Stop reordering dead stock: the six-factor buy gate for every purchase order line

Dead stock is usually created on the second order, not the first. The first buy is a small hypothesis; the reorder is a habit, and habits don't check velocity. The fix is a gate: six factors — sell-through, age, margin, MOQ, lead time, launch risk — each with a threshold you set on a calm day, applied line by line before the PO goes out. Most routine POs have one to three lines that fail one factor and pass the other five, which is exactly the pattern that turns a reasonable order book into eighteen months of shelf occupancy.

Why the reorder is the expensive decision

Every SKU you own was bought twice: once as a test, once as a commitment. The test is cheap to be wrong about — you bought 40 units, they didn't move, you learned something. The reorder is where the cost lands, because it multiplies a position you already have rather than starting a new one.

And it gets made under pressure that has nothing to do with demand:

• Case packs and minimums make a full case feel like the natural order quantity.
• Price breaks reward spending more per transaction.
• Supplier reps resend the previous order sheet because it's easier than asking what changed.
• The completeness instinct says a six-colorway line needs all six to look like a real assortment.
• Stockout fear from the one SKU that ran dry in July makes every line feel under-bought in August.

None of those is a demand signal. The gate's job is to outrank them for sixty seconds per line, using numbers that came from your store rather than your supplier's catalog.

The six-factor gate

Pass/fail logic: factors 1, 2, and 5 are gating — fail any and the line is NO-GO. Factors 3, 4, and 6 are moderators — they can reduce the quantity rather than kill the line. That distinction matters, because a lot of good lines fail a naive all-or-nothing checklist and get skipped for the wrong reason.

What the gate looks like on real lines

All figures illustrative. A phone-accessory seller is about to send a PO with two lines on it.

PS-BLK, matte black. Sell-through 62% over 30 days, 9 units/week on in-stock days. On hand 22 → about 2.4 weeks of cover. Oldest stock 3 weeks. MOQ 100 at $3.80; 100 units clears in about 11 weeks at current velocity. Margin $8.20/unit. Projected cover after the PO: 122 ÷ 9 ≈ 13.6 weeks, inside a 16-week ceiling.

Verdict: passes 1, 2, 5 outright; 4 is borderline-acceptable; 3 is comfortable. GO at full quantity — this is what a healthy reorder looks like, and it should be fast to clear.

PS-SGE, sage green. Sold 6 units in the last 90 days ≈ 0.5/week. On hand 41 → 82 weeks of cover, roughly nineteen months. Oldest stock 8 months. The case pack would add 50 units, taking cover past 180 weeks. Margin is identical to black.

Verdict: fails factor 1 (borderline at best), fails factor 2 hard, fails factor 5 catastrophically. NO-GO — and it took forty seconds to see. The cash view makes it concrete: 41 units × $3.80 = $156 already parked, and the reflex reorder would have put another $190 in front of it.

That is the whole mechanism. Same product, same margin, same supplier, same day — one line is a good buy and one is a slow-motion write-off, and the only difference is two numbers nobody looked at.

Sell-through is the one everyone gets wrong

Factor 1 deserves its own section because the number most stores use is biased.

Count only in-stock days. A SKU that sat dark for two weeks after a promotion looks slow on calendar math. A genuinely slow SKU that never went dark looks identical. Those two cases need opposite decisions — one is a replenishment problem, the other a demand problem — and calendar-based velocity can't separate them. The correction is sales velocity measured during out-of-stock days, and it matters most on exactly the borderline lines this gate is judging.

Net out returns. Shopify's quantity-sold figures are gross. In apparel and home goods, a 20% return rate turns an apparent 5 units/week into 4, and that difference decides borderline NO-GOs.

Use the right window for the category. Thirty days is fine for fast consumables and useless for a $200 item that sells twice a quarter. For slow-turning categories, use 90 days and set the threshold accordingly — otherwise the gate eliminates your entire high-ticket range and you'll stop trusting it.

The MOQ trap, worked

MOQs fail lines that pass everything else, and the failure is arithmetic rather than judgment. A line has 6 units/week of velocity, a case pack of 50, and $2.10 of landed cost per unit. The case is $105 of cash and 50 ÷ 6 ≈ 8.3 weeks of demand — fine. Now the same line with a MOQ of 250: $525 of cash and 41 weeks of demand, which is most of a year's stock purchased to save a phone call.

