Supply Chain Management or SCM refers to the handling of an entire flow of goods or services, starting from the raw components all the way to delivering the final product to the consumer. This domain consists of various activities including Demand Planning, Sourcing, Production, Inventory Management and Storage, Logistics, and returning excess or defective products.
Various Enterprise Resource Planning (ERP) solutions are available to streamline these steps. In fact, ERPs have become an important part of SCM. Their integration not only allows manufacturing and distribution businesses the ability to gain greater visibility into all operations, it has also increased speed, efficiency and overall customer satisfaction. ERP’s feature-rich working environment combined with an efficient workflow of an SCM provides a range of important advantages. But along with the many advantages, setting up ERP and working with it brings some issues as well.
Challenges with ERPs
ERP software is a business technology, on its own it doesn’t cause failure, but inefficient implementation and handling can severely hamper the ROI of an ERP system. Some of the Key areas that can cause ERP implementation failure are:
- Data conversion while moving data from an old system and mapping it to the new ERP is the biggest challenge a business face during an ERP Project.
- If you plan to use ERP, then early in the project itself you’ll need to think about what other systems need to integrate with your ERP system and have a solid plan for integrations.
- Since ERP’s ROI depends on the trained hands that are using it, training becomes an integral part. Internal teams and end-users have to be made comfortable with the new software.
- Project planning is crucial to the success of ERP implementation. Setting up a methodology for its use and following it is also fundamental to avoiding failure.
Even with these setting up issues, ERP solutions have spread it’s reach to almost all businesses. But just because it is widely used does not mean that it is the only solution to carry out an efficient SCM.
KNIME to the rescue
KNIME is an open source data analytics, reporting and integrating platform. It provides an easy to use GUI to create your small modular pipeline by assembling different types of nodes. You can integrate these pipelines with other components for various data mining, analytics or Data science activities.
KNIME finds its use case in a variety of different scenarios, from simple pre-processing of Big Data to creating forecasting solutions. But apart from these, KNIME also finds its place in the domain of Supply Chain Management.
Using KNIME, you can create our own customized efficient modular pipeline for either one or more steps of Supply Chain Management. And if the need arises, you can easily plug that pipeline in an existing ERP solution in place. From Demand Planning, Logistics Planning, Sales Forecasting, KNIME can be used in any of these steps of SCM. But the one we will be focusing on is Inventory Management.
Inventory Management with KNIME
Inventory Management is the process of ordering, storing, and using a company’s goods to meet consumer’s demand levels with an appropriate amount of supply. Thus a part is to have an efficient Inventory Movement Plan in place. This plan describes an approach to move the goods from a central warehouse to distribution pods. The mechanism used to create a movement plan is central to each organisation’s process. It can vary from having an Excel spreadsheet to having a sophisticated ERP solution.
An alternate approach can be to use KNIME to not only create an inventory movement plan, but also update it according to new factors that might become available in the future.
With KNIMEs different types of nodes, ranging from simple data transformation to creating reports, an inventory management plan can be easily built. It’s simple drag and drop mechanism gives the flexibility to easily update a solution with any new condition.
Inventory Movement Plan
One of India’s most prominent On-Demand Food Delivery Platform wanted to create a pipeline that given the current stock in central warehouse and distribution pods, would decide on how many units of each item should be shipped to different pods. Along with that, they wanted the flexibility to add more factors in calculation process in an easy approach without hampering the existing pipeline. As a status quo, they were using Excel sheets for their calculation, but this approach made the process tedious with a lot of manual intervention.
As an alternate, we proposed to leverage KNIME to develop the pipeline. This pipeline takes a user selected file containing each pods available stock units for each item and its warehouse units. The pipeline performs calculations using some factors, and updates the units to be transferred for each item to different pods. As an extra advantage, with KNIME we deployed the pipeline on the server as a Web Guided Application, that allows the end user a friendly interface to track each step of the process and visualise the final result. Also, KNIME deployment allowed us to schedule the workflow execution to further reduce manual intervention and send final reports to interested stakeholders.
KNIME supports ingesting data from various sources, from a simple excel file to reading data from a number of databases. For our use case, the workflow consumes data from an excel sheet that can be selected by the user at run time.
The different data sources we will be handling are:
Warehouse and Pod Stock Units: This data sheet contains information about each items stock units in warehouse and distribution pods. Also, it contains information about the current and projected sale rate of each item in different pods.
Pod Threshold: Each pod has a threshold value for different categories of items. This signifies the lower limit of an item that should be available in a pod.
Transfer DoH: The number of days an item category takes to reach different pods.
CutOff DoH: The number of days beyond which an item should be initiated for movement.
Priority for Movement: Priority of Pods to which item should be sent.
Following this, the second step in the workflow is cleaning the data. This involves transforming it’s structure in a way that can be easily manipulated. KNIMEs core nodes comes in handy in such basic data cleaning, restructuring and filling in empty values.
Once cleaned, the initial data looks like this:
Once we have data in the right format, several manipulations can be implemented according to the use case. For our case, we have done the following:
- Allow the user to select the initial input file and then select one of the items for which movement plan is to be created.
- Once selected, calculate the current holding for that item in each pod. This step calculates the number of days an item with its current pod stock and run rate can sustain.
- On the basis of the item’s current DoH and cutoff DoH, we decide the pods to initiate transfer for.
- Using the transfer DoH and projected run rate, we then calculate the number of units to transfer for each pod.
- If the total number of units to transfer is less than the warehouse stock, we finalise the movement plan.
- In case the requirement is more than the warehouse stock units, then we take in account the current DoH for each pod and it’s priority in the priority list to finalise the movement plan.
Share the Plan
Once the workflow completes execution on the server, it can be configured to share the plan with other shareholders via a notification to the e-mail.
Deploy the Solution
Once done developing, the inventory management workflow is deployed on KNIME server as a Guided Web Application. This gives user an interface to interact with the workflow. Using the web application, the user can
- Select an item
- Change thresholds for an item that are by default set at the item class level.
- Finally, visualise the Final Movement Plan
Once Deployed as a Guided Web Application, different stakeholders of the business are brought together under one platform. The end user can finalise a movement plan for any item interactively and intelligently.
The complete workflow can be found here.
Advantages of using KNIME
A Supply Chain Management Component developed with KNIME has various advantages over traditional solutions:
- Configurable: Traditional approaches can be too rigid to adapt to a specific workflow and business process of some companies. On the other hand, KNIME allows us the flexibility to change the parameters, data set or underlying process in a matter of few hours.
- Faster Processing with Big Data: KNIME provide nodes to easily integrate various Big Data Technologies to the workflow. Whether it’s reading large amount of unstructured data from a NoSQL Database, or performing complex analytical operations on Big Data, KNIME provides a support for all such cases.
- Inclusion of Random events: Since KNIME solutions are configurable, adding factors related to random effects can be easily done. Not only developing a workflow, but updating it too is an easy task.
Since KNIME provides support for various ingestion, transformation, calculation and reporting use cases, it makes developing a workflow a much easier task. We can create these workflows or pipelines independently and finally plug them together for an end to end implementation. Like Inventory Planning, KNIME can be easily used to create a workflow for sales forecast, demand planning, or any other step of Supply Chain Management. And since you will be developing this solution from scratch, it will be fully customised to your business requirements.