Most warehouses will be sitting on data which can create immense value for their operations. Whether its tracking current performance, looking out for exceptions to manage, improvement opportunities or a basis to use for the design of the replacement warehouse, all these situations would benefit by utilising the information contained in your operational data. That’s where Warehouse Analytics comes in.
It’s important to know what to look for in the data
Similar to how Supply Chain Analytics (as its name implies), covers the analysis and modelling for various functions throughout the supply chain – such as inventory analysis and forecasting, network modelling and route optimisation, Warehouse Analytics can be considered as a subset to Supply Chain Analytics, focusing on the warehouse. As an example, this can cover analysis of current operational throughput, order profiles, material flow within the warehouse and modelling such as warehouse sizing (for new warehouses), storage equipment selection and SKU slotting. It’s important to know what to look for in the data and which measures create the most value to your operations and provide the most insight.
Analytics and similarly Warehouse analytics can be broken up into 4 broad categories or types of analytics:
- Descriptive – visualise and summarise data. Display trends and exceptions.
- Diagnostic – analyse and identify problems, trends, exceptions and why these occur.
- Predictive – predict future performance requirements.
- Prescriptive – provide optimised decision support.
It’s important to note that the above categories are not mutually exclusive. Especially when it comes to descriptive analytics. Descriptive analytics should always be the first step in any analytics process and which is why I’ll focus on it today.
Descriptive Analytics
Descriptive analytics is the stage when the current situation is visualised and summarised allowing you to identify any patterns, trends or exceptions. If it’s part of a greater analytic solution, for example, creating an optimised SKU slot recommendation system (which can be classified as a prescriptive analytics solution), descriptive analytics will help define the problem and refine the potential solution. Within Warehouse analytics, Descriptive analytics would entail the calculation of a variety of measures to summarise performance, order profiles and SKU profiles. Providing numerical and visual summaries allows the reader to identify trends and patterns within the data. Listed below are some of the measures that can be calculated from warehouse operational data, which we like to call Standard Analytics:
- Daily/ Weekly/ Monthly Performance (Order lines, orders, units, SKUs)
- Pareto analysis
- Order profiles
- Lines per order
- Units per Order
- Units per Order Line
Depending on the current/proposed processes and available data, it’s also a good idea to analyse these measures categorised by order type, SKU category, or pick type (eg: is an orderline for 10 units picked as two cases of 4 units each and two eaches or is it picked as 10 individual units) and any other categorical measure unique to your organisation. This creates an operational summary of your warehouse which you can use to identify specific categories of orders or SKUs that may require further attention or have improvement potential.
If you use a dashboard to keep track of warehouse performance, including some of the above measures will complement your existing KPIs and assist you in tracking and explaining operational trends and exceptions. Remember to keep in mind some of the Dos and Don’ts when it comes to Data Visualisations which I covered earlier.
Descriptive Analytics at Fuzzy LogX
At Fuzzy LogX, we work with multiple customers each with their own unique requirements and goals. With that comes different WMS/ ERP systems, each with their own data fields and formats. For example, ‘SKU ID’, ‘Item Code’, ‘Product No’ and a few other combinations and terms all mean one thing – a unique identifier for a SKU. And did we mention the increasing amount of data as well? For all these customers, we always start with identifying what the current situation is and looking to see what the trends, problems and exceptions are – that’s Descriptive Analytics! Similar to how we always look to improve processes for our customers, we wanted to improve our internal processes as well, so that we can spend more time helping you!
Say hello to SAM!
SAM
Our Standard Analytics Module (or SAM as we like to call him) is our solution to speed up and improve the analytics process for our customers. SAM essentially provides output for Descriptive analytics, but as part of achieving that, includes a standardised foundation for any other analytics we may perform on the data, be it diagnostic, predictive or prescriptive.
standardised data model of customer data
SAM first creates a standardised data model of customer data so that all our downstream models and algorithms can refer to a standard set of attributes and require a minimal amount of changes for routine analysis we may perform. We have a list of attributes we use by default and we initially map customer data to our list of attributes. It’s important to note that prior to giving SAM the data to process into the standard model, we ensure the data is clean and arranged in a suitable form. Then SAM does the rest of the job, setting up the various data tables and views within the model. Whether creating visual summaries of data, setting up dashboards, or even conducting a SKU slotting exercise, it helps having a standard model to refer to rather than wasting time changing all references from ‘SKU ID’ to ‘Product Number’.
Once the data tables and views are setup, SAM can then proceed to provide us with the Standard Analytics we mentioned earlier. On every project, we would also supplement the standard analytics outputs with additional visualisations and summaries depending on the specific goals and requirements of the customer.
Automating the process not only makes it faster but allows for reproducible analysis. That is, changes can be made to the underlying data and the results and visualisations can be updated promptly thereafter. The changes also flow-on to downstream models and algorithms . Which means, once we complete the initial setup for example, annual data for 2017, you can come back to us with annual data for 2018 and it would be a much quicker process to obtain the same outputs.
SAM has also been setup to allow easy scaling up, so we can easily scale up our computing resources to perform more complex analysis or work with larger datasets. For this we can use our very powerful in-house PC or even leverage the capabilities and resources of Cloud Computing providers.
If you want to see what SAM can do for you, contact us here.
About the author: Yohan Fernando is the Manager – Systems & Data Science at Fuzzy LogX who are the leading warehouse, logistics, and process improvement consultants in Australia. Fuzzy LogX provide project management & consulting services, leading-edge data analytics, process improvements, concept design & validation, solution/software tendering, implementation and solution validation services to businesses with Storage & Distribution operations looking to improve their distribution centres.