VMI and Collaborative Supply Chain Management

The Business Situation
A nationally branded multi-million dollar manufacturer for the retail home improvement market was involved in a growing strategic relationship with a major home improvement center retailer. Both the manufacturer and retailer were interested in implementing some form of Vendor Managed Inventory (VMI) as the next step in the ongoing process to reduce costs and improve service levels. A supply chain analysis and best practices review of the current business processes was conducted as the first step in development of a comprehensive business technology and action plan for the implementation of VMI.

The Discoveries
The primary benefit of the analysis and review was the bringing together of both parties to "let the left hand know what the right hand was doing." This revealed several "truths" that both parties sensed but could not accurately articulate. Chief among the discoveries were:

Excess Inventory. The manufacturer's production was increasing while the retailer's store sales were declining and, at the same time, on-hand inventory levels were more than double the demand and growing. This resulted in a total system inventory of four to five times sales, which translated to five to six month's inventory across all product lines.

Lack of Shared Data. The excess inventory was, in part, the result of the manufacturer's lack of retailer sales visibility and retailer's lack of the manufacturer's inventory visibility. This lack of shared data explained the volatility of the overall planning.

Lack of Inventory Reliability. A contributor to the poor supply chain dynamics of the relationship was the inaccuracy and unreliability of the inventory at the store level. Errors and misinformation at this point were only made worse as they moved up and through the planning process.

Too Many Variables. Throughout the analysis process it became evident that the manufacturing and distribution process was rather complex, involving over 24 variables. Although each step in the process only used a selected few, collectively the entire process was affected by all of the variables, which compounded the planning complexity exponentially.

Insufficient Checks & Balances. The replenishment planning process was a one-way process with little emphasis on system driven checks and balances. Most checks and balances came from human judgment and experience and were typically made after the fact. Very little information flowed upstream to validate what was coming down stream in order to update inventory, productions, and other information databases.

The Recommendations
The discoveries led to the following recommendations for improvement.

Exception management. There were too many product combinations and delivery variables to manage all of the possibilities. Therefore, it was recommended that a small group of key indicators be selected for hands-on management leaving the majority to be processed by automated information systems. The trick was to define those attributes of product sales that were the best predictors - to pick the right ones and let the rest 'follow' in step.

Collaboration. In this fast changing world, both manufacturer and retailer have to work together to stay in business. The solution here was to use Collaborative Planning Forecasting and Replenishment (CPFR) as the model to bring the manufacturer and the retailer together in a collaborative environment.

CPFR is the recognized VICS standard for the retail industry and was therefore already 'acceptable' to both parties. Selecting a standard model also meant that the retailer did not have to have a different system for each vendor. Doing CPFR with selected key predictors also made it unnecessary to try and manage all 600,000 item/store combinations. Less was more.

Inventory reduction. All business processes live and die on the quality and availability of the information. Since inventory management is at the center of the manufacturer/retailer relationship, the accuracy of inventory information becomes essential to the success of the rest of the project. The means equal sharing of inventory data through a common database accessible by all interested parties. Here there was room for significant reduction in total system inventory by more than 60%. The solution, therefore, was to improve inventory/information accuracy and management of the entire system inventory through a collaborative process.

For this project, the sharing of information about safety stock was particularly important. Although acknowledged by both parties as major driver of the process, there was some confusion as the number of safety stocks, the exact calculation of safety stock, and which safety stock (the manufacturers or the retailers) was actually being used. Because of this, the safety stock calculations were independently made and resulted in what is know as safety stock compounding -- the number one contributor to over stocking and run-away inventory.

Checks & Balances. The success of collaboration and management by exception is based on the ability to identify the collaboration points and the right exceptions. Implementing the right check points and metrics in a dependable, repetitive, automatic environment can reduce human error to near zero and let mangers know when events, good and bad, occur as they occur. This is where information technology and computerized systems are at their best. Thus, the recommendation here was to create a series of collaborative checks that will provide alerts when the systems are out of balance, making sure that metrics are uniform across all operations.

Observations and Conclusions
The right analysis of the supply chain does more that just define and describe the current state (the what and the where), it also helps open the door to understanding the underlying philosophies, practices and procedures, and root concepts that make the current process function as it does (the why). This is important information that is needed before moving to corrective action phase where the road map to the future state is created (the who, when and how).

The identification of the root concepts can make finding the correct solution not only easier but much more effective. In this case study, it was important to understand how accurate forecasting was critical to the replenishment process and why unpredictable seasonal factors such as the weather had such considerable influence. This lead to further understanding of how the money ended up on the selves and in the warehouses when the systems get out-of-sync, and why management by exception was the best practice for correcting the situation.

The sales/inventory plots that were part of the analysis process were also helpful in focusing in on the key underlying forces that seem to drive the current operation - historical planning (lack of collaboration) and supply chain dynamics.

Historical Planning. Planning by the both the manufacturer and the retailer was based on historical data. On the retailer-side, order planning was based on POS, which was limited to what was already sold. On the manufacturer's side, orders from the retailer were only used to schedule deliveries. Most of the planning and production scheduling was driven primarily by historical data for previous production runs. The fulfillment process was like driving a truck using only the rear view mirror. It was anything but collaborative.

Supply Chain Dynamics. Process volatility (and therefore the primary area of concern) was most prevalent during the selling season when certain products were in high demand. Although only 5% of the SKUs (but 25-30% sales) were in this category, it was here where significant process swings took place that fed the dynamics of the supply chain that drove planning and inventory "out of control".

The Bottom Line
There are always several good opportunities for improvement in any supply chain operation, all reasonably identifiable but all not necessarily correctable without cooperation from all parties and/or a major change in the process architecture. Therein lies the challenge in going forward -- working in a collaborative environment and developing better ways of doing business. Simplification may be the first step, which is to say, don't try and manage everything at once, but stick to the exceptions. Use a smaller number of carefully selected predictors to manage the forecasting (where forecasting is a factor) and use automated processes (plan-o-grams and scanners) to manage the rest. Then collaborate to manage the exceptions. Doing so will make it easier for both parties and provide a win-win for everyone. David A. Hough, Precision Systems Concepts, Inc.
Newsletter Signup
Newsletter Signup
Get the latest industry news, insights, and analysis delivered to your inbox.
Join your peers