For those of us in the process and facility design business, data is the life-blood of our existence. It is our primary raw material. But, it certainly can be a fickle friend.

The general attitude about data is, "the more, the better." But, we can easily become overwhelmed with the detail—and, with verifying its accuracy.

I once asked a group of university professors who teach subjects in materials handling and warehousing—"If you had to choose just one piece of data that might be used to design a handling process, what would it be?"

Of course, it was a hypothetical question. But the answer would certainly be useful if the problem was a simple point-to-point move. In a more complex environment, the answer could be used as a measure of scale, especially when the industry is already well understood.

But, this seemingly trivial question does raise the issue of a data gathering hierarchy. And this is important not just to consultants but to people such as yourself who need to better understand your operations. So here are a few suggestions on making data a friend.

Get the big picture first. How many truck loads come in, and how many leave? How many products are there? And, how many are active? How many orders are processed? Then, gradually dig into the details to evaluate specific hypotheses and alternatives.

When the analysis doesn't match your "gut feel," you need to check to make sure that there isn't some data missing, or that a spreadsheet formula might be wrong. On the other hand, don't be lulled into a false sense of security when the analysis quickly confirms your favorite hunch. Check again here, too, for you still could have made an error. Continually reconfirm the results against the big picture to be sure that you still have all of the data—and, the correct units of measure. It's always a big surprise when the linear feet in a spool of wire show-up as cubic feet.

Of course, beware of averages. Grouping data into "buckets" that are meaningful for materials handling system design helps you see the problem in real handling units. How many of the orders will fit into a single carton, or tote? How many will fit on a single pallet? These help you to visualize the picking process.

For instance, grouping the products by cubic feet of daily throughput for each stock keeping unit (SKU) is the most useful measure for designing pick face sizing. Pick lines per cubic foot of throughput helps us to concentrate on the fast movers.

Then there's the matter of pinpointing peaks. Many times it has turned-out that the peak day was a result of reporting Saturday's work along with Friday's or Monday's. Time stamps from scanning transactions tell us not just the peak hour of activity, but the peak minute, as well.

Fast computers have made it possible to analyze huge databases. However, they will answer only the questions that you think to ask. To get a real feel for activity, there is no substitute for manual sampling and tallying of data from process documents; receiving packing lists, production routings, replenishment records, to name a few sources. These samples and observations usually prompt you to search the data again to quantify another transaction type.

As much fun as it is to collect and use, don't get too comfortable with the data. Data is about yesterday. But, you're designing for tomorrow. Projection of future volumes and analysis of changes in product and order mix are critical to making effective use of history.