AI uses for records
Supplier minimum order reasoning data: when the lowest unit price is the wrong buy
By SourceX Editorial · Updated
10 sections
Short answer
Judge a purchase in two passes. First drop any option that breaks a hard constraint: an unapproved supplier, a superseded specification, a shelf life shorter than the stock will sit, missing storage or a customer source rule. Then rank the remaining options by cost per unit actually used, counting expected expiry and carrying cost, not by quoted unit price. Record the evidence for each step so the decision can be checked later.
Value drivers
Supplier minimum order
Key takeaways
- A price break only matters after an option passes every hard constraint, each proven by its own document.
- Compare options on cost per unit you expect to consume within shelf life or obsolescence limits, plus holding cost.
- U.S. federal law already frames the rule: the most advantageous total and unit cost, without exceeding the quantity reasonably expected to be required.
- Relative prices and percentages let a decision packet keep its reasoning without exposing confidential supplier pricing.
- Last-time buys and supplier allocation are legitimate exceptions, recorded with the document that triggered them.
How do you evaluate a purchase when the lowest unit price is the wrong choice?
Evaluate it in two passes. First remove every option that breaks a hard constraint: an unapproved supplier, an outdated specification revision, a shelf life shorter than the time the stock will sit, storage you do not have, or a customer contract that names the source. Then compare the options that remain on the cost of each unit you expect to actually use, including carrying cost and expected write-offs, not on the quoted unit price.
The unit price on a quote is a fact about the supplier's price list. Whether the purchase was right is a fact about your demand, your constraints and what happened after the goods arrived. A purchase order that took the cheapest price tier can still be a bad decision if half the lot expired on the shelf, and a buyer who paid more per unit for a smaller lot may have made the better call.
That gap between price and outcome is what makes these records valuable as reasoning data. A procurement file that shows the constraint, the options, the approval and the delivery result lets someone check the decision instead of guessing at it. This guide turns that file into a task specification a team can fill in for each minimum order quantity (MOQ) decision.
Which constraints disqualify the cheapest option before price is compared?
A constraint disqualifies an option when no price advantage can compensate for it. Treat these as pass or fail checks, record the document that proves each one, and only then rank what survives.
| Constraint | Record that proves it | What it rules out | Common trap |
|---|---|---|---|
| Specification revision | Item master revision, engineering change notice, customer drawing | Stock or quotes built to a superseded revision | The quote references a part number that did not change when the revision did |
| Supplier approval | Approved supplier list entry, qualification record, audit date | A cheaper source that is not yet approved for this item | An approval exists for the supplier but not for this item or site |
| Shelf life or obsolescence | Lot expiry data, product life notes, end-of-life notice | Quantities that cannot be consumed before they expire or become obsolete | Demand is measured in months while shelf life is recorded in days |
| Storage and handling capacity | Warehouse capacity plan, hazardous material limits, cold chain slots | Lots that physically do not fit or need special storage | Capacity is checked for the pallet count but not for temperature class |
| Customer or contract source rule | Customer contract, quality agreement, approved manufacturer list | Any source the customer has not approved | The rule lives in a contract annex the buyer never sees |
| Need date versus lead time | Vendor lead time on the pricelist, open sales orders, production schedule | An option that arrives after the date the goods are needed | Lead time on the quote differs from lead time in the item master |
| Approval authority | Delegation of authority matrix, approval log | Order values or quantities above the buyer's authority without sign-off | The large tier crosses a threshold that requires a second approver |
What public purchasing rules say about quantity and unit price
U.S. federal procurement law states the trade-off directly. Under 41 U.S.C. 3310, an executive agency buys supplies in the quantity that gives the most advantageous total cost and unit cost where practicable, and that does not exceed the quantity reasonably expected to be required. The second limb is the one a cheapest-tier purchase most often breaks: a lower unit price does not justify buying more than will be needed.
The Federal Acquisition Regulation provision at 52.207-4, Economic Purchase Quantity, invites offerors to say whether the requested quantity is economical and to recommend other quantities with a total and a unit price for each. It defines an economic purchase quantity as the quantity at which a significant price break occurs, and says the information is collected to avoid buying in disadvantageous quantities. The Revolutionary FAR Overhaul class deviations for Part 7 issued in 2025 trimmed the surrounding policy text, and acquisition.gov's Part 7 deviation guide lists 52.207-4 among the provisions kept without text changes.
FAR 15.101 describes a best value continuum: when a requirement is clearly definable and performance risk is low, price may dominate, and as requirements get less definite or riskier, technical and past performance factors weigh more. Even the lowest price technically acceptable process in FAR 15.101-2 selects the lowest evaluated price only among proposals that meet the acceptability standards. Both passages are the codified FAR text; agencies that have adopted the Revolutionary FAR Overhaul class deviations for Part 15 move the continuum guidance into the FAR Companion and relocate the lowest price technically acceptable approach, so check which version an agency follows. Private companies are not bound by any of this, but the logic transfers: price ranks options only after they qualify.
