Benchmarking When Measurements Becomes the Crime
“But mom, all the other kids are doing it!” We’ve all heard and probably quoted that line at one time or another. And just as inevitably, we have all faced one of the two most likely retorts, “I don’t care what other kids are doing, they’re not mine,” or my personal favorite, “If all the other kids jumped off a bridge, would you?” To which, in the days when my wit overshadowed my judgment, I usually answered, “Well, that depends on how high the bridge is, whether there’s water under it, how deep the water is ….” I’m not sure how I survived my childhood.
Either way, the answer was clear — what other children did never justified duplicating their actions. In all likelihood, it was the weakest argument one could proffer to a supervising adult when seeking permission for some as yet never before allowed behavior or activity. So given this almost universal youthful experience, why then does the exact opposite logic apply to most applications of benchmarking in the adult arenas of distribution and logistics?
Most of us have had conversations about operational performance benchmarking with our colleagues or have reviewed it within our industry with the hope of understanding how our operation performs. Confronting a productivity rate from another operation that is significantly higher than the same functional rate your operation achieves can be disheartening or motivating, but it should be neither until you know more about the benchmark rate than just the number of cartons or pallets picked per hour that it purports to be the ‘industry standard’ or ‘best in class.’ The value of a benchmark productivity rate, or an order cycle time rate, to the operation evaluating itself is proportional to the similarities between the operation or operations that established the benchmark and the operation benchmarked against it.
There are some very simple questions to pose when evaluating any benchmark, the most important of which is: “Does that operation’s benchmark apply to my operation?” not, “How can I duplicate or surpass that performance?” Answering the first question will keep you from wasting time chasing performance benchmarks that are completely irrelevant to your operation. In turn, that allows for the devotion of effort to benchmarks that are bona fide targets for a particular operation and that should be pursued to elevate an operation to higher levels of performance.
To be clear, the benchmarks we are talking about here are distribution center facility operating benchmarks. These are most commonly metrics around productivity, but we also frequently focus on throughput, order cycle time, accuracy, storage density and even shrink percentages. Ground level or narrow benchmarks are the simplest to discuss, which occur when only one or two variables have an impact compared with broader benchmarks.
Broader benchmarks, such as cost per unit shipped, are affected by all the variables of all the productivities for all the tasks required to receive, store, replenish, pick and ship an item, as well as the wage rate, facility operating costs, rent and other fixed overhead. Those broader benchmarks are no less subject to misapplication, but it is much easier to point out the pros and cons of benchmark application when discussing a straightforward or narrow benchmark — such as piece picking rates.
First, let’s agree on a unit of measure for our benchmark rate — order lines picked per hour. In the benchmark examples listed above, let’s assume the operation being evaluated is picking 100 order lines per labor hour, and the benchmark rate that’s been published by the industry or other esteemed expert (either as an average rate from many surveyed operations, or as singular point of comparison to another operation) is 200 order lines per hour. A well-intentioned manager, upon learning another seemingly similar operation is picking 100 more order lines per labor hour, would be remiss if they didn’t ask why their operation was not equaling that performance. But let’s examine the operational characteristics around those rates. We might want to understand not only why the rates should be different but also why the 100 order lines per hour rate might represent a more efficient operation than the 200 order lines per hour rate.
SKU count — It’s very simple math. The more SKUs an operation must pick from, all other things being equal, the longer the pick path will be. If one operation has 200 SKUs and another has 2,000, the latter’s pick path may be 10 times longer than the former. Given that half of pick labor is travel along the pick path, it is easy to understand how pick rate would suffer from a higher SKU count.
Pieces picked per order line (order profile) - Assuming the pieces per order line are still significantly less than a full case, an operation that must pick one each per order line will be less productive (from a piece/hour perspective) than an operation that picks three units per order line. Ironically, the operation picking three pieces per line might have a lower order line per hour productivity rate, especially if the three pieces require three ‘grabs’ or ‘reaches’ per line, as opposed to only one reach for a single piece order line. Finally, the number of pick location stops to gather all the items for an order or order package (considering the same pick path distance) will affect the overall order line per hour productivity. All other things being equal, more ‘stops’ in a pick path to assemble an order translates to lower productivity.
Item characteristics consistency — At the simplest level, size does matter! The size of the item picked affects pick rate. From how many units can be collected before a new pick container is needed, to how many pick faces an operation can present per linear foot of pick aisle, productivity, throughput and order cycle time can easily be affected by the size of the picked unit. High variability in size or other characteristics, such as fragility, can force either more careful picking or force a picking sequence to avoid product damage. You wouldn’t want to pick the light bulbs before the crow bars! Confirm that the benchmarks to which you are comparing your operation are roughly equivalent for both overall size and consistency in size. The more akin they are to your operation, more likely the benchmark is one against which you should measure your operation. This is also a further complexity sometimes driven by a high SKU count when compared to a low SKU count. The more SKUs an operation must support, the more likely it is that those items will differ in their size, weight and levels of fragility.
Overall volume — It is not so much the volume itself that might make one operation more productive than another (although there is something to be said for simple economies of scale) but the higher volume provides an operation with picking strategies that are less effective in low-volume operations. Strategies such as batch picking, even in a manual environment, pay greater dividends and offer more productivity economies, with a greater population of order from which to build a productive pick batch. For example, an operation with six pick aisles and enough volume may be able to batch their orders by aisle and achieve more significant productivity gains than an operation with lower volume demands.
Mechanization and automation — When order volume and labor are high enough, there is the potential to apply mechanization or automation. Mechanization might entail a conveyor system transporting order cartons to pickers who pick only from a select number of items into those cartons presented to them by the conveyor delivery system. Automation might include an automatic storage and retrieval system (ASRS) pulling out only those items needed, in the sequence appropriate to that day’s order requirements, and present only those items to a picker for order fulfillment. The productivity potential can approach 10 times a manual pick approach, but the investment in technology only makes sense if there is enough labor avoided to pay back the outlay. A low-volume operation, with 10 employees, could not pay back a multimillion-dollar ASRS application necessary to house all its items, but you might misconstrue a comparison of only the order line pick rate achieved in each operation as a failure on the part of the smaller operation with the lower pick rate. The fact might be: the smaller operation would fail if they invested in technology they could not fiscally justify.
Though not exhaustive, all these factors can contribute to a particular metric, such as order line pick rate, having very different productivity levels in two operations and there being nothing wrong with those differences. The level of similarity between the operations used to develop a benchmark and the operation to which the benchmark is compared will determine if it represents valuable insight that can allow an operation to work toward world class capability, or if that particular benchmark for your operation is nothing more than what the other kids are doing.