At a hypothetical Tier-2 stamping supplier in Aichi, the production planner assumed better demand visibility would require a new data-sharing agreement with the OEM. In practice, the first useful signal was already present in the assembly plant’s pull schedule, the Tier-1 replenishment pattern, and the supplier’s own ERP history. The harder task was organizing those signals by tier and connecting them to the parts that the line would actually consume.
This distinction matters for suppliers that cannot change their commercial relationships upstream. A Tier-2 or Tier-3 company may not receive a direct forecast from an OEM. It may receive a kanban call-off, a revised MRP requirement, or an urgent phone request from a Tier-1 plant. That does not mean the supply chain has no usable demand information. It means the information is distributed across operating systems, planning routines, and human workarounds.
Visibility can start with the pull schedule
An assembly plant creates a demand trace every day. Its pull schedule reflects planned production, model mix, takt time, line stoppages, changeovers, and sometimes the recovery plan after an andon event. When that schedule is connected to a bill of materials, the signal can be translated into requirements for Tier-1, Tier-2, and Tier-3 parts.
The translation is not simply a matter of multiplying vehicle volume by a standard quantity. A model changeover can alter the mix of brackets, stampings, fasteners, or molded components. A service part may have a different pattern from a line-fit part. Pack quantities, scrap assumptions, lot sizes, and shipment calendars also affect the quantity a supplier needs to prepare.
For a Tier-2 planner, the practical question is often narrower: which part numbers are likely to be pulled during the next one to three weeks, and how much uncertainty remains? A forecast that answers that question can be useful even if the OEM does not share a new portal, alter its supplier policy, or participate in a joint planning project.
Lean Path’s approach begins with the assembly plant pull schedule and works tier by tier. Its time-series model uses pull schedule history, seasonality, and model changeovers, then propagates the signal through the structure from OEM to Tier-1, Tier-2, and Tier-3. The aim is to give each supplier a view of the demand relevant to its own BOM and lead time, rather than asking every company to interpret the full vehicle schedule.
Collaboration does not have to mean a new relationship
Supplier collaboration is often discussed as if it requires a new meeting cadence, a shared planning portal, or a formal agreement among every company in the chain. Those arrangements can help, but they are not a prerequisite for improving the first forecast.
A smaller supplier can begin with the files and records it already handles:
- Assembly plant pull schedules or customer call-offs
- ERP order history, shipment history, and open purchase orders
- Part numbers and quantities in the relevant BOM
- Kanban quantities, container cycles, and shipment calendars
- Known lead times, minimum lots, safety stock, and supplier constraints
- Dates for model changeovers, shutdowns, and planned maintenance
The operating relationship remains familiar. The Tier-1 still sends the order. The Tier-2 still confirms capacity and ships against the agreed schedule. The production supervisor still uses the same escalation route when a press, die, or inspection process is unavailable. What changes is the preparation behind those actions: the planner can distinguish a recurring demand pattern from a one-off panic order before deciding whether to add a shift or expedite material.
This is also more compatible with kaizen than a large process redesign. A team can start with a small set of high-impact SKUs, compare forecast quantities with actual pull, and review the exceptions in its existing production meeting. The model becomes another input to the daily management process, not a replacement for the people who understand the press schedule, tool condition, and customer priorities.
What the assembly signal can reveal downstream
Consider a synthetic example. A Tier-2 stamping supplier in Aichi produces reinforcement brackets for a Tier-1 seat assembly plant. The supplier has an eight to fourteen day material and production cycle, carries safety stock for normal variation, and uses an ERP system that records shipments but not the final assembly schedule. Its planner sees irregular orders: several quiet days, followed by a large call-off and a request for partial delivery.
When the assembly pull schedule is mapped through the BOM, the pattern looks different. The high call-off is associated with a model changeover, while the preceding quiet period reflects a temporary mix shift rather than a sustained reduction in demand. The forecast can therefore flag the likely bracket requirement earlier, while also showing that the demand peak may not repeat at the same level the following week.
That information supports several ordinary decisions. The supplier can reserve coil material before the order arrives, check whether the die schedule is compatible with the forecast, and avoid treating every spike as a permanent increase in safety stock. If the expected pull falls, it can hold a conversation about capacity and shipment timing instead of reacting only after an urgent request appears.
The same structure can expose dependencies at Tier-3. A fastener supplier may not need the vehicle schedule itself, but it may need a credible view of the bracket or seat assembly demand that drives its own shipments. Signal propagation makes the relationship visible without requiring each tier to exchange every internal planning detail.
Lean Path’s typical first forecast is available in the third week of onboarding. A pilot is structured as an eight-week engagement covering up to 15 SKUs. In design-partner pilot data, the observed result was approximately 35% fewer unplanned reorders. That figure is specific to the pilot context and should be treated as a qualified operating measure, not as a guarantee for every plant or part family.
Where the method has limits
Demand visibility cannot correct every supply problem. If the pull schedule is frequently overridden without a reliable record, the model may see noise rather than a stable demand signal. A new vehicle launch with limited history can make seasonality and changeover effects difficult to estimate. A sudden quality hold, die failure, labor shortage, or logistics interruption can also invalidate a forecast that was reasonable the day before.
There are structural limits as well. A BOM may be incomplete, outdated, or maintained differently in separate ERP systems. Part substitutions can obscure the relationship between the assembly schedule and the actual component consumed. A supplier with a long capacity constraint may need a commercial decision, tooling investment, or APQP intervention rather than a better prediction.
The boundary is important: this approach does not claim to replace MRP, customer instructions, kanban rules, PFMEA controls, or the judgment of a production planner. It also does not claim that a Tier-2 supplier can infer confidential OEM decisions that are absent from its available data. Its purpose is narrower, to organize existing demand evidence so that suppliers can prepare earlier and respond with fewer avoidable surprises.
Human review remains necessary. A planner should be able to ask why the forecast changed, check whether a model changeover was entered correctly, and record an operational reason when actual demand departs from the expected pattern. In automotive production, traceability and escalation matter as much as statistical accuracy. The forecast should sit inside that discipline.
A practical starting point for supplier teams
A useful first step is to map the structure before choosing a model or setting a target. Identify the assembly plants that generate the relevant pull schedules, the Tier-1 customers that place orders, and the Tier-2 and Tier-3 parts that sit behind each assembly. Then document lead times, BOM relationships, pack quantities, shipment days, safety stock policies, and known points where manual intervention enters the process.
Next, select a limited SKU group. Parts with repeated demand, clear customer ownership, and a meaningful cost of expedite are usually easier to review than one-off components. Compare the forecast with actual pull by week, not only against the order placed after the fact. Track unplanned reorders, late changes, excess stock, and the operational reasons for misses.
This structure mapping also creates a shared language for collaboration. A Tier-1 can discuss the signal it receives from the assembly plant. A Tier-2 can show how that signal reaches its press schedule and material plan. A Tier-3 can identify where the lead time or lot size creates a delay. None of this requires the participants to change their contracts or establish a new joint organization.
Lean Path is a bootstrapped team working directly with Japanese automotive suppliers. Our company description is simple: “Lean Path keeps automotive supply chains stocked without overbuying. Our AI forecasts demand across each tier of parts suppliers, so a plant orders what the line will need, not what it panicked over last week.” If your supply chain has a difficult handoff between the assembly schedule and the parts that support it, contact [email protected] to discuss the specific structure, including where the Tier-1, Tier-2, and Tier-3 relationships begin and where the current signal becomes unclear.