Purchasing

Your open PO book answers “how much?” three ways

An on-hand report tells you what is on the shelf. Your open purchase orders tell you what else you have already promised to buy — but “how much?” has three answers there: the units still to arrive, the value of those units, and the cash still to leave your account. They sit on different bases, they can rank your suppliers differently — here two of the three agree and one does not — and a fourth figure gets quoted as though it were all three. The six orders below are invented so every figure can be checked.

Ask what you have of a product and the answer is the on-hand figure. It is correct, and it is half the position. Behind it are orders you have signed off, units a supplier is already cutting fabric for, and deposits that have cleared your account. None of that is on the shelf; all of it is committed. The difficulty is that “how much is committed?” does not have one answer, and the answers will not reconcile until you say what each one is counting.

The open book, laid out once

Six invented open orders across three suppliers, at $12, $15 and $45 a unit. Every later table derives from this one.

Order 1063 is the awkward one, and it is the reason for the extra column. It was raised for 900 units at $15; 600 have been received and 300 remain. Its full value is $13,500, but only $4,500 of that is goods still to come. The other five orders are untouched, so for them the two dollar columns agree — which is exactly why a book with one partial receipt in it is so easy to misread.

Three measures, and a fourth number that is neither

The first two are the same goods counted in two units of measure: units and dollars, the same 6,400 items. The third is on a different axis altogether, because money moves on payment terms rather than on deliveries. The fourth is a size-of-book figure and belongs to neither pair.

The $120,000 splits two ways, and the two splits do not line up. By receipt: $9,000 of goods already in, $111,000 still to come. By payment: $39,150 already paid, $80,850 still owed. Same total, different cuts, different questions.

Where the two cuts cross is the number no report has a column for. Of the $111,000 of goods still to come, $30,150 has already been paid — deposits, plus one order settled in full ahead of delivery — and $80,850 has not: $30,150 plus $80,850 is $111,000. That $30,150 is cash out of the account against goods not yet on the shelf. It is not stock, not bank balance and not a payable, and it appears only when you hold two of these measures against each other.

A single number labelled “committed” is not so much wrong as unanswerable. The first question to ask any open-PO report is which of these four it is showing you.

Reading 1 — Units still to receive

Units still to receive is coverage you have already bought, and it belongs next to on-hand stock in any conversation about running out. The trap is the definition. The column you want is ordered minus received, floored at zero, on orders that are still open. Order 1063 shows why. It was raised for 900 units, 600 are booked in, 300 remain. Those 600 are on the shelf and already inside on-hand stock, so adding the ordered figure counts them twice.

Six hundred units against a real figure of 6,400 is an overstatement of 9.4%. It does not announce itself, because both numbers look plausible and both come from the same orders.

Units still to receive is the one reading of the three that can be double-counted against a report you already trust. That is what makes it the easiest to get wrong.

Two rules keep it honest. Count only orders that are genuinely open, and floor each line at zero rather than letting an over-receipt run negative.

Reading 2 — Open order value

The second reading puts money against the same 6,400 units: $111,000 of goods ordered and not yet received. That is the figure competing with your next order. Whether to place another $40,000 is not a question about today’s bank balance on its own; it is whether you can carry $40,000 on top of $111,000 of goods already bought and not yet arrived, part of which you have paid for and none of which you can sell. What that has to clear, and when, is a question of timing, and the ladder further down is where it gets answered.

SKU Compass carries a summary card on the purchase orders page labelled Committed Capital · on order. Treat a label like that the way you would treat any column heading: as a name, not as an accounting definition. What an approved but undelivered order represents in your books — an obligation, an encumbrance, nothing yet — depends on how your system and your accountant treat it, and that is a conversation to have with them rather than a point this article can settle. The narrower question is the checkable one: before you act on any dollar column on an open-PO report, find out whether it is on the full-order basis or the still-to-receive basis. On these six orders the two differ by $9,000.

Value per unit falls out of the same two numbers, and it doubles as a check that you have them on the same basis. Across the book, $111,000 over 6,400 units is $17.34 a unit once rounded. Per supplier, nothing like it:

Those three per-unit figures are the actual unit costs on the orders, and that is the point: when numerator and denominator cover the same goods, the division gives back a price you recognise. Put full order value over remaining units instead and supplier B comes out at $31,500 divided by 1,500, or $21.00 — a number that is not what B charges for anything, because 600 units left the denominator while their $9,000 stayed in the numerator. If a per-unit figure on your own report is not a price you recognise, that mismatch is the first thing to look for.

Supplier C holds 14% of the units still to come and 36% of the open order value. Supplier A is the mirror image: 62.5% of the units and 43% of the money. A dominates your unit exposure; C is disproportionately large in dollar exposure. Neither is only one kind of problem — a late order from C is still stock you do not have, and A’s $48,000 is still capital exposure you have already committed to — but the two are large in your book for different reasons, and a report sorted one way will show you only one of them.

Reading 3 — Cash still owed, by supplier

The third reading is on a different axis from the first two, and saying so is most of the work. Units still to receive and open order value both describe goods that have not arrived. Balance owed describes money that has not moved, and money moves on payment terms. The two can come apart completely: order 1063 still has 300 units to come and owes nothing at all, because it was settled in full against its partial delivery.

Of the $120,000 of orders still open, $39,150 has been paid — 30% deposits on 1041 and 1052, 50% on 1074, and 1063 paid in full. That leaves $80,850 still to go out.

