What is the highest-return two weeks of work in pallet management? It is not a repair line upgrade or a better supplier. It is putting one person on a dock with a clipboard and counting for ten working days.
The count is also the best preparation for talking to suppliers. Anyone you request a quote from can quote far more accurately if you can tell them what you actually use.
What gets counted
Inbound pallets by footprint. Not "mostly 48x40". Actual counts, every shift, for ten days. Sites are routinely surprised; a facility that believes it runs almost entirely GMA frequently runs a noticeable share of other footprints.
Inbound condition profile. Of what arrives, what proportion is reusable as-is, repairable, or genuinely finished. This is the number that produces the biggest surprise.
Outbound consumption by grade. What pallets actually leave on, and whether that matches what was ordered.
Current spend. Purchase invoices for twelve months, broken by grade and footprint.
Current disposal cost. Waste hauling, skip pulls, whatever is being paid to make pallets go away. This line is frequently filed under general waste and nobody has ever separated it.
Loss. Pallets that leave and do not come back, where a return was expected.
What it typically finds
Repairable stock going in a skip. Many sites find a meaningful share of scrapped pallets are single-repair candidates, usually a companion stringer or one deck board. That is simultaneously a disposal cost and a replacement purchase, paid twice for the same pallet.
Grade over-specification. Almost every site buys one grade for everything, and that grade is often higher than most of the volume requires. Splitting the buy, with Grade A for customer-facing and Grade B for internal, can reduce purchase cost with no operational change at all.
Emergency purchasing. Rush orders cost more than planned ones. Most sites make several and nobody has ever totalled them, because they arrive as individual small invoices rather than as a line.
Footprint drift. A facility standardised on one footprint years ago and has been slowly accumulating others, each requiring its own stock position and none of them managed.
Nobody knows how many they own. This is near-universal. A count that reconciles pallets on site against purchases and returns frequently finds units nobody had recorded.
Why a spreadsheet will not do it
The data does not exist in a system. Pallets are the one material in most warehouses that moves without a transaction attached. They arrive under product, they leave under product, and nothing counts them in between.
That is why the answer is a person and a clipboard. It is genuinely low-tech and there is no substitute for it. Attempts to derive pallet consumption from ERP data tend to produce a number that is wrong by a factor that makes it useless.
Ten days, not three
Two weeks matters because pallet flows are weekly-cyclical. Mondays and Fridays look nothing like Wednesdays in most distribution operations. Three days of counting gives you a snapshot of whatever pattern those three days happened to sit in.
Ten working days catches two full weekly cycles, which is enough to see the shape. Longer is better but the marginal return drops off quickly.
What comes out of it
A baseline: consumption by footprint and grade, condition profile, current all-in cost per pallet issued including disposal, and a saving estimate with the assumptions written out.
And sometimes the conclusion is that nothing complicated is needed. If a site's volume is modest, a simple standing collection arrangement through a partner supplier may capture most of the benefit with far less complexity.
The part that is hard
The count is easy. Acting on it is not, because the findings usually implicate a decision somebody made. The grade was over-specified because a customer complained once. The skips are there because the previous site manager set them up. Emergency orders happen because a stock level was never set.
None of that is anybody's fault and all of it is fixable. But a baseline is uncomfortable reading and it helps to know that going in.

