Demand Pull Programs (VMI) (White Paper)
Background: The Drive for VMI
In most manufacturing environments, components and subassemblies are ordered through a purchase order, acknowledged by the supplier, delivered to the manufacturing entity, invoiced, and paid. Sometimes, the manufacturer provides a forecast to its suppliers in order to provide visibility to future demand. These forecasts are used for general planning purposes and unless defined otherwise, are not commitments to purchase. Typically, forecasts are sent through an electronic link to the supplier such as EDI, a web-based model, or via spreadsheets. When changes to demand occur such that the requirements stated on the purchase order need to be modified, the manufacturer requests a pull-in, push-out, cancellation, or modifies the quantity. In most manufacturing entities (or at the central purchasing organization where orders are sometimes managed), an MRP/ERP system runs on a weekly basis, triggering orders to be placed or modified as described.
Problems with Traditional PO Approach:
The approach of placing hard copy purchase orders has several problems as described below for both the buyer and the component supplier:
- Visibility: The component supplier has little meaningful visibility on which it can plan its production, particularly when the component is technologically complex and has a long manufacturing lead-time. When orders are pushed out, pulled in, or cancelled, the component supplier does not fully understand what is happening inside the buyer’s entity and when the next order will be placed and for what quantity. It does not know if different entities of the buyer are transferring inventory from one site to another, if there is excess inventory at the procuring site itself, or if true demand is following forecasted demand. Everything outside of real purchase orders or contractually committed forecasts is an estimate at best. In other words, the component supplier is in the dark beyond the PO and/or the forecast. The only way it can sort through these issues is by spending inordinate amounts of time analyzing every forecast change and every change request. Given the limited resources that are available, it is nearly impossible to analyze every situation.
- Execution: The employees of the buyer or the component supplier are not perfect. With hundreds, perhaps thousands, of transactions every week, there is a good chance that something will go wrong: paperwork is lost, misrouted or never sent and people make input errors on both sides. In addition, processing delays can mean precious days are lost during a time of change. Betting your success on a manual process is a formula for failure in a fast-paced business environment.
- Accountability: When there is excess inventory, it is not always clear who is responsible for the excess. For example, who is accountable when an order is lost? Who is accountable when demand doubles or drops to zero in one run of the MRP, while the supplier is responding to a month-old forecast? OEMs, contract manufacturers, and component suppliers have different perspectives to offer on these questions. In addition to these administrative problems, these stakeholders have different opinions on how much of the excess exposure belongs to them. In the event of a cancellation, OEMs normally believe their liability should assume that the CM and component supplier acted on their change requests immediately, and that their exposure is limited to the cost that the component supplier has incurred in the manufacturing process until the day the cancellation request was issued. Component suppliers, however, have more strict cancellation clauses that do not reflect the same assumptions made by the OEMs, particularly when they have strict limitations placed on them by the foundries. The same reasoning can be applied to unexpected spikes in demand that often lead to shortages.
The net result of these three process gaps is that action is delayed and errors are made – the net impact is excess inventory and/or supply shortages.
An Alternative Approach: Demand Pull Programs or Vendor-Managed-Inventory (VMI):
Vendor-managed-inventory is an inventory management and supply flexibility model that leaves inventory in the possession of the component supplier until there is a need to use that component in a production build. Only at that time does the inventory transfer on to the books of the buyer and trigger payment to the supplier. There is usually a clause that requires enough supply to meet an immediate spike in demand (typically 25-30%). At the same time, a separate clause is used to push-out deliveries while requiring the buyer to accept deliveries after the freshness period has ended. Again, requiring that the component supplier “hold on” to inventory on its books until a “pull” signal is generated by the customer does not necessarily mean that the component supplier is fully responsible for the inventory that it has reserved for use by the buyer. The program puts the entire process of pull-ins, push-outs, and cancellations on auto-pilot. A “liability” clause is structured to reflect the order lead-time, simulating what would have been a PO in a traditional environment.
This approach addresses much of the fundamental problems above:
- Better Visibility: Up-to-date visibility is achieved by providing a real-time view into the buyer’s demand and inventory levels. In an outsourcing environment, information starts at the OEM, flows to the contract manufacturer, and to the component supplier. Since the information flows electronically, it can be disseminated more quickly to the component supplier’s own production line so that adjustments can be made quicker.
- Eliminates Manual Execution: In VMI, since there are no hard copy POs and acknowledgments, the chance for human error is drastically reduced. Human intervention is only needed to analyze the information from time to time and ensure that the system is working properly.
- More Defined Accountabilities: In a VMI model, accountabilities are defined upfront through a contract. This contract clearly articulates how information is to be sent, which commitments are “soft” and which ones are “hard,” how escalations and exceptions are to be addressed, and most importantly, who is responsible for what portion of the forecast. All stakeholders will have visibility to the same information and the contract terms and conditions will establish how information will be used to trigger action.