The test isn't "can I afford the MOQ" but "does one MOQ clear inside 90 days at current velocity?" If not, three options exist, in order: negotiate a smaller first run (suppliers will often do this for a new account or a mixed order), buy the MOQ once and accept a longer cover on a line you believe in, or skip the line. What must not happen is filling the case with units you already know are slow in order to reach a total the supplier prefers — that converts a supplier's convenience into your inventory.

Lead time matters in the same direction. A 45-day lead time with a 14-week cover ceiling means the reorder lands when cover is already high; a 120-day lead time means you're ordering against a season that hasn't started, and the ceiling needs to be read against the season rather than against today's velocity. The sizing logic for both is worked through in how much inventory to order.

Launch risk: the new-SKU version of the gate

New products can't pass factors 1 and 2 — there's no history — so they get a different test:

Two high-risk signals don't mean don't launch — they mean launch smaller. A first order sized to 8–12 weeks of plausible demand, rather than to a freight threshold or a supplier's case pack, keeps the cost of being wrong low enough that you can afford to test three things instead of betting on one.

Running the gate without it becoming a ritual

Three practical notes:

Run it at the PO screen, not in a separate review. A gate that lives in a spreadsheet nobody opens on ordering day is decoration. The six factors belong as columns on the PO template you already use, so failing a line is a visible act rather than a silent omission.

Exempt deliberately, not accidentally. New SKUs under 60 days get a pass on factors 1 and 2 by design. Seasonal lines get judged against season-adjusted velocity, and a winter item failing in July is a "not this PO" verdict rather than a lifetime ban. Everything else runs the gate every time.

Expect lines to flip. A NO-GO isn't permanent. Velocity, cover, and age all move, so re-run the gate on every PO — some NO-GOs become GO again in Q4, and treating the list as a blacklist loses that.

What happens to a NO-GO line

Nothing dramatic. The SKU keeps selling until it's gone, nothing new stacks behind the old stock, and the position winds down on its own. What happens next belongs to the clearance routine: the age data that fed factor 2 comes from your aging report, and stock that ages past 180 days with weak velocity graduates to the markdown-or-exit decisions in dead stock on Shopify. The gate is the upstream half — the cheapest place to stop dead stock is before it's purchased.

Common questions

How do I know if a product is dead stock before I reorder it? Check sell-through over the last 30 in-stock days, the age of the stock you already hold, and projected weeks of cover after the order. If a line is selling a fraction of a unit per week while you hold more than about 20 weeks of cover, it's effectively dead regardless of the label — and adding units makes the position worse, not better.

Should I reorder a product that sold well last year? Yes, if the season supports it — compare against last year's same-weeks curve rather than against trailing off-season numbers. A seasonal SKU that fails the gate out of season isn't a permanent verdict; revisit it as its window approaches, using seasonal velocity rather than today's.

What if the supplier's minimum order is larger than I need? Test whether one MOQ clears inside 90 days at current velocity. If not: negotiate a smaller first run, accept a longer cover once on a line you believe in, or skip it. Don't fill a case pack with slow units to reach a total — that's the most common way a routine reorder becomes long-term stock.

How much inventory should I hold before reordering? Reorder when weeks of cover falls to your lead time plus a buffer, and cap the top end with the gate's ceiling — 12–16 weeks of projected cover is a workable default for ordinary goods, tighter for trend items, wider for staples with flat demand. Below the floor you risk gaps; above the ceiling you're buying storage rather than sales.

Is it better to order smaller quantities more often? For lines near the gate's edge, yes — smaller, more frequent buys keep cover honest and reduce the cost of being wrong. For proven staples with stable demand, larger buys save ordering effort and sometimes unit cost. The gate handles both because it judges cover after the order, whatever the quantity.

Next step: run the six factors retroactively on your last PO and count the NO-GOs. Multiply that count by the cash each line committed, and you have the value of sixty seconds per line — then add the six columns to your PO template so the next one gets checked before it's sent.