Lineage Logistics is the second largest North American temperature-controlled warehousing company with more than 500 million cubic feet of freezer and cooler capacity. It has over 6000 employees and 110 warehouses across USA.
Lineage Logistics specializes in the cold storage of all major food commodities including beef, poultry, pork, seafood, ice cream, fruits and vegetables. Lineage Logistics warehouses are essentially gigantic refrigerators that store perishable goods. The cost of building and running such cold storage facility is 3-4 times the cost of non-refrigerated warehouses and as such, any in-efficiencies are magnified.
Those warehouses are kept (generally) at 0F, requiring hundreds of megawatts of industrial refrigeration. The design of the refrigeration system and their control systems have enormous implications on how much energy they consume.
As a lot of food consumption is seasonal, its facilities act as “buffer” between production and consumption of many perishable commodities. Examples of extreme demands are during holiday seasons such as Thanksgiving and Christmas when turkey and ham consumption experience massive surge. The only efficient way to handle this is to build inventory over a few months to cater for such this peak requirement during short periods.
They must design the Warehouses have to be designed to maximize the amount of products they can store while allowing operators to move product in and out. They must arrange products in the warehouses to minimize the travel time of forklift operators picking orders and putting away product.
Despite the obvious potential for data-driven decision making, Lineage, like the rest of the agriculture supply chain, has historically made decisions – via notional data and “gut feeling”. Warehouses were designed based on notions of product ﬂow and rough estimates of throughput requirements.
Summary of Results
Lineage Logistics’ data science teams were called in to help when one particular warehouse was experiencing capacity shortage. Faced with this option of extending the warehouse (at great cost/time), the analytics team were called in and decided to use data science approach to solve the issue and optimise usage within warehouse.
- 25% increase in warehouse density
- Re-configure warehouse at 10% cost of building new warehouse
- Reduced number of aisles traversed from 5 to 2
Warehouse Before Optimising
Lineage had to understand the distribution of pallet heights and how it changed over time. Such understanding was previously unfeasible for Lineage, as it did not have the underlying data in an accessible location and could not manage the joins / processing of the millions of records required.
Computing the distribution requires (literally) “playing back” every movement inside a warehouse over a period of years.
25% increase in
warehouse at 10%
cost of building
of aisles traversed
from 5 to 2
Warehouse After Optimising
To summarize, Lineage was able to increase warehouse utilization by around 25% by designing racks to match customer’s product profile and order patterns. An additional benefit, the reconfigured warehouse consumed limited additional electricity for refrigeration, while any new building would have required significant additional electricity.
After distribution analytics, Lineage finalised on different racking configurations for small, medium and large pallet loads to specifically cater for the diversity at its warehouse.
It was also able to precisely plan the frequency of small/medium/large pallet holes and distribution of them around the warehouse.
2/3rd Reduction in Picking time and Effort
At a particular distribution centre, where order pickers traverse the facility picking a specified number of cases from pallets containing various items and assemble those cases into outbound pallets. Such operations are extremely labour-intensive; instead of one forklift move to move one pallet, case-picking may require moving to 50 separate locations to assemble a single outbound pallet.
Lineage designed the processes in the warehouse to store items that move together close together, minimizing the number of aisles that the warehousemen have to traverse to fulfil a single order. Achieving this required analyzing all historical order data (several million records for a single customer in a single facility) and understanding the covariance structure of the item movements.
Lineage was able to drop the average number of aisles traversed to fill an order approximately by approx. 2/3rd. The stock clustering has resulted in significant labor savings and throughput increases, as operators are travelling significantly less to fulfil orders than in a traditional setting.
Lineage has implemented Cloudera’s Hadoop solution to build a business data hub in a public cloud environment. The business hub contains details of warehouse movements and is used for warehouse optimization and ongoing operation reporting.
Lineage has historically grown through acquisition and has multiple disparate Warehouse Management Systems (WMS). These WMS have inconsistent schemas and mismatching tables making data consolidation difficult. The initial task was to build a framework for standardizing data ingest and consolidation which could be easily replicated for the hundreds of tables in the 10-15 odd schemas within the WMS systems. This framework would serve as a baseline for future integrations.
This was achieved by writing custom tables and schema mapping scripts and Map Reduce jobs to create a consolidated final view of the dataset. Cross-links between tables and custom views were generated using Impala. Notable examples here are real-time views of incoming and outgoing product inventory, arrangement within pallets, pallet arrangement by rack within warehouse known.
Daily summarizations of state of warehouse, expected load, historical load and movement were used as key performance indicators. This was previously an ad-hoc process up to the discretion of the warehouse staff.
Cloudera Manager monitoring was critical in detecting errors and generating exceptions for Hadoop cluster on public cloud. Cloudera also helped Lineage develop a strategy for analyzing and optimizing transport routes for vendors storing inventory, based on historical and current origin-destination routes.
Prior to the Hadoop implementation, it was prohibitive to get data from WMS sources; one would have to ask multiple system administrators and receive data in different formats. Now it is automated and happens directly in the data hub.
The organization has shifted from “gut-driven” to “data-driven” after implementing Hadoop. New warehouses and re-designing existing warehouses were done using learning’s from millions of WMS transactions.
Delivering the Results
Such projects represent significant investment in technology and collaborations across multiple areas of Lineage Logistics. Lineage made significant investments in IT infrastructure (particularly in Cloudera) to provide the data critical to such design exercises. Once armed with the data, Lineage’s Data Science team analyzed the data and performed all of the computations necessary to determine the optimal rack design. Lineage’s warehouse operators collaborated on the design to ensure it was feasible and, once complete, ensured the rack was operated as designed.
Success in this and other projects has allowed IT the trust and support from the C-Suite necessary to allow it to operate as a “data-driven SWAT team,” called in by different business units as needs arise. Lineage’s Data Science team, for example, now consists of several PhDs in physics, mathematics and master-level engineers from different disciplines. This is a great example of multi-discipline, cross collaboration approach that can deliver breakthrough results.
BigInsights is a research & advisory firm focused on Big Data analytics solutions and technologies. BigInsights is focused on helping companies harness data driven innovation for customer and operational insights. See www.BigInsights.co for further details.
Cloudera delivers the modern platform for data management and analytics. Cloudera provide the world’s fastest, easiest, and most secure Apache Hadoop platform to help you solve your most challenging business problems with data. See www.Cloudera.com for further details.