Big Data for Big Inventory
While traveling recently, I had a chance to read an article (Meet a Start-Up With a Big Data Approach to Hiring) in a magazine about how Big Data is being used to help HR divisions in organizations predict the outcome of new hires. Whitetruffle provides a data service using a proprietary model which analyzes 50 categories of “signals” in a job candidates profile. The claim is that the more data they get, the smarter the model becomes. The result is that you end up hiring better employees. While I’m not an expert in the “Big Data” trend, I summarize its purpose as simply “more data to analyze, means better results”. Big Data analysis can be a valuable tool to selecting the best performers for your company, but it can also be used for selecting the best performing method of picking, receiving, and bin strategies.
In my career, I have spent years reviewing all types of companies that have inventory. They create it, they buy it, they bundle it, and ship it. The truly amazing thing about it, is that I rarely see the same solution needs for two companies. Even in one company, recommending the best method for a process can be very perplexing. Using volumes of data grouped into inventory related categories of “signals” we can match up which method of picking for example should provide the best results.
During my trip, I visited a company that employs temporary workers whose jobs are to pick orders in the warehouse. Some pickers are very knowledgeable about how to find products and pick orders; others pickers are relatively new. The warehouse has over 40,000 SKUs and many of the cases are shipped in and arrive at the dock as a hodgepodge of boxes. Some combinations of pallets are more random than others.
Should the receiving dock spend the time to break down the pallets or should be pallets be put away as is?
Using volumes of data to analyze, the system could consider the profile of the receivers, and pickers at the facility. The answer to whether to breakdown pallets may be made in real time using this data. It may even vary by pallet.
To select the best answer, the system might consider signals such as the workers skill sets, number of bin locations versus empty bin locations, the ordering statistics such as number of lines and frequency of these items, and whether FIFO picking is used or Primary Pick locations.
Big Data certainly has a place in giving distributors and manufacturers the best methods to accomplish a set of tasks. In the future, like the start-up company Whitetruffle, using Big Data to provide a service to help make hiring decisions, I can see companies subscribing to a service that collects member data securely and in return provides key data analysis across all the members to help each one make the best decisions on whether to single pick, batch pick, or wave pick even down to the pallet level. Sounds pretty cool!