Apparel inventory planning has one rule that generic advice misses: the unit of planning is the variant, not the product. Your "Pima Tee" doesn't stock out — the M / White does, while XS / Olive gathers dust. Plan at the style level and you'll simultaneously overstock slow variants and lose sales on the popular ones. Everything else — size curves, broken runs, case packs, long mill lead times — follows from getting that unit right.
Why apparel is harder than the average Shopify catalog
- Variant explosion. 40 styles × 6 sizes × 4 colors is 960 SKUs, each with its own sales rate. No one maintains 960 reorder points in a spreadsheet.
- Size curves aren't flat. M and L might be half your tee sales while XS and XXL are single digits — and the curve differs by style and by color.
- Broken size runs quietly kill conversion. A product page missing its two most popular sizes still looks "in stock" in your admin totals, but to half your shoppers it's sold out.
- Suppliers sell in packs. Mills and factories quote MOQs and case packs — often prepacked size ratios — not the 13 loose units your math suggested.
- Lead times are long and mixed. A domestic knitter might restock in a week; overseas production is 60–90 days. One safety buffer can't fit both.
- Seasons flip the demand curve. Chino shorts in March and fleece in October don't sell like their annual averages.
Plan every variant like its own product
The mechanics are the same reorder-point math any store uses — daily sales × (lead time + safety days) — just applied per variant:
| Variant | Daily sales | Lead time + buffer | Reorder point | Available | Status |
|---|---|---|---|---|---|
| Pima Tee · M / White | 0.9 | 7 + 14 days | ≈ 19 | 11 | Reorder |
| Pima Tee · S / Olive | 0.3 | 7 + 14 days | ≈ 6 | 4 | Reorder |
| Chino Shorts · 32 / Charcoal | 0.5 | 45 + 14 days | ≈ 30 | 26 | Stockout risk |
| Chino Shorts · 36 / Charcoal | 0.1 | 45 + 14 days | ≈ 6 | 31 | Healthy |
Two things to notice. The shorts' reorder points dwarf the tees' — same math, longer lead time (an overseas supplier vs. a domestic knit mill). And the 36-waist is healthy at 31 units while the 32 is at risk at 26: identical product, different velocity. That's the size curve doing its work — and why the style-level total ("57 shorts in stock") tells you nothing.
Ordering: respect the pack, watch the curve
When you do reorder, two apparel-specific rules:
- Round to how the supplier sells. If Loomstate Mills ships that tee in case packs of 12, a suggested 33 units becomes 3 cases (36). Build the rounding into the plan, or you'll do case math by hand on every line of every PO.
- Reorder the curve, not the total. Refill each size to its own target rather than "topping up the style." This is what keeps runs unbroken — the M/L core deep, the tails shallow but present.
Multi-location, counts, and the rest of the loop
If you run retail and warehouse (or two shops), a size can be dead in one location and sold out in the other — transfers are often faster than a new PO, and they're free. And because apparel accuracy erodes fast (try-ons, mis-scans, returns to the wrong shelf), regular cycle counts on your top styles keep the "available" number the whole plan rests on honest.
This whole loop is what Stockwik runs for Shopify apparel merchants: velocity and reorder points per variant, statuses (Stockout risk / Reorder / Watch / Healthy) to surface what needs attention, POs that round to case packs and MOQs, receiving, transfers, and cycle counts — with seasonal peaks detected from each variant's own history.
FAQ
How do I stop breaking size runs?
Set reorder points per variant, sized to each variant's own velocity, and reorder when core sizes hit them — don't wait for the style total to look low. The style total is the last place a broken run shows up.
How much of each size should I order?
Order to each size's own target: daily sales × days of cover, minus what's available and incoming, rounded to the supplier's pack. Your sales history is your size curve — use it rather than a generic ratio.
What about brand-new styles with no history?
Borrow the size curve from your most comparable existing style, order conservatively, and re-order from real data once a few weeks of sales exist. New-variant velocity starts working as soon as there's history.
