Product
Ordering to a Budget: Spending on the SKUs With the Most at Stake
When cash won't cover the whole suggestion list, the question stops being "what should I reorder?" and becomes "what do I fund first?"
The suggestion list assumes you can afford all of it
Replenishment math is good at exactly one question: how much of this SKU should I order? Assisty's Smart Buying Table computes a suggested quantity per SKU from sales velocity, lead times, and incoming stock, then rounds it to your case sizes and caps it at your maximum stock levels. That is a per-line answer, and it is the right per-line answer.
Cash is not a per-line question. It is a single shared pool, and every line draws from it. The selection footer on the Buying Table page shows the shape of the problem in one string — its example reads "3 selected · 164 units · $1,276". Three good decisions, one number that has to clear the bank.
When that number exceeds what you are willing to spend this cycle, no amount of per-SKU precision helps, because no line knows about the others. You have crossed from a replenishment problem into an allocation problem — what retail has long called open-to-buy, the practice of setting a merchandise budget for a period and buying within it (Shopify).
The cut gets made either way. The only real choice is whether it is deliberate or accidental. Accidental cuts look like this: order the top of the list until the card stops working, push the rest to "next week," or trim every line by a flat percentage because that feels even-handed. The across-the-board trim is the worst of the fair-seeming options — it under-funds the lines that were about to run out and keeps funding the ones that weren't.
"Most at stake" is not "biggest number"
The instinct when a list is over budget is to sort it by something. Which something you pick decides where the money goes, and three of the tempting choices misallocate in predictable ways.
| Sort the over-budget list by… | What it optimizes for | Where it misallocates |
|---|---|---|
| Largest order value | Getting the most inventory value onto the shelf per decision | Funds whatever is expensive, not whatever is urgent — one slow-moving high-cost line can absorb the whole budget |
| Lowest stock on hand | Filling the emptiest bins first | Ignores demand rate and lead time — a slow seller at four units may still hold months of cover |
| Highest sales rank | Protecting the products you think of as the business | Rank is history, not the gap — a bestseller with stock already inbound needs nothing this cycle |
| Money at risk before the next order | The demand you would actually fail to serve if this line waits | Needs a margin and a next-order date per line; the boundary of the list still needs judgment |
Only the last row asks the question the budget is really posing: if I skip this line until the next buying cycle, what does skipping it cost? That is a forward-looking figure, and it is the one that makes lines comparable to each other.
Fund down the list until the money runs out
The cost of skipping a line has a standard shape: the units you fail to serve times the margin on each (Verve AI). Rewritten for a buying cycle, where the relevant window is the gap until your next order can land:
money at risk = expected unserved demand before the next order × margin per unit
funding priority = money at risk ÷ cost to cover it
The second line matters more than the first. Ranking by money at risk alone tells you which line hurts most; dividing by what it costs to fix tells you which line retires the most risk per budget dollar — the standard greedy approach to a constrained spend.
Line A is short 20 units at $18 margin — $360 at risk, $200 to cover. Line B is short 120 units at $4 margin — $480 at risk, $600 to cover. Line B carries more money at risk and would win a straight sort. But A returns $1.80 of retired risk per dollar spent and B returns $0.80, so with a tight budget A is funded first and buys back more exposure with less cash. (Illustrative numbers.)
Then fund straight down the list until the money is gone, and review the boundary — the two or three lines straddling the cutoff. Greedy funding is sound in the middle of a list and unreliable at its edge, because real order lines are only partly divisible: minimum order quantities, case packs, and freight thresholds mean a half-funded line is sometimes not orderable at all, and sometimes cheaper per unit if you push it over a break. Seasonal one-shots also break the ranking, because "defer to next cycle" isn't available to them.
Where a line is divisible, partial funding is usually better than a binary cut: buy the weeks of cover the budget affords rather than none.
A budget is a constraint, not a one-time edit
Budgets move. A supplier offers terms, a payment clears late, a category gets an extra allowance. If the cut is a one-time manual edit — rows deleted, quantities typed over — every move means rebuilding the plan from scratch, and the reasoning behind the first cut is gone.
That is why the constraint belongs in the plan rather than applied to it. On the Smart Buying Table, the ranking, the constraint, and the cut are all part of the same live document.
| Step | What Assisty does |
|---|---|
| Know what's at stake | Surfaces findings ranked by money at stake — items that run out before a reorder can land, capital tied up in overstock — each carrying its one-tap fix |
| Apply the constraint | State the budget in conversation and the whole table recomputes (the page's example: "I have $1,000 to spend") |
| Make the cut | Fit-to-budget defers the least-urgent items and shows exactly what got cut — the page's example: $1,276 → $844, "Deferred: Travel Tumbler ($432)" |
| Change your mind | Every change shows a was→now chip with per-row undo, and nothing is committed until you create the order |
| Commit | Select rows and Atlas splits the selection into one PO per supplier and asks which location receives it — or add them to the Purchase List |
Two properties do the real work here. The cut is re-runnable, so a changed budget is a restated number rather than a rebuilt spreadsheet. And it is reversible — the was→now chip and per-row undo mean the boundary review is something you can actually perform, overriding the ranking on the two lines where you disagree with it without losing the ranking everywhere else.
The output that outlives the cycle is the deferred list. A deferred item is not a cancelled item; it is a line whose money at risk has been growing since you last looked at it. Kept explicit — as the fit-to-budget cut keeps it — it becomes the opening agenda for the next buying cycle instead of a surprise stockout three weeks out.
An in-table data expert that answers "why 84?" on any row, computed from the exact values in your grid, is listed as coming soon.