Vendor Managed Inventory Suppliers: How to Evaluate

Vendor managed inventory only works if the supplier on the other side can actually run it. Plenty of suppliers will say yes to VMI because it locks in volume, then discover they cannot read your stock file, cannot forecast, or have no one to visit and count. This guide is a selection checklist: the capabilities to test, the categories that suit the model, how to structure the commercial deal, what to ask in an RFI, how to design a pilot, and the KPIs that belong in the agreement.

Why supplier choice decides the outcome

Most failed VMI arrangements are not failures of design. The levels were sensible, the agreement was readable and the internal case stacked up. What went wrong is that the chosen supplier could not sustain the work. Someone had to pull a stock report manually every Monday, the replenishment decision drifted to whoever happened to be free, and within two quarters the arrangement had quietly reverted to the buyer chasing the supplier. If you want the mechanics of the model itself, our vendor managed inventory guide covers how VMI works and how it differs from consignment. This article assumes you have decided VMI is right and now need to choose who runs it.

The shift in thinking is that you are no longer buying a product with a service attached. You are outsourcing a planning function, and you should assess candidates the way you would assess any outsourcing provider: on process, systems, people and financial strength. A supplier can be excellent at making and shipping the item and still be a poor choice to plan your consumption of it.

The six capabilities to test

These are the areas where VMI arrangements actually break. Test each one with evidence rather than a statement of intent, because every supplier will claim all six.

• Systems and data integration. Can they consume a stock and consumption file on a schedule, in a format you can produce, without a six-month project? Ask what their other VMI customers send them. A supplier who only accepts a proprietary portal login, or who wants a spreadsheet emailed by a human, has a fragile process. Reliable exchange, whether by API, scheduled file or EDI, is the foundation everything else sits on.
• Forecasting competence. Ask how they turn your consumption history into a replenishment plan. You are listening for a method, not software names: how they handle seasonality, what they do with a single large order that distorts the average, how far ahead they plan. A supplier who says they will simply top up to maximum every week is not forecasting, they are reacting.
• Replenishment discipline. Who reviews the numbers, how often, and what happens when the review flags a problem? The strongest answer names a role, a frequency and an escalation route. The weakest says the system handles it.
• Local stock-holding or hub capability. VMI compresses your buffer, so the supplier needs one of their own within reach. Establish where the stock physically sits, what the transit time is from there, and whether that location is dedicated or shared with every other customer in the region.
• Field service. For anything held in bins, stores or on a shop floor, someone has to visit, count and correct. Ask who that person is, how often they attend, and whether the visit is a contractual commitment or a favour that disappears when their diary fills.
• Financial capacity. Carrying stock ties up the supplier's working capital. A supplier already stretched will either refuse consignment, hold less than agreed, or push the cost into your unit price. Check the accounts before you get to the commercial conversation, not after.

A capability scorecard you can use

Score each candidate out of five against the criteria below, weight them for your category, and keep the completed sheets. The scorecard forces the conversation past relationship history, and it gives you a defensible record if the choice is questioned later.

Which categories and suppliers suit the model

Category fit and supplier fit are separate tests, and both have to pass. The category needs steady, predictable consumption, a reasonable number of line items, low enough unit value that a small overstock is not painful, and a shelf life long enough to absorb a slow month. Fasteners, packaging, cleaning consumables, personal protective equipment, maintenance spares and basic raw materials all qualify. Anything highly seasonal, customised, perishable or subject to engineering change usually does not, because the supplier cannot forecast it better than you can and the whole rationale disappears.

On the supplier side, three profiles tend to work. Manufacturers with genuine stock-holding and a distribution arm can run VMI on their own range well. Specialist distributors are often the strongest candidates, because consolidating many lines is their core competence and they already hold regional stock. Large national wholesalers can do it, though the risk is that you become a small account inside an automated process nobody looks at. Our inventory vendor guide goes deeper into how these supplier types differ in practice.

Structuring the deal so the cost does not come back

The commercial risk in VMI is simple. You hand the supplier the carrying cost, they add it to the unit price, and you end up paying for the inventory you thought you had removed, with less control than before. Guard against this in three ways.

First, benchmark and fix the unit price before the arrangement starts, and hold it for the pilot term. Second, if the supplier needs a service charge for stock-holding or site visits, insist on it as a separate visible line. A visible charge can be weighed against the working capital you release and the admin you remove; a charge folded into the rate cannot be measured at all. Third, agree what happens to stock at the end. Who owns slow movers, what obsolescence threshold triggers a review, and who bears the cost of items the supplier pushed in beyond the agreed maximum.

Volume commitment is the other lever. Suppliers will carry stock more willingly against a committed share of category spend, and that commitment is worth more to them than a small price concession. Trading commitment for a fixed rate and a visible service line is usually the cleanest structure, and it keeps the inventory management economics legible to both sides.

