Decentralised News · Agentic Finance · Decision research
By Heath Muchena · Source review 2026-11-11 · Instrument version 1.0.0
What Matters
Stock on hand is not the same as stock available for a new order, and available stock is not a complete delivery promise. Subtract existing commitments, preserve hold states and check the remaining fulfilment capacity for the required window. DN’s model tests one narrow scenario. It does not verify inventory or guarantee that a supplier can deliver.
A supplier can answer several quote requests correctly in isolation and still promise the same scarce capacity more than once. The unresolved decision is whether this new commitment fits after earlier obligations, unavailable stock and the work required to fulfil it.
DN’s existing supply-chain coverage addresses architecture and control. This article focuses on the units and timing behind a proposed delivery promise. It uses current merchant documentation to separate stock states, then provides a deliberately conservative local capacity calculation. It does not infer live supplier capacity from public product pages.
DN Evidence Block
Sources reviewed: 2026-11-11 · Evidence type: Primary documentation review and original DN method; examples are explicitly synthetic.
Scope: A single product, fulfilment location and delivery window, with stock and processing capacity entered in matching units.
- Shopify distinguishes on-hand, available, committed, unavailable and incoming stock: Shopify: Inventory states.
- Shopify documents explicit draft-order reservation behaviour: Shopify: Creating draft orders.
- The capacity model is DN’s synthetic method. Other systems can use different state names and reservation rules.
Not established by this work: No supplier inventory, factory schedule, simultaneous quote experiment or delivery performance was measured.
Original DN instrument · version 1.0.0
DN Capacity-Aware Delivery-Promise Model
Check a single product and location against current stock and already-netted free fulfilment capacity for one window.
Enter your assumptions or load a labelled example. Required blank numbers stay missing; unknown evidence remains unknown. Use nonpersonal labels only. This instrument makes no network requests and uses no persistent browser storage.
Use your own inputs
No current result. Complete inputs or load an illustrative example.
Uniform units, one location, one window and static inputs only. The model does not reserve inventory or prevent simultaneous commitments.
1. Ask which stock is actually available to promise
The first check is the inventory state. Shopify’s current inventory documentation defines on-hand stock as including available, committed and unavailable units. Incoming stock remains separate until received and made available. Its committed category includes stock reserved for draft orders under the documented current behaviour. These distinctions are platform-specific evidence, not universal accounting vocabulary.
The buyer’s agent should ask for the state relevant to a new commitment, not simply a large stock number. Record product variant, location, unit and timestamp. A warehouse count for a different variant or a location unable to fulfil the buyer’s order does not settle the delivery question.
In DN’s narrow model, usable stock equals on-hand minus committed minus unavailable stock. If the supplier already gives a correctly scoped available quantity, do not subtract those categories again. The input contract matters as much as the arithmetic. A negative result indicates inconsistent inputs or oversubscription that requires review, not a quantity to silently clamp to zero.
Check the timestamp’s meaning as well as its age. A physical count, a last inventory synchronization and a quote creation time can refer to different states. Record which event produced the quantity and whether later commitments have been applied. The buyer does not need every internal warehouse detail, but it needs enough evidence to understand what the supplier is promising at acceptance. Where that evidence cannot be supplied, label the capacity claim as supplier-reported rather than independently verified.
The acceptance record should state the quantity basis explicitly: raw on-hand stock with separate committed and unavailable categories, or an already-netted available balance. Use one representation consistently. A supplier can expose a useful availability statement without sharing every internal allocation, but the buyer should know the statement’s scope and the event at which it becomes binding.
2. Distinguish a quote from a reservation
The second check asks what the quote actually holds. A saved cart, draft order, stock indication and confirmed reservation can have different effects. Shopify’s draft-order guidance is a concrete example of explicit reservation mechanics. Do not assume that every quote removes inventory from sale or that every reservation lasts until the buyer is ready.
Record the reservation identifier, quantity, expiry, release conditions and required next action. Preserve whether the supplier confirmed the hold or the agent merely requested it. When expiry is unknown, the order remains exposed to a change before commitment. A stale hold should not be carried forward as a current assurance.
A buyer considering several suppliers should also avoid creating unnecessary blocking holds. Within the permitted process, release reservations that are no longer needed and retain confirmation. The method here evaluates one proposed commitment. It does not authorize an agent to reserve stock across merchants simply to improve its own optionality at their expense.
For a simultaneous-quote test, use a merchant-approved environment and synthetic orders. Observe whether accepting one quote changes the quantity still available to the next request, and whether an abandoned hold is released under the stated rule. A successful single quote does not answer the concurrency question. The test needs to preserve the sequence and the resulting shared state.
3. Check the work needed inside the promised window
The third check is remaining fulfilment capacity. Stock may still need configuration, quality checks, packing, installation or a dispatch slot. Translate the proposed order into the same capacity unit used by the supplier’s schedule, and subtract work already committed for the relevant window before entering the remaining amount.
The local calculator assumes a simple one-unit-to-one-unit relationship: each product unit consumes one unit of the entered free fulfilment capacity. That is appropriate only when the work is sufficiently uniform. If one order requires specialist preparation or several bottlenecks, use a resource-specific schedule instead of forcing it into this simple model.
Do not add incoming stock merely because it appears on a purchase order. A fuller model may include documented arrivals with explicit timing and uncertainty, but the narrow calculator excludes them. It is designed to reveal whether current available stock and free fulfilment capacity are already sufficient, not to predict how uncertain future replenishment will resolve.
