I attended a useful session involving medical products manufacturer Arrow International along with a few consultants and academics on using detailed warehouse activity data to improve decisions around layout, storage modes, processes and automation.

At one level, there’s nothing new here, as many of us have used this technique for many years. Nonetheless, it’s always good to hear how someone has profited from such an exercise, and these kinds of sessions reinforce the truism that only a small percentage of distribution managers realty have or use order profile and activity data regularly – and that almost always, some assumptions about how the business is running, aren’t supported by the facts.

Warehouse activity profiling involves linking data from three sources:

	Line-item level order data
	Item master data, including weight and cube
	Warehouse layout and storage location data

Arrow used this information to support a new DC design, one which would end up using less than full pallet storage for many SKUs, selection of a full case picking process that used a medium choice of automation versus the full automated sortation system that was also considered, and SKU-based rather than discrete order picking for parcel shipments. The design also substantially reduced the warehouse footprint originally envisioned by company execs reacting to perceived needs for storage and fulfillment capabilities driven by sales growth.

A few key takeaways from the session:

	Cube data, even when availability in the item master, is generally the most unreliable data you’ll get – you must validate it.
	Plan on 30-60 days for working with IT to get the data, while the data analysis itself generally only takes 1-2 weeks.
	It’s helpful to plot the activity data over the warehouse layout, color coding to show levels of activity (says number of order picks). Several low cost programs exist to help do that. Georgia Tech is also almost complete with a free tool that enables companies to graphically plot pick travel paths across the facility – it’s available from Dr. John Bartholdi – email him at john.bartholdi@isye.gatech.edu.
	Make sure everyone is on the same page on what the terms mean – for example, in looking at activity volume by SKU, is the activity based on units, number of picks, or cube movement?
	In today’s world, focus on flexibility. Perform sensitivity analysis around potential changes in assumptions around volume growth and order profiles. In Arrow’s case, they had to take on customer direct business after the company stopped using many distributors, which changed order profiles, and had to consider possibly closing a west coast DC.

What do you find are the keys to using activity and order profile data to improve DC design and process decisions? Should we be using it more frequently