improve productivity without impacting customer satisfaction — without having to actually run the experiment.
Examples of prescriptive analytics
Prescriptive analytics is especially suited for situations involving massive data sets, such as GPS information or digital consumer behavior. For example, a large e-commerce retailer could use prescriptive analytics to:
- Run simulations to test the impact of operations decisions, such as reducing staffing or adding more sorting equipment.
- Respond more quickly to changing conditions, such as how to change staffing in the distribution center to deal with an unexpected surge in unreadable bar codes.
- Take corrective action more quickly. For example, if throughput in a particular facility is slowing down, predictive analytics can suggest or even initiate responses, such as shifting volume away from a tunnel that is experiencing the slowdown.
Why is prescriptive analytics important?
By crunching more data with more sophisticated models, prescriptive analytics can be used to provide specific recommendations to supply chain managers rather than just provide predictions on what might happen. This approach is often particularly useful for understanding how different scenarios will impact downstream inventory levels, allowing for smarter procurement and staffing decisions. This approach often leads to faster decision-making, enhanced productivity and reduced financial losses.
Diagnostic analytics gives supply chain managers the tools to understand when data is telling a story they don’t understand. This form of analysis uses in-depth data mining and correlation analysis to answer “why” questions, such as why have no-reads increased for products from Acme Co. When paired with powerful visualization technology, it can be used to explain data anomalies, and to better understand deviations from averages, trends, expectations or norms. It differs from other types of analytics, like descriptive analytics, in its ability to isolate specific events within the supply chain and answer important questions managers may have about how and why sales in a certain region have been impacted, for example.
Diagnostics analytics is increasingly important in the logistics realm, as a way to continually improve automation within distribution centers. The job of these facilities is to make sure products reach their ultimate destination as quickly as possible, whether it’s a home on the other side of the country, a store across town or a shelf in the warehouse next door. To do this in an automated way, hundreds of elements must work according to standardized processes, including conveyors, cameras, sensors, data capture software and the hardware it runs on, and the underlying networks and hardware, not to mention the policies and procedures followed by distribution center personnel. When automated systems break down, diagnostic analytics can quickly identify the root cause and suggest fixes, whether the problem is a sorter in need of maintenance, a broken sensor or a worker who is not following procedures for how to place products on a conveyor.
Examples of diagnostic analytics
Diagnostic analytics is increasingly needed to maintain the ability of logistics networks to operate efficiently, and in an automated fashion.
- Quickly troubleshoot a problem within the distribution center, such as an increase in no-reads or an increased percentage of products requiring human intervention.
- Suggest new supply chain configurations to address a problem at a particular line or facility – say, by adding more sorting equipment or increased staffing temporarily.
- Identify non-standard products, and provide insight on how to optimize the supply chain to handle them more efficiently.
Why is diagnostic analytics important?
Without insight into the cause and context of problems within data sets, businesses would be flying blind. This method of analytics can be used to troubleshoot areas of concern before they become big problems, more quickly spot emerging opportunities for competitive advantage and improve supply chain efficiency and effectiveness.
Cognitive analytics is an emerging technology that is focused on drawing out important, contextual conclusions from data — ideally, to approximate what a human being might do when presented with the same data. It relies on methods including data correlation, contextual awareness, semantic relationships and machine learning to make nuanced conclusions from varied data sets, such as groups of phrases or specific objects within images.
Examples of cognitive analytics
In the supply chain world, cognitive analytics is being used in a number of ways to integrate advanced technologies such as robotic process automation (RPA) and AI, essentially allowing systems to “see” packages and make logic-based actions without perfect information.
- Robots are able to sort, pick and package silicon chips, car assemblies and other parts and final products, based on quality assurance standards, and reject those that fall short in some way.
- AGVs (autonomous guided vehicles) can avoid or navigate around unexpected obstacles, to deliver products where they need to be in a distribution center.
- Cameras and other sensors in sortation tunnels can determine if a shipping label was positioned properly but was obscured by the manufacturer’s packaging. Organizations sometimes use cognitive analytics to make decisions on how to repackage products to prevent such problems in the future.
Why is cognitive analytics important?
Cognitive analytics allows businesses to extract more meaning from data, effectively turning previously unusable or superfluous information into actionable insights. It represents the cutting edge of supply chain analytics, and promises to let companies scale the intelligence, experience and judgment of top employees. While it will never fully replace human expertise, it will undoubtedly help large supply chain operations make better decisions.
Understanding — and leveraging — the different types of supply chain analytics will become indispensable as market forces pressure businesses to adapt quickly to changes in the market. As big data technology and artificial intelligence becomes more affordable and more scalable, all of these types of supply chain analytics will play an increasing role in solving complex problems and ensuring long-term operational success.