A single aggregate processing number can also hide a deadline bottleneck. Forty free units across the next week do not prove that forty can be dispatched tomorrow. Match the free capacity to the required window, and include the sequence of work where one step depends on another. For mixed orders, a simple unit count may be inappropriate: ten configured machines can require more work than a hundred standard accessories. Use the relevant constrained resource instead of treating all physical units as equivalent.
Capacity evidence should identify whether the figure is gross capacity or already-netted free capacity. The calculator expects the latter. Subtracting existing work twice makes the supplier appear less capable than the inputs justify; omitting it makes a promise easier to satisfy on paper than in practice. Retain the supplier’s definition beside the number.
4. Test the next order against the remaining commitment
The fourth check is the joint limit. In a synthetic example, a location has 100 units on hand, 30 committed and 10 unavailable, leaving 60 usable units. It can process only 45 additional units in the required window. The maximum supported quantity under these inputs is therefore 45, the smaller of stock and free processing capacity.
A proposed order for 50 units has a five-unit shortfall even though the warehouse contains twice that quantity. A proposed order for 40 fits the entered constraints, leaving five units of processing capacity. That result is conditional: it does not inspect the warehouse, transport connection or actual customer requirements.
Now imagine two agents each receive a quote for 40 units before either quote creates a reservation. Each isolated answer appears feasible. Together they require 80 units and exceed both relevant limits. The supplier needs a commitment mechanism that updates the shared state when an order is accepted. A buyer-side snapshot alone cannot eliminate this concurrency problem.
A reservation mechanism needs a release path as well as an acceptance path. If a buyer abandons a confirmed hold and the state never returns to available, the supplier can appear short of capacity that is no longer commercially committed. Conversely, releasing a hold before the customer’s authorized payment completes can expose the order to a race. A permitted test should inspect both transitions and record the rule that resolves them; this article does not assume one universal ordering.
The order decision should preserve both the constraining resource and the remaining quantity. In the 50-unit illustration, processing limits the accepted quantity to 45 despite 60 usable stock units. If a later request consumes the same processing window, the earlier snapshot must be refreshed. Adding another warehouse unit does not solve a processing bottleneck by itself.
5. State what would invalidate the promise
The fifth check is the change trigger. Identify which events require a revised promise: a hold expiry, damaged stock, missed inbound delivery, capacity outage, changed specification or an earlier commitment taking precedence under the contract. Assign an owner and communication path. An agent cannot protect a buyer from a changed condition that no system reports.
Use a dated promise record containing the approved quantity, required arrival window, supplier confirmation, reservation state and evidence scope. Keep dispatch and arrival separate. Completion may depend on a carrier, installation team or buyer access window outside the supplier’s warehouse. The calculator does not cover those dependencies; the record should show them as additional checks.
For a measured study, compare promises with actual completion on a defined set of authorized orders and classify the reason for every miss. Do not treat silence as on-time delivery. The commercial benefit of capacity-aware quoting would be fewer unsupported commitments and less repair work, but this article does not claim that benefit has been observed. It supplies the method needed to test it.
Keep the proposed delivery date distinct from the confidence of the evidence. A date can be specific while resting on several unresolved assumptions. The agent should identify the one assumption most likely to invalidate the purchase decision and seek that confirmation first, instead of presenting a precise date as if precision itself established reliability.
Keep the revision and notification rule with the promise. If a supplier reports a new constraint, identify the affected order lines, the previously approved window and the supported alternatives. The buyer then has a concrete choice about revised terms. Silent editing of a delivery date destroys the evidence needed to understand what changed.
Methodology, limitations and corrections
Define the product, unit, location and required window. Reconcile stock categories to the supplier’s actual system definitions and record whether quoted capacity has already subtracted existing work. Preserve the timestamp and supporting record for every input.
The example is synthetic. A real evaluation should test reservation creation, expiry and release in a permitted environment, then reconcile accepted promises to completion. For several products or shared bottlenecks, use a scheduled resource model instead of adding incompatible unit counts. A quantity that fits this calculator still needs the commercial and delivery conditions checked.
For corrections, identify the claim, source and date when contacting Decentralised News. A new record-creation timestamp does not refresh its underlying evidence.
Change log: version1.0.0, 2026-11-11. Initial documentation review, original instrument and labelled examples.
Frequently asked questions
Is on-hand stock enough to promise delivery?
No. Some stock may already be committed or unavailable, and processing or transport can constrain completion.
Can I enter an available-stock number as on hand?
Only after adapting the categories correctly. Otherwise you may subtract commitments twice. This tool expects the explicit on-hand definition shown.
Does a quote reserve stock?
Only if the supplier’s actual process confirms that effect. Preserve the hold identifier and expiry where applicable.
Why exclude incoming stock?
It keeps the model narrow. Future arrivals require timing and uncertainty assumptions that this calculator does not represent.
Primary sources
- Shopify: Inventory states. On-hand stock includes available, committed and unavailable quantities; draft-order reservations currently affect committed inventory. Reviewed 2026-11-11.
- Shopify: Creating draft orders. A draft quote and a documented inventory reservation are distinct; current reservation mechanics need checking. Reviewed 2026-11-11.
Turn the finding into a decision
The DN Research Desk can scope a capacity and reservation-evidence review for a named purchase workflow, subject to source access and a written quote. Start with one unresolved decision and the records available to answer it. This article does not promise a completed audit, integration or independently verified result.
Related DN research
Disclosure: Analysis and instrument results are editorial. Where labelled affiliate links appear, DN may receive compensation when you sign up or buy through them. Commercial relationships do not determine findings. No tested integration or guaranteed outcome is implied.