The questions to ask in an RFI

Keep the request short and answerable. Ten pointed questions produce better information than forty generic ones. The ones that separate candidates are these.

How many customers do you currently run VMI for, and what is the smallest? What data do you need from us, in what format, at what frequency, and what happens if a file is late or wrong? Who inside your business owns the replenishment decision for our account, and what else does that person do? Where will our stock physically sit, and what is the committed transit time? How often will someone attend our site, and is that frequency contractual? How do you propose min and max levels are set and reviewed? What KPIs do you already report to VMI customers, and will you accept service credits against them? What is your proposed exit process and how long does it take? How is the carrying cost reflected in your pricing? Name two references we may contact.

Answers to the last three tell you most. A supplier who has thought about exit has thought about the whole lifecycle, and a supplier who answers the carrying cost question directly is one you can negotiate with.

Designing the pilot

Pilot one supplier, one category, one site. Choose the site with the cleanest stock records, not the one with the worst problem, because a pilot running on unreliable data tests nothing but your patience. Set the min and max levels from real consumption history rather than opinion, and write them down item by item.

Run in parallel for the first few cycles. Keep raising your own orders while the supplier produces the replenishment they would have made, and compare the two. The gaps are the most instructive conversation of the project, and they surface forecasting weaknesses before they become stockouts. Only when the two lines track sensibly should you switch the manual route off.

Fix the review date, the success criteria and the exit terms in writing before day one. A pilot without an agreed way to stop tends to continue by default, which is how a trial becomes a dependency nobody chose.

The KPIs to write into the agreement

Measure the supplier on outcomes you can verify from your own records. Fill rate or service level against demand is the headline. Stockout events and their duration, since a short stockout on a critical consumable matters more than a long one on something you barely use. Stock accuracy at each count, comparing the supplier's figures to yours. Days of cover held, tracked against the agreed range in both directions, because too much stock is a failure as clearly as too little. Obsolete and slow-moving value, reported monthly rather than discovered at year end. Emergency or expedited deliveries, which should fall to near zero if the model is working. Order and invoice accuracy, because VMI moves data entry to the supplier and errors follow.

Set a review cadence for these figures, monthly at first and quarterly once the arrangement is stable, and give the review a named owner on your side. VMI removes transactional work, not accountability. Within wider supply chain management practice this is the one arrangement where taking your eye off the numbers is most costly, because by design nobody on your team is watching stock day to day.

Making the decision

Bring the scorecard, the RFI answers and the reference calls together, and be honest about the weighting. A supplier who scores well on data and forecasting but poorly on field service can still work if the stock sits in a controlled store you count yourself. A supplier who scores well on everything except financial capacity is a risk that will surface at the worst time. Rank the candidates, then sanity-check the top one against the simplest question of all: if this arrangement went wrong in month four, how quickly could we take replenishment back in-house?

Most of the practical friction sits in the data. If your stock, consumption and supplier records live in one place, sharing a reliable file becomes a routine export rather than a project, and the capability bar for candidate suppliers drops considerably. That is what makes VMI worth attempting with more than one supplier. If you would like to see how teams keep those records together and run supplier performance reviews from the same source, take a look at what ProcureWave covers, or get in touch and we will talk it through against your categories.

Frequently asked questions

What makes a supplier suitable for vendor managed inventory?

Four things, roughly in order of importance: they can receive and use your stock and consumption data without a bespoke project, they hold or can hold stock close enough to serve your lead time, they have people who will look at the numbers every week rather than only when something breaks, and they have the balance sheet to carry inventory on your behalf without needing it back in the price. A supplier missing any one of those will struggle, however good the relationship is.

Should I ask existing suppliers or run a market search?

Start with your incumbents. They already have consumption history with you, the commercial terms are known and the switching cost is zero if the pilot fails. Only go to market if none of your incumbents in that category can meet the capability bar, or if the category is concentrated enough that a specialist distributor would bring real stock-holding advantages. Our supplier management guide covers how to assess incumbents properly before you look elsewhere.

How do I stop the supplier pricing the carrying cost back into the unit rate?

Benchmark the unit price before you start, fix it for the pilot period, and make any change to it a separate negotiation with open figures behind it. If the supplier wants a service charge for holding stock, ask for it as a visible line rather than buried in the rate, because a visible charge can be measured against the savings and a buried one cannot.

How long should a VMI pilot run?

Long enough to cover several replenishment cycles and at least one demand swing, which for most consumable categories means three to six months. Anything shorter tells you the supplier can set the arrangement up, not that they can run it. Write the review date and the exit terms into the pilot agreement at the start.

Can a small supplier run vendor managed inventory?

Yes, and sometimes better than a large one, because a small supplier will often assign a named person rather than an account queue. The limits are financial capacity to carry stock and the ability to exchange data reliably. If both are workable, the attentiveness of a smaller supplier can outperform a larger competitor running an automated process badly.