SKU Compass groups the open book by supplier in a collapsible card headed Open value by supplier, whose columns run Open POs, Units On Order, On Order $ and Balance Owed against each supplier. Grouped that way, the six orders look like this — with the dollar column on the still-to-receive basis, and the full-order total carried alongside so the two stay visibly apart:

Sort it and the answer changes. By units still to receive: A, B, C. By open order value: A, C, B. By balance owed: A, C, B again. Two of the three agree and one does not, and which one breaks ranks is not something you could have predicted from the other columns — B drops to the bottom on money because one of its two orders happens to be paid in full, which has nothing to do with how many units it still owes you. Whichever column your report sorts on decides which supplier you look at first, and it will not tell you that another sort hands you a different name.

Plan the next order against the commitment, not the shelf

Two invented SKUs:

On shelf stock alone, K‑310 looks two and a half times safer: 50 days against 20. Now add what is already ordered. J‑220 has 1,200 units due in 12 days and will sell 288 over those 12 days, leaving 192 on hand when the delivery lands: eight days of cover still underneath it at the moment of arrival. That is delivery slack — how late an order you have already placed can run before you are empty.

K‑310 has nothing on order, so it has no delivery slack to measure. Its 45-day lead time measures something else: order today and the goods arrive on day 45 against cover that runs out on day 50, leaving five days of cover at arrival. That is a latest safe order date, and it is a deadline on your desk rather than a tolerance you hold over a supplier.

Eight and five are close in size and are not the same kind of number. J‑220’s eight days are tolerance you already own, sitting on a supplier who has given you a date. K‑310’s five days are a countdown you are running yourself, and they assume the 45-day lead time holds and that you are content to receive goods with five days of cover left underneath them.

So the ranking inverts — but say what it inverts on. On shelf days, K‑310 is the comfortable one. On how close it is to the point where ordering no longer fixes the problem, K‑310 is the one in front of you. Whether that means order today is a decision the arithmetic cannot make: five days of cover at arrival is the zero-buffer version, and if you want a fortnight on the shelf when the goods land, 50 days of cover less a 45-day lead time less a 14-day buffer puts the order date nine days behind you already.

Both numbers are worth having on the same screen, as long as they are labelled apart. Delivery slack tells you which inbound orders to chase. The latest safe order date tells you which SKUs to raise orders for. Run them under one heading and you lose track of which of the two you are looking at.

SKU Compass makes this a choice rather than a default. The Counting stock setting decides whether days of stock counts units already ordered on open POs, with including them or counting on-hand only as the two options. Its help text states the trade-off: including them assumes those POs arrive as expected, on-hand only is stricter and tells you to order sooner, and the same SKU can look comfortable on one setting and urgent on the other.

Neither setting is the true one. They are two scenario views. On-hand only asks what happens if nothing arrives. Including open orders asks what happens if everything lands on the date and in the quantity it claims, which is an assumption rather than a fact: orders get short-shipped, dates slip, suppliers cancel. The gap between the two answers on a SKU is the size of the bet you are placing on your open book.

The balance owed has a shape, but it is not a payment schedule

Start with what the next table is not. The expected-delivery date is not the payment date. If a supplier’s terms run net 30 from receipt, that line lands 30 days to the right of where the ladder puts it; if a deposit falls due on order, part of it went out long before. A cash plan runs on contractual due dates, and if you have those on file, build the ladder on them instead. Cash is one of the three headline readings, and the version of it a finance conversation can actually use is balance owed by payment due date.

What an open-PO report generally hands you is delivery dates, so that is what this ladder uses. Sequenced that way it is a stock-arrival ladder with the money attached: good for seeing where commitments cluster, not for telling anyone when the account gets debited.

The total was $80,850 either way. The ladder says what the total could not: $31,410 of it sits behind deliveries expected inside the first two weeks, and the last $18,000 sits on a single order at the end of October. Shift every line right by its own terms and those dates change. Whether the clustering survives the shift depends on the terms themselves, which is the argument for building the second version rather than letting the first stand in for it.

Where the open book distorts the position

All three readings are only as true as the state of the orders behind them. Four failure modes distort the open position in predictable ways, and it is worth knowing which direction each one pushes.

Orders closed in reality and open in the system. A cancellation agreed in an email, a supplier that quietly stopped producing, a sample order nobody closed. This one can inflate units and open value, and may inflate balance owed depending on payment status.

Receipts booked late. Stock that physically arrived but has not been received against its order is missing from on-hand while still appearing as inbound. Two errors in opposite directions, and they do not cancel — you plan as though goods already in the building were still in transit.

Lines with a quantity and no agreed price. These add to units still to receive and nothing to the money columns, so the unit figure looks right while open order value runs low. Not a defect in itself, but the two are then computed over different sets of lines. If they have to reconcile, find those lines first.

Expected dates entered once and left. The ladder and the date arithmetic both run entirely on expected dates. A date that was optimistic when the order was raised and was never revisited makes both look calmer than the situation is. This one degrades quietly, because nothing about the number changes to signal it has gone stale.

One thing sits behind all of this and is deliberately out of scope here. Approval determines whether an order may proceed; where that gate belongs relative to the actual commitment point is covered in the previous article on approval.

What to do with the three numbers

Before raising the next order, read the open book three times, and name the basis each time.

1. Units still to receive, against the SKUs you were about to order. Some are already covered, and the cover is dated.
2. Open order value, against the size of the order you are contemplating. The value of goods still to come, not the total of every order still open.
3. Balance owed, laddered by the best date basis you have—payment due dates for cash planning, delivery dates only as a stock-arrival proxy. Not to decide whether to order, and never on both date bases at once.

If a report hands you one figure and calls it your commitment, the useful question is not whether the figure is right. It is which of the four it is.