The VMI agreement: clauses, data-sharing terms, and the exit clause (template)

The usual VMI setup advice starts with tooling: pick software, set min/max levels, let the vendor take it from there. That's not where these programs fail. They fail in the paper — at the count that never matches, the shrinkage nobody agreed to eat, and the exit that neither side wrote down because asking felt awkward. So this article runs setup in contract order: a clause-by-clause agreement template you can adapt this week, the data-sharing terms a Shopify seller actually needs to send, and an exit clause with template language. Whether VMI makes sense for you at all, and how the day-to-day works once it's live, is owned by our vendor managed inventory guide; who owns the goods is a separate axis, covered in our VMI vs consignment comparison. This article owns the paperwork.

What the agreement has to settle

VMI hands reorder decisions to the supplier. In VMI, the vendor monitors your sell-through and initiates replenishment instead of waiting for purchase orders — that's the standard definition (Investopedia, accessed November 2026) — and the agreement is the only thing standing between that arrangement and "the vendor ships whatever, whenever." Four disputes kill more VMI programs than bad software ever has:

• Count mismatches. Your shelf says 31, their system says 40, and neither side's replenishment logic is right until someone owns the reconciliation.
• Shrinkage. Units damaged in transit, lost in your back room, or missing at count — each has a different natural owner, and the default is whoever kept worse records.
• Stale data. The vendor ships against last Monday's file while a weekend campaign emptied the shelf. Without a latency clause, the vendor did nothing wrong and you still ran dry.
• No exit. Six months in, the vendor's shipments don't fit your demand anymore and there's no written way out. Programs without exit clauses don't end; they decay.

The template below addresses each in contract language you can shorten for a small supplier.

The agreement, clause by clause (template)

Numbered clauses, with the fill-ins that make it work for a small program:

1. Scope and SKUs. List exact SKUs — e.g., PET-CARRIER-M-02, PET-CARRIER-L-01 — covered at one location. Add: "No new SKUs enter this agreement without a written amendment."
2. Data sharing. Weekly file, every Monday by 09:00 store time, covering: on-hand quantity per SKU, units sold over the trailing 7 and 28 days, inbound purchase orders, open returns. Format: CSV. As-of timestamp: Monday 00:00.
3. Min/max levels and review. Starting values agreed in writing (an illustrative table below). Changes require written consent with 14 days' notice. Joint review quarterly.
4. Replenishment authorization. "Vendor ships to restore max level when on-hand plus in-transit falls below min. Shipment confirmations within 2 business days. Buyer may pause replenishment up to 30 days per calendar quarter with written notice, for campaigns, launches, or cash timing."
5. Lead times. Stated per agreement, not assumed: "Lead time is 32 calendar days, order to sellable stock. Material changes require 30 days' notice." (Illustrative — use your measured number.)
6. Title and payment. "Title passes on receipt at buyer's dock; invoices payable net 30 from receipt." This sentence is where VMI ends and consignment begins — if you want the vendor to own stock until sale, that's a different model with different clauses, covered in the comparison guide linked above.
7. Shrinkage and damage. Loss between receipt and sale sits with the buyer (title follows clause 6). Units damaged before receipt are replaced or credited by the vendor within the claim window.
8. Counts and audit. Vendor may count monthly with 5 days' notice; buyer cycle-counts quarterly and may audit the vendor's replenishment records. Discrepancies above 2 percent on any SKU trigger a joint recount within 10 days.
9. Term and exit. 12-month initial term. Either party exits with 90 days' written notice under the exit clause below — which is spelled out in full, not referenced.

Data-sharing terms for a Shopify store

Clause 2 is easy to fill from Shopify: the admin exports inventory and sales as CSV (Shopify Help Center, accessed November 2026), and a saved view per VMI SKU group makes the export a two-minute Monday task. Two Shopify realities shape the clause, so set expectations with the vendor up front:

• There is no native low-stock email alert. Shopify's native options are saved views plus Shopify Flow automations (Shopify Help Center, accessed November 2026), so the agreement must name who watches levels between the Monday sends — "both parties" means nobody.
• The Monday file has a known blind spot. A Friday flash sale that empties PET-CARRIER-M-02 doesn't reach the vendor until next Monday. The 30-day pause right in clause 4 is how you cover launch weeks without arguing about stale data afterward.

If the vendor wants machine-readable data instead of CSVs, Shopify shipped a read-only Purchase Order API on September 17, 2026 (shopify.dev changelog, accessed November 2026) — read-only, so it exposes your PO history without letting the vendor write anything to your store.

