Vendor-Managed Inventory (VMI): How It Works and When to Use It

What VMI actually is

In a traditional supply relationship, the buyer watches their own inventory, generates a purchase order when stock falls below a reorder point, and sends that order to the supplier. The supplier fulfills it. Each step introduces delay and information distortion.

VMI inverts this. The supplier is given visibility into the buyer’s inventory — sometimes down to the shelf level at a retail store, sometimes at the distribution center. The supplier then decides when to replenish, what quantity to send, and how to route the shipment, all within parameters the buyer has approved (minimum service level, maximum inventory cap, agreed lead time). The buyer receives goods without placing an order.

Walmart and Procter & Gamble piloted this arrangement in the late 1980s and documented dramatic results: P&G’s in-stock rate at Walmart stores improved and both companies reduced inventory. The model spread quickly through consumer goods and has since moved into manufacturing, healthcare, and industrial distribution.

How the process works

A typical VMI cycle has four steps.

First, the buyer shares data. This is usually point-of-sale data from retail scanners, inventory-on-hand at the warehouse or point of use, or consumption data from a production line. The sharing mechanism is typically EDI (electronic data interchange) or a shared web portal. Some programs use IoT sensors on bin levels or weight-based shelf monitors.

Second, the supplier analyzes the data. Their replenishment system — or a human planner — reviews current stock against agreed minimums and projected consumption. They decide whether to ship, when, and in what quantity.

Third, the supplier ships. Often without a purchase order from the buyer. The advance shipping notice (ASN) serves as the transaction record. The buyer’s system receives the ASN and reconciles it against the delivery.

Fourth, the supplier invoices. The billing trigger varies: some programs invoice on shipment, some on consumption (consignment VMI), and some on a fixed schedule. Consignment VMI means the supplier retains ownership of the goods until they are used, which changes the balance sheet treatment for the buyer.

VMI vs traditional replenishment

The difference is not just administrative. It changes who holds information risk and who absorbs forecast error.

Under traditional replenishment, the buyer forecasts their own demand, sets a reorder point, and places orders that reflect that forecast plus a safety margin. If the forecast is wrong, the buyer either runs short or sits on excess inventory. The supplier sees only the order, not the underlying demand — a filtered and delayed signal that drives the bullwhip effect.

Under VMI, the supplier sees actual consumption or point-of-sale data. Their replenishment decisions are based on real demand, not on a buyer’s order that already reflects the buyer’s safety stock logic. The signal is cleaner. The supplier can plan production and logistics more smoothly because they are not reacting to lumpy purchase orders.

The trade-off: the supplier now carries more planning responsibility and may need to hold more finished goods inventory to fulfill VMI commitments on short notice. This needs to be reflected in the commercial terms.

Benefits for the buyer

When VMI is working well, buyers see several concrete gains.

Reduced safety stock. The supplier is reacting to real consumption data with a shorter effective reaction time than a buyer who orders weekly. This tightens the replenishment cycle and reduces the buffer needed to cover uncertainty. Use the Safety Stock Calculator to quantify the reduction — the standard deviation of lead time drops when the supplier controls replenishment timing.

Lower ordering cost. Buyers in VMI programs often eliminate the purchase order entirely for covered items. The procurement team’s time shifts to managing the VMI contract and exception handling rather than routine ordering.

Improved service levels. Suppliers with visibility to real demand can position inventory proactively before a seasonal peak or promotion rather than reacting after the buyer’s order arrives late. In-stock rates typically improve.

Reduced bullwhip contribution. Because the buyer is no longer placing lumpy, delayed orders based on forecasts, the demand signal reaching the supplier is smoother and more accurate.

Benefits for the supplier

VMI is not a one-sided arrangement. Suppliers gain as well.

Production planning stability. Real consumption data is more accurate than order data, which reflects the buyer’s forecast plus safety stock logic. Suppliers can level their production schedules instead of reacting to volatile order patterns.

Stronger customer relationship. A supplier with deep data access and replenishment authority has a higher switching cost for the buyer. This is a competitive moat, though it requires the supplier to perform or risk losing the program.

Potential for larger share of wallet. VMI often expands to additional product lines once trust is established. The supplier’s visibility into the buyer’s operations creates natural opportunities to identify unmet needs.

When VMI makes sense

VMI works best when several conditions are present at once.

High transaction volume with low unit value. The overhead of generating purchase orders for thousands of low-cost items — fasteners, consumables, MRO supplies — is disproportionate to the value. VMI eliminates that overhead.

Stable, predictable demand. The supplier needs enough demand signal stability to make good replenishment decisions. Items with extreme seasonality or high demand variability require sophisticated forecasting on the supplier’s side that many suppliers cannot support.

Trusted supplier relationship. The supplier needs to have access to sensitive inventory and consumption data. This requires a relationship where the buyer is confident the supplier will use that data appropriately and not, for example, share it with competitors.

Supplier capacity to manage it. Small suppliers sometimes lack the IT infrastructure or planning capability to run VMI effectively. The program requires the supplier to invest in systems or staffing. If they cannot, the arrangement degrades into the buyer doing most of the work anyway.

When VMI does not work

Several conditions reliably cause VMI programs to underperform or fail.

Poor data quality. If the inventory data the supplier receives is inaccurate — because of receiving errors, unreported shrinkage, or warehouse management system gaps — the supplier’s replenishment decisions will be wrong. Garbage in, garbage out. Audit your inventory accuracy before starting a VMI program, not after.

