Operations · Supply Chain

Just-in-time on paper, push in practice

Mohit Saini · 1 May 2024 · 4 min read

Just-in-time on paper, push in practice

Walk through most automotive plants today, whether a vehicle maker or a component supplier, and you will see the symbols of just-in-time manufacturing everywhere: kanban cards, marked bins, limited storage space, clean floors, tidy cells. Incoming stores often hold only a few shifts’ worth of material.

It looks like a supply chain that runs on pull. But appearances can mislead.

Two tests of a stable supply chain

There are two simple ways to judge whether a supply chain really flows:

  1. How much expediting goes on?
  2. How stable is inventory across the chain?

Expediting. Auto plants are assembly operations: each product needs many parts to arrive together. A good indicator of stability is kit coverage: for how many days of planned assembly are complete sets of parts available? In many plants, the answer is less than a day for exactly the items customers want most urgently. Purchase managers spend the whole day chasing parts to complete today’s kits. Production schedules change within a day, sometimes within a shift.

Inventory. Dealers’ stocks balloon in every downturn, and vehicle makers take months to adjust production, sending a bullwhip through the chain. Inventory drops sharply at every month end with a rush of dispatches. Component suppliers keep a month’s stock in warehouses a few hours from the assembly plant, yet the vehicle maker still can’t find the part it needs today.

A true just-in-time chain wouldn’t show either symptom.

Why bins alone don’t make pull

Limiting space or bins works when demand for each item is relatively smooth. Toyota achieved that through standardisation and by smoothing demand for each item over time.

Most vehicle makers can’t impose such smoothing on their markets. Demand for individual models and variants swings widely. On top of that, frequent changes in emission norms and models mean components are replaced often, so the number of bins for each part keeps having to be redefined. That makes the classic kanban improvement cycle (gradually removing bins to expose problems) almost impossible.

So a compromise has taken hold:

  • Bins move material between the supplier’s warehouse and the vehicle maker’s line.
  • But production at the supplier is still triggered by the vehicle maker’s monthly forecast, not by the bins.
  • Dealers are still supplied according to sales targets, not according to what they actually sold.

The result is a push system with a few visible pull features. The vehicle maker keeps very little stock and asks the weaker partners (its suppliers) to hold enough near its plant to absorb every swing. Suppliers, planning to monthly forecasts, can’t respond to daily changes, so they still miss parts on the day while holding large stocks overall. Judged against the original monthly plan, they score themselves highly; judged on daily needs, they fail.

Stock in the wrong form

There is a structural oddity in many of these chains: finished-goods stock is many times higher than component stock. Yet components are where flexibility lies. The same component can go into many end products. When components run short, a part needed today may already have been built into a product that won’t ship for weeks.

Reversing this profile, with more stock held as components and less as finished goods, gives far better availability for the same money.

Picture a supplier that makes eight variants of a brake assembly, all using the same housing. It holds two weeks of each finished variant, but only two days of housings. When the vehicle maker suddenly calls for double the usual quantity of variant C, the supplier has plenty of A, B and D on the shelf, and no housings left to make more C. Hold a week of housings and a week of each variant instead, roughly the same money, and the same call is met in a day: the housings go where demand actually turned up.

Where the change should start

The best starting point is the vehicle maker’s demand side: supply dealers according to consumption, hold more stock centrally, and produce to refill the central stock. That smooths the load on the plant and on every supplier tier.

But a component supplier doesn’t have to wait for that. Even while the customer sends monthly schedules, the supplier can switch its own production from push to pull:

  1. Produce to actual off-take, not to the monthly schedule. Hold a buffer of each finished part sized to the maximum consumption during its replenishment time. Daily consumption from the buffer triggers production. The buffer level is a lock: nobody produces above it.
  2. Make the plant fast by securing components. Most component plants can make a part quickly; what stretches their lead time is waiting for bought-out components from tier-2 suppliers. Give purchase stores enough space to hold buffers of these components, and trigger purchases by daily consumption, not forecasts. Component stock may rise at first, but finished-goods stock falls by much more.
  3. Prioritise by colour. For each finished part, compare stock with its buffer: below a third is red, a third to two thirds is yellow, above two thirds is green. Produce reds first. In a healthy system, most parts stay in yellow. The same colours tell you when to raise or lower a buffer.
  4. Help tier-2 suppliers shorten their lead times, which in turn lowers the stock needed at tier-1.

Suppliers who work this way often find their inventory turns several times faster, with fewer missed deliveries. It does require a partnership approach with sub-suppliers, rather than constant renegotiation on price.

Clean floors and kanban cards are the visible part of just-in-time. The invisible part, producing only what is consumed, is what actually makes it work.

Related reading

Mohit Saini

Strategy, supply chain & sales transformation consultant. About · Book a discussion