One tooling note, stated plainly: we build InventoryX, and its weekly Monday Buying List runs on the same typed inputs this agreement fixes — per-SKU days of cover and lead times you enter per supplier. The limit: V1 produces the list and one-click draft POs; it doesn't transmit anything to a vendor, so the weekly data send stays a manual step in your process either way.

Min/max clauses without the math fight

The fight over min/max is usually a fight over whose forecast is real, so the clause's job is only to freeze who can change the numbers and how often. The values themselves come from your demand and lead time — the arithmetic is the standard reorder point method covered in our reorder point formula guide. Illustrative starting values for the SKUs in clause 1:

The mins cover the lead-time demand with a small buffer; the maxes cap the pipeline so a vendor's production-push can't park 100 days of stock in your back room. When the vendor proposes raising the min from 40 to 70, the clause gives you the procedure — written proposal, 14 days' notice, quarterly review — instead of an argument in an inbox.

The exit clause (template language)

The clause most agreements leave out, in full:

Exit and wind-down. Either party may terminate this agreement on ninety (90) days' written notice. Upon notice: (a) the Vendor completes goods already in production and ships them on the agreed schedule; (b) the Vendor pauses new replenishment shipments except to restore minimum levels during the notice period; (c) the Buyer pays invoiced amounts under standard terms; (d) the Vendor returns, in machine-readable format, twelve months of shipment and inventory history within 15 days of the final shipment; (e) replenishment control reverts to the Buyer on the effective exit date; (f) confidentiality and the data-use limits in Clause 2 survive for 24 months. Neither party owes a termination fee; each bears its own transition costs.

Why each part earns its line:

• The pause on new shipments. Without it, a vendor can fill your pipeline to the max in week one of the notice period and leave you holding the stock.
• The data handback. Your reorder history is your asset. Without it, your first self-run cycle is blind — you'd re-derive lead times and velocities from scratch.
• The reversion date. Names the day replenishment becomes your job again, closing the "nobody's ordering" gap that turns an exit into an empty shelf.
• No termination fee. Symmetric exits keep the threat of leaving from becoming a bargaining chip inside the relationship.

The first 30 days, as a checklist

• Week 1: pick one supplier, five to ten SKUs. A small pilot on your most predictable supplier proves the model without betting the catalog.
• Week 2: agree the data spec and starting min/max in writing. The two clauses that require the most thought are 2 and 3; everything else is standard.
• Week 3: sign, then send the first data file on schedule. Sending on time the first week sets the cadence the whole program runs on.
• Week 4: first vendor-driven shipment, then reconcile counts. Compare their shipment record to your received counts immediately — discrepancies found in week 4 are process problems; found in month 6, they're disputes.

Common questions

What data do I share with a VMI vendor? On-hand per SKU, units sold over the trailing 7 and 28 days, inbound purchase orders, and open returns, on the fixed weekly cadence stated in the agreement. Add lead-time changes and planned campaigns, since both change what the vendor should ship. Share only the SKUs in scope — nothing else.

Who sets min/max levels in VMI? Both parties, in writing. The vendor proposes from its production view; you adjust from your demand reality; the agreement should require written consent, 14 days' notice for changes, and a quarterly joint review. Undefined edit rights are the quietest cause of overstock in small VMI programs.

How do I get out of a VMI agreement? Through the exit clause: written notice (90 days is common), completion of in-production goods, a pause on new shipments, final payment, and a data handback so your first self-run cycle isn't blind. If your agreement has no exit clause, negotiate one while relations are good — once a dispute starts, terms get worse, not better.

Do I need special software to run VMI? No. A weekly CSV export, a shared spreadsheet, and a kept cadence cover small programs. Shopify offers saved views and Flow automations but no native low-stock email alert (Shopify Help Center, accessed November 2026), so the agreement must name who monitors between sends. Software helps at scale, not at week one.

Is VMI the same as consignment? No. VMI assigns reorder decisions to the vendor; consignment assigns ownership of the goods to the vendor until sale. They're independent axes that can be combined. The full comparison, including who pays when and the accounting consequences, is in our VMI vs consignment guide.

Next step

Draft the exit clause before anything else and attach it to your supplier conversation this week — a program that can't end cleanly shouldn't start. Then fix the data spec: pick your Monday send time, export your on-hand and 28-day sales, and check the numbers you'd hand over are ones you'd act on yourself. The shipments the vendor schedules arrive like any other inbound order — the receiving steps are in our purchase order process guide.