Misaligned incentives. If the supplier is measured only on fill rate, they will over-stock the buyer to protect that metric. Carrying cost is the buyer’s problem. The contract needs to include inventory caps and a shared measure of inventory turns to align incentives. Check Inventory Turnover as a shared KPI.

Too many SKUs, too early. Some programs try to cover an entire product catalog on day one. The complexity overwhelms the supplier’s planning team. Start with a pilot of 20–30 high-volume items and expand once the process is stable.

No governance process. VMI agreements need regular reviews — weekly or monthly dashboards, a clear escalation path when service levels fall, and a defined process for adding or removing SKUs. Without governance, the program drifts until a stockout or inventory pile-up triggers a crisis review.

Key metrics to track

A VMI program needs a small set of metrics tracked consistently, not a sprawling dashboard.

Service level (fill rate). Are orders being fulfilled at the agreed rate? A fill rate below target is the primary alert that the supplier’s replenishment decisions are failing.

Days of supply on hand. Is inventory at the buyer’s location staying within the agreed range? Chronic high days of supply means the supplier is over-stocking; chronic low days of supply means the supplier is cutting it too close.

Forecast accuracy. Track how well the supplier’s replenishment quantities match actual consumption. The Forecast Accuracy Calculator gives MAPE and bias — bias is particularly important in VMI because a consistently low bias (under-shipping) drives stockouts while high bias (over-shipping) drives excess.

Inventory turns. Compare the turns on VMI-covered items against non-VMI items in the same category as a sanity check that the program is actually reducing inventory and not just shifting who places the order.

Implementation steps

1. Audit inventory accuracy first. VMI built on bad data fails. Cycle count VMI candidates and get accuracy above 95% before starting.
2. Define the data-sharing mechanism. EDI 852 (product activity data) and EDI 855 (purchase order acknowledgment) are the standard formats. Agree on frequency — daily point-of-sale is ideal; weekly is the minimum for most programs.
3. Set the inventory parameters in the contract. Minimum stock trigger, maximum inventory cap, service level target, and what happens if the supplier misses the service level. These are the guardrails the supplier operates within.
4. Pilot on a small SKU set. 20–30 items, one supplier, one location. Run for 90 days and measure against baseline before expanding.
5. Build a governance cadence. Weekly operational review (fill rate, days of supply), monthly strategic review (turns, contract performance, expansion candidates).

Common mistakes

• Treating VMI as a technology project. VMI is a relationship and process change that technology enables. Companies that buy VMI software without fixing the underlying data quality or commercial alignment fail regardless of the platform.
• Not establishing baseline metrics before starting. Without a pre-VMI baseline for service level, days of supply, and turns on the covered items, you cannot demonstrate the program’s value or identify where it is failing.
• Skipping the ABC segmentation step. Not all items are good VMI candidates. Use the ABC Analysis Calculator to identify the high-volume A items where VMI overhead savings are largest, and leave C items out of the initial program.
• Letting the contract become stale. The inventory parameters agreed at program launch (min/max levels, service targets) need to be updated as demand patterns change. A contract written for normal demand will be wrong during a seasonal peak or a product launch.

Frequently asked questions

¿Who owns the inventory in a VMI program?

It depends on the commercial model. In standard VMI, the buyer owns the goods on receipt and the supplier invoices on shipment. In consignment VMI, the supplier retains ownership until the buyer consumes the goods, which keeps the inventory off the buyer’s balance sheet. Consignment VMI requires more legal and accounting work to set up but gives the buyer a working capital advantage.

¿What is the difference between VMI and consignment stock?

VMI is about who controls the replenishment decision — the supplier does. Consignment is about who owns the inventory — the supplier does until it is consumed. These can be combined (consignment VMI) or separated. A supplier can manage replenishment under VMI while the buyer takes ownership on receipt, or a supplier can hold consignment stock without taking on VMI planning responsibility.

¿How does VMI reduce the bullwhip effect?

The bullwhip effect is driven by each supply chain level ordering based on its own forecast plus a safety buffer, amplifying demand variability as the signal moves upstream. VMI bypasses the buyer’s ordering layer — the supplier sees actual consumption data, not the buyer’s order. This removes one layer of demand signal distortion and the associated safety buffer inflation. The supplier’s replenishment decisions are anchored to real demand rather than to a filtered, delayed purchase order.

¿Is VMI only for retail?

No. VMI originated in retail (Walmart-P&G is the classic case) but is widely used in manufacturing (supplier-managed bins of components at the point of use), healthcare (supplier-managed hospital supply rooms), and industrial distribution (MRO consumables managed at a plant). The mechanics differ, but the core principle — supplier sees real consumption and manages replenishment within agreed parameters — applies in any context where the buyer has high transaction volume and the supplier has the capacity to manage it.

¿What data does the supplier need for VMI to work?

At a minimum: current inventory on hand at the buyer’s location, recent consumption or sales data, and agreed min/max inventory levels. More sophisticated programs also share demand forecasts, promotion calendars, and seasonal plans. The supplier also needs to know the buyer’s lead time expectations and any constraints on delivery frequency. The quality and timeliness of this data is the single biggest determinant of VMI program success.

¿How do you measure whether a VMI program is working?

Compare pre- and post-VMI performance on four metrics: service level (fill rate), days of supply on hand, inventory turns, and ordering cost. A working program should show equal or higher fill rates with lower days of supply and fewer purchase orders processed. If turns are flat or lower, the supplier is over-stocking to protect fill rate — a common early failure mode that the contract’s inventory cap should prevent.