The parcel peak should not set the factory's pace

Every autumn China's parcel network prepares for the largest shopping events of its year. Sorting hubs add temporary staff, line-haul carriers book extra trucks, and the franchisees who deliver the last kilometre buy vans. I spent five Novembers on the carrier's side of that preparation, at Kuanhe Express in Hangzhou, where daily volume at the peak reached 58 million parcels. I remember waiting anxiously for new vans in September, and I remember telling manufacturers that if they could not deliver by the end of the month we would buy from someone who could.

Now I am the manufacturer, and I have come to believe that the parcel peak should not set a factory's pace. The carriers are right to want their vans on time. But the sprint we used to run to meet them was expensive in ways our monthly reports did not show, and there is a better way to give them what they need.

The hidden bill for building in a rush

When I joined Tanzhou in 2021, our Changsha plant built 42 per cent of its annual output in the third quarter. It worked six-day weeks from July to September, and dispatched workers supplied by agencies on short assignments reached 2,300 at the peak, many of them new to vehicle assembly. The plant met its delivery dates, and the third quarter always looked like our best.

The bill arrived later, in other departments' budgets. First-pass yield, the share of vans passing end-of-line inspection without rework, fell from 93 per cent in spring to 84 per cent in August and September. Lost-time injuries in the third quarter ran at more than twice the rate of other quarters. Vans built in the autumn generated 1.6 times the warranty claims of spring-built vans in their first ninety days, and those claims arrived during the peak itself, when a broken-down van costs a franchisee most. Overtime premiums ran to more than RMB 70 million a year. None of these costs was charged to the decision to sprint. Each appeared as a separate problem with a separate owner.

I knew the pattern from the carrier's side. At Kuanhe the vans that failed in the first weeks of the peak were disproportionately the newest. At the time I blamed the manufacturers' quality. I now understand that their quality was partly a product of our own ordering habits.

A yard of finished vans is not waste

The usual objection to a level build plan is that it means holding inventory. Lean manufacturing, as many of us were taught it, treats finished stock as waste: it ties up cash, occupies land and can lose value. All of that is true. But the comparison that matters is not stock against no stock. It is stock against the full cost of the sprint, and in our case the stock was cheaper.

From 2023 we built at a constant daily rate from February to November and parked the vans built ahead in a 9,000-place yard at Changsha. Carrying them costs about RMB 38 million a year in financing, space and handling. Overtime premiums fell by RMB 74 million, dispatched workers at the peak fell to 640, first-pass yield averaged 94.1 per cent across the year, lost-time injuries fell by 58 per cent and early-life warranty claims by 31 per cent.

Holding stock well is a skill, and we learnt it expensively. In our first winter, 140 stored vans suffered deep discharge because their systems drew a small current while parked and we had not scheduled maintenance charging. Twenty-two needed new battery packs. The yard now has its own charging rota, monthly inspections and stock controls that our auditors examine like any other inventory. An operations team that treats a yard as a car park will lose money in it.

The yard also changed how the plant thinks about its year. With a steady daily rate, supervisors can plan maintenance, training and line improvements in every month rather than cramming them into the winter lull. Team leaders keep the same crew for most of the year instead of supervising strangers in August. Our suppliers benefit as well: their schedules flatten with ours, and the two that used to add weekend shifts for us every autumn no longer need to. A level build plan travels upstream, and some of its savings appear in parts prices that no longer carry a peak premium.

Asking customers to commit in February

A level plan works only if customers commit earlier than they are used to. We offer large fleet buyers early allocation agreements, signed in the first quarter, which reserve their vans and schedule deliveries to their depots in August and September. Title passes on delivery. If our list price falls before then, the customer receives the difference. Customers may also move up to 15 per cent of their committed volume between models until the end of May, because a carrier's fleet mix changes as it learns what the coming peak will look like.

That price protection is our main exposure. In 2024, during the price war, it cost RMB 52 million on the models where we did cut prices. Our chief financial officer warned about exactly this when we began, and he was right to. I still prefer it to the alternative, because it lets a customer commit without betting against us, and because it makes us think harder before cutting a price.

Not every customer accepts. Smaller carriers and franchisees often cannot forecast their needs in February, and we gave up about 3,400 orders in 2023 from buyers who would not commit. Some came back the following year, having been let down by a manufacturer that promised September and delivered in November.

My advice to other operating chiefs who serve seasonal fleets is to add up the costs of the sprint across every budget they land in: overtime, rework, injuries, early warranty claims and the damage done to customers' own peak. Set the total against the cost of holding stock carefully. If, as in our case, the stock is cheaper, the factory should run at the pace of the year and let the yard absorb the season. The carriers do not need us to sprint. They need their vans to arrive, work and keep working through the busiest weeks of their year.