If executed properly and with mutual agreement, quicker action can lead to reduced inventory and more supply flexibility for all stakeholders.
Common VMI Complaints:
Understandably, component suppliers have generally been the least receptive stakeholder in the supply chain to accept the VMI model. Naturally, pushing the inventory burden on them and subjecting their revenue plan to uncertainty until late in the quarter is of concern, especially if they have no control of their own production environment (e.g. fab-less semiconductor companies). If poorly executed, VMI can actually cause a great deal of strain for all stakeholders throughout the supply chain and end up damaging relationships, resulting in a lose/lose or win/lose situation.
When it comes to VMI, component suppliers' most common complaints are as follows:
- Significant Revenue Risk: Since VMI is a demand pull program, the trigger for invoicing comes only after the customer requests a part out of inventory. This means that quarterly revenue projections are fluid and at the full mercy of the buyer. For small, specialty fab-less semiconductor companies, this can have dire consequences on their business.
- Significant Inventory Risk: When revenue is unexpectedly pushed out, the component supply in question turns into excess inventory on the balance sheet.
- Push-out Clause Execution: When the push-out clause kicks in and requires the buyer to take possession of the inventory, the burden is still on the buyer to actually “pull” the material. If the buyer does not execute as agreed (which is typically the case), this clause becomes meaningless and can leave inventory on the books on the component supplier for several months.
- Lack of Monitoring: There are two information related problems that sometimes arise in the migration to an automated replenishment program like VMI. One problem is the technical glitches that might occur in the submittal of information, which can go unnoticed for weeks. This happens due to the assumption that since VMI is an automated replenishment program, it must run on its own and without monitoring. When these glitches are not caught in time, component suppliers end up absorbing an unfair share of the impact. Another problem is that the human intelligence that normally goes into reviewing the information and creating, rescheduling, or canceling a PO is lost when data is submitted electronically. Unless a consistent and institutionalized review process is in place on both sides to ensure that the information being transmitted is completely accurate, the risks of VMI can outweigh its benefits.
All of the points above can be negotiated and addressed with a win/win mindset and without placing excessive burden on one stakeholder. A prerequisite for a successful VMI program is that there is alignment on the organizational goals of the component supplier, the EMS company, and most importantly, the OEM responsible for originating the forecast. In fact, one of the major pitfalls of VMI is that they are developed and negotiated with minimal involvement and support from the OEM, causing the program to fail only a few months after implementation. Also, it is important to realize that all parts are not the same. A component supplier’s ability to offer flexibility on part varies depending on whether the part is custom, semi-custom (e.g. semiconductor ASSPs), or standard. This has a strong impact on defining what is win/win and what is not.
Summary:
Implementing VMI can lead to success or failure, depending on how it is implemented. There is no doubt that the paradigm shift associated with migrating away from a traditional PO based environment to VMI will have noticeable impact on all stakeholders. Whether this impact is positive or negative depends on how well it is understood by the stakeholders, the robustness of the contract that is negotiated, the mechanics for information exchange, and most importantly, its financial implications for everyone involved.
Symphony's Value:
Our consultants are intimately familiar with the VMI program, its risks and rewards. We can demonstrate, via case studies and quantitative modeling, how the risks stated above can be mitigated to allow a successful implementation. Symphony can support its clients’ implementation efforts through the following 8-step process:
- Assess impact of VMI program on company financials, i.e. revenue and inventory. This will ensure that prior to launching the program, the financial impact is understood upfront, defining what should and shouldn’t be within the scope of VMI.
- Identify scope of program. At this point, the company determines which OEMs, contract manufacturers, or grouping of products fit this program. It helps the company review where a product may be in its life cycle, possibility of transition from one contract manufacturing location to another, etc.
- Validate with OEM and obtain support. This step is very critical and often overlooked. It is imperative that the OEM that originates the forecast knows and buys into how the model of inventory liability and supply flexibility that impacts its products are being modified.
- Develop a win/win VMI business model, including the negotiation of mutually acceptable terms and conditions that are executable. In this model, critical items such as inventory protection and inventory liability will be determined.
- Define the mechanics of information and material flow. Contrary to what many may believe, this program cannot be placed on autopilot and left to its own device, particularly since it can have such a large impact on these small companies. Data needs to be reviewed, validated, and signed off on an on-going basis to ensure that human intelligence is applied to the information that is being used to build expensive products.
- Test the mechanics through a pilot program and work out bugs. Implementation of a new program never goes 100% as planned. A beta test needs to take place and the system needs to be debugged before going “live.”
- Implement program and expand as appropriate. Internal and external business processes need to be established as part of the implementation to handle issues such as product obsolescence, migration from one manufacturing plant to another, invoicing and payment, along with others.
- Review program and adjust to meet the changing business environment. The success of this program needs to be reviewed and modified as appropriate. The business climate is dynamic and what is optimal today will not be optimal tomorrow.