Where MOQ, order multiples and price breaks are recorded in an ERP
Most of the inputs already exist as fields; the work is to snapshot them as they stood on the decision date. In Odoo's Purchase app, for example, the vendor pricelist on a product form carries Quantity, Unit Price and Delivery Lead Time columns by default, which is where quantity price breaks and lead times are kept per vendor.
Replenishment rules add the demand side. Odoo's documentation describes reordering rules with a Min, the forecast level that triggers a replenishment order, and a Max, the level stock is replenished to. Its Multiple field rounds the order quantity up to the nearest multiple, so a vendor that sells only in cases of six turns a need of 32 units into an order of 36 and leaves stock above the maximum. That rounding is a small, built-in version of the MOQ problem.
Odoo defines vendor lead time as the number of days from purchase order confirmation to receipt. Other ERPs use different names for the same ideas, so the task specification below uses neutral field names and asks you to map each one to the system's own field. Keep the system export as evidence; do not retype values.
The MOQ demand and carrying cost task specification
Each task is one purchasing decision: the options that were available, the constraints in force, the choice made, who approved it and what happened afterward. Fill one packet per decision. Fields marked as outcome must come from records created after the order, because they are what a grader checks the decision against.
At a glance
The MOQ demand and carrying cost task specification
| Field | Source record | Rule for the task | Why it matters |
|---|---|---|---|
| decision_id and decision_date | Purchase requisition or RFQ | One packet per decision; date is when the order was approved | Every other field is snapshotted as of this date |
| item_and_revision | Item master and change notices | Record the revision in force on decision_date | Catches orders placed against a superseded design |
| options | Quotes and vendor pricelist rows | Each option lists supplier, quantity tier, relative unit price, lead time and approval status | The grader needs every choice, not only the one taken |
| demand_basis | Consumption history, open orders, forecast version | State the demand per period and the forecast version used | Shows whether the quantity matched expected need |
| coverage_limit | Shelf life, end-of-life notice, storage capacity | Express as a duration or unit cap; cite the record | Turns expiry and space into a hard ceiling on quantity |
| carrying_rate_assumption | Finance policy or stated assumption | A percentage per year, labeled as company policy or assumption | Makes holding cost explicit instead of implied |
| constraint_checks | Approved supplier list, contracts, specification | Pass or fail per constraint with the document reference | Removes invalid options before any price comparison |
| choice_and_rationale | Purchase order, approval comments, email thread | Selected option plus the reason as written at the time | Captures the reasoning, not a reconstruction |
| approval_trail | Approval log, delegation of authority matrix | Approver role and whether a threshold was crossed | Shows the decision was authorized |
| outcome_evidence | Receipts, consumption, expiry write-offs, stockouts, returns | Measured over the coverage period after delivery | Lets a grader judge the decision by what happened |
How to compare surviving options on cost per unit actually used
Rank qualified options by the effective cost of each unit you expect to consume, expressed relative to a baseline option so no currency figures are needed. The steps below use only relative prices, quantities, durations and percentages, which keeps the packet shareable even when supplier pricing is confidential.
Checklist · 6 items
- Usable units: the smaller of the order quantity and the demand expected within the coverage limit (shelf life, obsolescence date or storage cap).
- Expected waste: order quantity minus usable units; treat it as a write-off unless the record shows a resale or return right.
- Relative purchase cost per usable unit: relative unit price multiplied by order quantity, divided by usable units.
- Holding cost per usable unit: carrying rate per year, multiplied by average units on hand, by the months they are held divided by 12 and by the relative unit price, then divided by usable units.
- Effective relative cost: purchase cost per usable unit plus holding cost per usable unit; the lowest effective cost wins among qualified options.
- Sensitivity check: rerun with demand 20 percent lower and 20 percent higher; if the ranking flips, record the decision as demand-sensitive.
0 of 6 ticked. Tick each item as you go.
Worked example (illustrative): a price break that loses to expiry
All figures here are invented for illustration. A plant consumes an adhesive at about 400 units a month. The approved supplier quotes two tiers: 1,000 units, or 5,000 units at a unit price 12 percent lower. The adhesive has a 9-month shelf life from receipt, and the company's stated carrying rate assumption is 20 percent a year.
The 5,000-unit tier covers 12.5 months of demand, but only 3,600 units can be used within 9 months, so about 1,400 units, 28 percent of the lot, are expected to expire. Relative cost per usable unit is 0.88 times 5,000 divided by 3,600, about 1.22 times the base price. Stock on hand falls from 5,000 to 1,400 over 9 months, an average of 3,200 units, which adds about 0.12 of holding cost per usable unit, for an effective relative cost of about 1.34.
