The problem ATP solves
When a customer places an order, they want an answer to a simple question: when will I get this, and how much of it? The naive answer — "we'll check and get back to you," or worse, a delivery date typed in optimistically without checking anything — is exactly what causes broken promises downstream. ATP is the discipline of answering that question properly at the point of order entry, by actually checking what is available: existing finished stock, planned production, incoming shipments, and open capacity, rather than assuming supply will simply appear when needed.
The stakes of getting this wrong run in both directions. Overpromise — confirm a date the business cannot actually hit — and the company breaks its word, damages trust, and often incurs real cost fixing the fallout (expedited freight, apologies, compensation). Underpromise — quote an overly conservative date "just to be safe" — and the company loses sales to competitors willing to commit to a faster, equally achievable date. ATP exists to find the honest answer in between: the fastest date the business can reliably deliver on.
It helps to be precise about what "available" means in this context, because it is not the same as "in stock." Stock that is physically present but already promised to another order is not available to a new customer — it has to be excluded from the pool ATP checks against. Conversely, stock that does not yet physically exist but is firmly scheduled — a confirmed production run, a container with a firm booking and sailing date — can legitimately be counted as available for a future date, even though nothing is sitting on a shelf today. ATP, done properly, is really a check against a time-phased supply picture, not a single static stock count.
ATP looks different depending on how you fulfil
What "available" actually means to check depends entirely on the company's fulfilment strategy — how far along the product is before a customer order even arrives.
| Fulfilment strategy | What ATP checks | Simple example |
|---|---|---|
| Make-to-stock | Existing finished-goods stock (on hand and firmly incoming) | A retailer promises next-day delivery on an item already sitting in the warehouse — ATP simply confirms enough units exist and are not already reserved for another order |
| Assemble-to-order / configure-to-order | Availability of the component parts, plus the time needed to assemble them | A furniture retailer promises a custom-colour couch in 10 days — ATP checks that the frame, chosen fabric and legs are all in stock, then adds the standard assembly time |
| Make-to-order | Full production lead time, from raw material through to finished, scheduled build capacity | A specialist equipment manufacturer promises delivery in 12 weeks because that is genuinely how long procurement plus production takes, checked against the current production schedule's open slots |
Notice the pattern: the further "upstream" a company fulfils from (make-to-order being the most upstream, make-to-stock the most downstream), the more of the supply chain ATP logic has to look into before it can give an honest answer. A make-to-stock promise is a simple stock check; a make-to-order promise is really a mini capacity-and-material feasibility check dressed up as a delivery date.
Allocation planning: deciding who gets scarce stock before orders even arrive
When supply is genuinely scarce relative to expected demand, simply promising stock on a first-come-first-served basis to whoever happens to order first can quietly starve a company's most important customers or channels — a small, low-priority order placed an hour earlier can lock up stock a key account needed later that day. Allocation planning addresses this by deciding, in advance of individual orders arriving, how available (or expected) supply should be split across customer segments, channels or regions, using a customer hierarchy (key accounts, standard accounts, specific channels) and a set of allocation rules the business has agreed.
In practice this might mean reserving a portion of incoming stock exclusively for contracted key accounts, or capping how much any single order can draw from a tight pool so no one customer exhausts the available quantity. ATP then checks a customer's specific order against the pool that customer or channel is entitled to draw from, not against the company's total supply — meaning two customers asking for the same product on the same day can legitimately receive different answers, because they sit in different allocation buckets.
Allocation rules also need periodic review, not a set-and-forget policy. A customer hierarchy built around last year's biggest accounts can quietly misallocate scarce stock if a smaller account has since grown into one of the business's most important relationships, or if a channel that used to be marginal has become strategically important. Because allocation decisions are made ahead of actual demand, and often ahead of full visibility into which specific orders will materialise, they inevitably involve some forecasting risk of their own — reserve too generously for one segment and stock can sit unused while another segment is turned away.
Order promising as a search, not a single lookup
Conceptually, order promising can be thought of as a search process rather than a single yes/no lookup. If the requested delivery date and location cannot be met from what is immediately available, the system does not simply say "no" — it works outward, checking successive future time periods (can it be delivered a few days later, once the next stock arrives?) and sometimes alternative locations (can a different distribution centre or warehouse cover this order instead?) until it finds the earliest combination of date and source that is genuinely feasible. The customer is then offered that feasible promise, rather than either an impossible best-case date or an unnecessarily conservative worst-case one.
