Reading a precast yard as a flow, not a warehouse

Placeholder hero for the precast-yard kaizen entry, to be replaced with a photograph of the yard and its stacked precast product. — The will-call wait wasn't a loading problem — it was overproduced inventory eating the yard space the loading needed. A five-day kaizen that cleared the yard and put the high movers by the trucks.

This entry describes a five-day kaizen at the customer-facing yard of an independent precast concrete manufacturer in the western United States, run in the spring of 2019 — the open-air yard where finished concrete drainage structures, manholes, and cast-iron rings and covers wait to be loaded onto the trucks that carry them to the job. The brief was the one the yard had carried for a while: customers were waiting too long at will-call, and the inventory system could not be trusted. The kaizen’s conclusion was that the wait was not, principally, a loading problem. The client’s name has been withheld at its request; customer names that appear in the yard’s own load-time records have been omitted.

The yard is the last operation before the customer. Precast concrete — drainage inlets, manholes, barrels and cones, the cast-iron rings and covers that finish a structure at street level — is cast and cured on the site and moved out to the yard to wait for a truck. Two kinds of truck arrive: the company’s own delivery trailers, and the customers who come to collect in person at will-call. The yard came into the event carrying three complaints. Will-call customers waited too long. The inventory system frequently disagreed with what was actually on the ground, so a customer could be promised a part that then could not be found. And the operators were writing too many manual purchase orders by hand for the small pick-up items — slings, banding, joint wrap, pipe — that customers bought alongside the concrete. Three complaints; a charter with five objectives spread across them. The temptation in a yard like this is to attack all five at once, and to measure the week by how many were touched.

The Tuesday of the event was spent measuring, not fixing. The team stood at the gate with a clipboard and timed a morning’s worth of customers, marking each one at every step: into the yard, into the office, out of the office, over to the product, loaded, out of the gate. The numbers were not what the complaint implied. On a smooth morning the average customer was in and out in about 30 minutes, and roughly half of that was the load itself; three-quarters of the trucks were already clearing the yard inside the 45-minute target the charter had set. The wait, where it existed, was not in the paperwork and was not a slow loading crew. It was in the distance the forklift had to travel and the order in which the product was stacked — and both were a function of how full the yard was. A walk of the ground showed why. Somewhere over the years the yard had stopped being a place product passed through and had become a place product accumulated. Seven cast-iron line items alone carried close to $120,000 of stock the system had no order for. There was an “Island of Misfits” of orphaned parts nobody was moving; a heap of rusting galvanized scrap in the weeds along the fence; drainage structures set down where the forklift had happened to leave them rather than where they belonged. The high movers — the parts most customers actually came for — were not up front by the load area. They were behind the slow ones. The binding constraint on how fast a customer got out of the yard was not the loading. It was the overproduction standing in the loading’s way.

From Wednesday the redesign treated the yard as what it was meant to be — a flow — and asked what stood in the way of it. The team re-drew the yard on an aerial photograph. A dedicated trailer-load area was cut close to the highest-moving product, so the forklift travel between stack and trailer fell. The scattered cast iron was consolidated into one location. The high movers were brought front-and-center. A staging area was marked off on the ground so it could not be colonized by whatever needed setting down that day, and an “EXIT” sign and routing were added to move truck traffic one way through the yard rather than against itself. The dead weight came out: the misfits island cleared, the galvanized scrap slated for disposal, the specials and the Armorcast parts each given a defined area of their own.

Placeholder for the aerial view of the yard after the kaizen, showing the new trailer-load area, consolidated cast iron, protected staging area, and one-way routing. — **The yard re-drawn from above.**·The redesign mapped onto the yard’s aerial footprint — a dedicated trailer-load area beside the high movers, the cast iron consolidated, a protected staging area, and one-way routing.·Aerial: house archive, May 2019.

Two of the charter’s other complaints turned out, on inspection, to be standards problems rather than the problems they had been filed as. The inventory “inaccuracy” was in large part a counting problem. When the team recounted three drainage-structure part numbers against the system’s figures, a count that had been done quickly showed a 26% error on one part — while the team’s careful count of the same part showed 2%. The system was closer to the truth than the yard’s own method of checking it. And the flood of manual purchase orders was a gap in how the pick-up items had been set up in the company’s ERP system: 11 of them were being hand-ordered simply because nobody had told the system to reorder them.

What went in during the week divided into what could be finished and what could be aimed. The purchase orders were the cleanest. Of the 11 pick-up items operators had been ordering by hand, nine were configured to reorder automatically in the ERP system; by Friday only two — dunnage and banding — still needed a hand-written PO, and the first automatic order for the truck knuckles had already gone out. The drycast labeling, filed as a fifth objective, came down to a broken stencil cutter: it was found and repaired, stencils for the drycast part numbers were cut and tested, and a run of parts was painted with its part number and cast date, so the yard could read at a glance what it was looking at. The layout itself was begun during the event — staging marked, the load area started — and carried out on a task list of 12 items, each with a named owner in the yard and a date, running out over the three weeks that followed. The yard’s area manager led the kaizen team himself; the task list was his team’s to close, not the practice’s.

What the week produced was less a finished number than a correctly aimed one. The yard had come in believing its problem was a slow will-call counter and a lying inventory system. It left knowing that its will-call was already fast three-quarters of the time, that its system was more honest than its counting, and that the thing actually slowing the customer down was the product it had cast and not shipped.

Enter-to-exit time (measured baseline): About 32 minutes on average; roughly 75% of customers already under the 45-minute target on a smooth day; the load itself about 14 minutes. Target of 90% under 45 minutes set as a standing KPI for the new layout to drive toward.
Manual purchase orders for pick-up items: 11 → 2 (dunnage and banding). Nine items moved to automatic reorder in the ERP system; the first auto-placed order went out during the event.
Inventory count accuracy: A careful recount cut a 26% error to 2% on the same part number. The count method, not the system, was the larger source of error.
Excess cast-iron inventory surfaced: Roughly $120,000 of stock beyond any system order, across seven line items — quantified and made visible.
Drycast labeling: Stencil cutter repaired; part-number-and-date stencils cut, tested, and piloted on a run of parts.
Yard layout: Redesigned around flow — dedicated trailer-load area, consolidated cast iron, high-movers forward, protected staging, one-way routing. Carried on a 12-item task list with named owners and dates.
Standing KPIs: Loads per day, as a measure of customer throughput; ten load times a week, as a measure of customer wait.

The overproduction was the story. A yard earns its keep by how quickly it can put a customer’s product on a customer’s truck, and this yard could not do that quickly because the ground the loading needed was occupied by product no one had ordered. Clearing it, and giving the high movers the ground nearest the trucks, did more for the customer’s wait than any change to the loading crew could have — the load time was never the constraint; the inventory in front of it was. The one thing the yard could not fix from inside its own fence was the tap that filled it. It asked, at the report-out, for someone upstream empowered to hear when overproduction was crowding the yard again. A yard can be cleared in a week. Keeping it clear is a question for the operation that fills it.

· · ·