The 1,000-unit tier covers 2.5 months with no expected expiry; an average of 500 units held for 2.5 months adds about 0.02, so its effective relative cost is about 1.02. The cheaper unit price ends up about 30 percent more expensive per used unit once expiry and holding are counted. A well-built packet records both calculations, the shelf-life record, the decision and, later, the actual write-off quantity, so a grader can see whether the buyer's reasoning held.
The exception: when buying above expected demand is the right call
Buying more than near-term demand can be correct when the alternative is not having the item at all. A last-time buy after an end-of-life notice, a supplier placing customers on allocation, or a customer contract that requires a set level of safety stock can all justify a quantity that fails the simple coverage test.
The task should not mark these as errors. Record the triggering document, such as the end-of-life notice or the allocation letter, as a constraint, and state the decision rule that applies: for a last-time buy, usable units are measured against the remaining lifetime demand of the product, not the next few months. A grader then judges the exception against the rule the company actually faced, which is what makes the case useful for training judgment rather than a fixed formula.
Grading a packet, and the signals that call for a procurement expert
Grade each packet on three things: whether invalid options were excluded with evidence, whether the chosen option had the lowest effective cost among qualified options under the stated assumptions, and whether the outcome records confirm or contradict the expected usage. Keep the grade separate from hindsight; a decision can be sound on the information available and still end badly.
Stop and bring in a procurement or supply chain specialist when the packet depends on a contract term you cannot read, when demand came from a forecast whose version you cannot identify, when regulated goods are involved, or when the carrying rate assumption swings the ranking. Narrow the scope instead of guessing: a smaller set of complete, evidenced packets is worth more than a large set with blank constraint checks.
Turning purchase decision files into licensable reasoning data
Procurement records also hold other parties' information. Supplier quotes and price lists are often confidential under the supplier's terms, approval comments name employees, and customer source rules sit in contracts with their own confidentiality clauses. Holding the records does not by itself give the right to license them, and relative prices may still need masking when a tier structure could identify a supplier.
Purchase decision files like these can become a licensed AI dataset once SourceX has run rights review, contracting, preparation and delivery on them. A procurement team checks fit through a short assessment without uploading data, and its files are released only once it approves the scope and terms and signs.
Frequently asked questions
What should an MOQ demand and carrying cost task specification include?
It should include the decision date, the item and specification revision, every option with its quantity tier, relative price, lead time and approval status, the demand basis, a coverage limit such as shelf life, the carrying rate assumption, pass or fail constraint checks with document references, the choice and its written rationale, the approval trail and outcome records from after delivery.
Which constraint most often makes the cheapest procurement choice invalid?
In practice it is usually a constraint recorded outside the quote: a specification revision, a supplier approval that does not cover the item, a shelf life shorter than the stock will sit, or a customer contract naming the source. Each can only be checked against its own record, which is why the specification asks for document references.
How do you compare price breaks without sharing confidential prices?
Express every option's unit price relative to a baseline option, such as 0.88 times the base tier, and express holding cost as a percentage. The ranking and the reasoning survive, while the actual supplier prices stay out of the dataset.
Is buying above the minimum order quantity ever correct?
Yes. A last-time buy after an end-of-life notice, supplier allocation or a contractual safety stock requirement can justify buying beyond near-term demand. Record the triggering document and measure usable units against the demand that rule covers.
When should a team stop and get specialist review?
When the decision depends on a contract term nobody can read, a forecast version that cannot be identified, regulated goods, or a carrying rate assumption that changes the ranking. Narrowing the dataset to complete packets is better than filling gaps with guesses.
Sources
- 41 U.S.C. 3310 requires agencies to buy supplies in the quantity giving the most advantageous total and unit cost where practicable, not exceeding the quantity reasonably expected to be required. Source
- FAR 52.207-4 invites offerors to recommend economic purchase quantities with total and unit prices, defines an economic purchase quantity as where a significant price break occurs, and aims to avoid disadvantageous quantities. Source
- The FAR Overhaul Part 7 deviation guide keeps provision 52.207-4 without text changes while trimming surrounding Part 7 policy text. Source
- FAR 15.101 describes a best value continuum in which price may dominate for definable, low-risk requirements and technical factors weigh more as risk rises. Source
- FAR 15.101-2 awards on the lowest evaluated price only among proposals meeting the acceptability standards for non-cost factors. Source
- Odoo's vendor pricelist on the product form has default Quantity, Unit Price and Delivery Lead Time columns. Source
- Odoo reordering rules use Min and Max levels, and the Multiple field rounds order quantities up, for example from 32 to 36 units in cases of six. Source
- Odoo defines vendor lead time as the days from purchase order confirmation to receipt of products. Source
- A Department of Labor class deviation implementing the FAR Overhaul for Part 15 relocates the tradeoff and lowest price technically acceptable approaches, and agency deviation memos move best value continuum guidance to the FAR Companion. Source
See if your company qualifies
A short company assessment. No data uploads are needed.