There is no need to understand this as an algorithm to appreciate why it matters: the quality of an ATP promise depends entirely on how wide and how current a view the search has into real supply. A search that only looks at one warehouse's stock, or that is working from data that is a day out of date, will systematically produce worse promises than one with full network visibility and near-real-time data — which is exactly why master planning, which already tracks planned supply across the whole network, is such a natural foundation for a company's ATP logic to build on.
The business cost of getting ATP wrong
Poor ATP shows up on both sides of the ledger. Overpromising — confirming dates the business cannot reliably hit, often because ATP logic is naive, out of date, or simply ignored under sales pressure — leads directly to missed delivery dates. Missed dates erode customer trust, and trust, once broken repeatedly, is expensive to rebuild; it also frequently forces expensive last-minute fixes such as air-freighting a shipment that should have travelled by sea, or re-routing stock from another region at the customer's or the company's cost.
Underpromising carries a quieter but real cost: a company that is systematically too conservative — quoting longer lead times than it actually needs "just to be safe" — loses orders to competitors who can honestly commit to a faster date, even when the underpromising company could genuinely have delivered just as fast. Both failure modes trace back to the same root cause: ATP logic that is not properly connected to real, current supply and capacity information. A reliable ATP process, closely tied to actual data, is what allows a company to make the fastest promise it can honestly keep.
The South African angle: promising delivery on stock still at sea
A South African importer's ATP problem is harder than a purely domestic one, because "available" cannot mean only "physically sitting in our local warehouse." A meaningful share of sellable stock, at any given moment, is somewhere on the water between the load port and Durban, Cape Town or Ngqura — tracked, for a business that stays on top of it, through container arrival alerts — or already discharged but still moving through customs clearance. A realistic ATP calculation has to treat that in-transit stock as genuinely available — but only as of its realistic arrival-and-clearance date, not the date it left the origin port, and certainly not "today."
This is where the honesty of the promise depends heavily on the honesty of the underlying estimated time of arrival (ETA) and the realistic customs-clearance buffer added on top of it, not just the shipping line's optimistic sailing schedule. An importer whose ATP logic assumes a container clears customs and reaches the warehouse the same day it berths is setting itself up to break promises the moment there is any hold-up at the terminal. Given how routinely South African port congestion turns a scheduled arrival into a delayed one, an importer's ATP system needs a realistic, not optimistic, buffer built in — and needs to re-check and, where necessary, walk back promises the moment a vessel's ETA slips, rather than letting a stale promise sit uncorrected in a customer's inbox. Getting this discipline right is directly what protects the customer-facing metrics covered in the perfect order concept — a promise that cannot be kept is, in effect, a defect before the order has even shipped.
Frequently asked questions
What does "Available-to-Promise" actually mean?
It means the quantity of a product genuinely available to commit to a new customer order, at a specific date, after accounting for what is already reserved for other orders. It is checked at the moment an order is placed so the delivery date quoted to the customer is one the business can realistically keep.
Why does ATP work differently for make-to-order versus make-to-stock companies?
Because "available" means something different in each case. Make-to-stock ATP is a simple check of existing finished-goods stock. Make-to-order ATP has to check full production lead time and current schedule capacity, since nothing exists yet until an order triggers production.
What is allocation planning and why does it matter for ATP?
Allocation planning splits scarce available or expected supply across customer segments, channels or regions in advance, using agreed business rules, rather than serving orders purely first-come-first-served. ATP then checks each order against the specific allocation pool it belongs to, so key accounts or priority channels are protected from being crowded out by unrelated orders.
Why can't an SA importer just promise delivery based on what's in the local warehouse?
Because a meaningful share of stock is often still in transit or moving through customs at any given time. A realistic ATP calculation includes in-transit stock, but only counted as available from its realistic arrival-and-clearance date — including a buffer for the port delays that are common in South Africa — not from the date it left the origin port.
What happens when a business consistently overpromises delivery dates?
Repeated missed delivery dates erode customer trust and often force costly last-minute fixes such as expedited freight. Over time, customers learn not to rely on the company's promises, which damages the relationship even when individual orders are eventually delivered.