How often should you actually count? A cadence decision for 100-5,000 SKUs
How often should you cycle count inventory? A weekly-to-quarterly cadence for small Shopify stores, sized to the hours a one-person team actually has.
InventoryX is in development and not available today. ClearStock is the live Shopify app.
Short answer: count your fastest, highest-value SKUs weekly, the middle tier monthly, and the long tail once or twice a year — and if you can only hold onto one rule, count 20–30 SKUs every single week. Frequency isn't about being fair to all your SKUs. It's a risk-and-hours decision: put the counting hours where phantom units do damage, and accept that the tail gets less of you.
Why frequency is a decision, not a habit
Your recorded quantities drift from the shelf constantly — receiving miscounts, damage nobody logged, returns that never restock, POS slips. Drift is also unevenly distributed. Three phantom units on your best-selling hiking sock is a reorder you never place and a listing that sells stock the shelf doesn't have. Three phantom units on a stationery variant that moves two a month is noise. Counting the whole catalog at one frequency wastes hours on the noise and starves the risk.
The expensive kind of drift is the sellable-looking kind: Shopify shows units available, the shelf is empty. When a shopper hits a genuinely empty shelf, they don't all wait — some buy the item somewhere else, some substitute, and some leave with the purchase unspent. Counting exists to catch drift before it turns into that. (Where phantom stock comes from in the first place is its own topic.)
What sets the frequency: three factors
Two numbers and one judgment call decide any SKU's cadence:
- Velocity — units per day. Fast movers burn through any discrepancy faster.
- Value — unit cost or retail. A variance on a $2 pen is a rounding error; the same variance on a $40 item is money.
- Loss exposure — the judgment call. Small, pocketable, high-demand items drift fastest: the SKUs that walk, get "borrowed," or get adjusted carelessly. These earn a higher frequency than their velocity alone justifies.
Put together, a workable tier table for a store in the 100–5,000 SKU range:
| Tier | Example SKUs | Cadence |
|---|---|---|
| Critical | Best-selling hiking socks, coffee bean bags, anything with 10+ units/day | Weekly |
| Middle | Enamel pin sets, phone stands, steady mid-price items | Monthly |
| Tail | Slow variants, seasonal leftovers | Once or twice a year |
| Watchlist | Pocketable high-value items, anything with a recent variance, new-staff territory | Weekly, regardless of velocity |
The capacity check that actually decides it
Every cadence plan dies on the same rock: hours. Run the math before you commit to anything (illustrative numbers — 1,200 SKUs, one person, a realistic shelf-to-screen pace of about 20 SKUs per hour in a stockroom):
The naive plan — count everything monthly. 1,200 ÷ 20 = 60 hours a month, which is 14 hours a week. For a solo operator, that plan is dead on arrival, and a dead plan gets no counting done at all.
The tiered plan. 12 critical SKUs weekly (0.6 h), 48 middle SKUs monthly (2.4 h ÷ 4.3 weeks ≈ 0.56 h/week), and the remaining 1,140 tail SKUs once a year (57 h, batched by collection into four seasonal half-days ≈ 1.1 h/week). Total: roughly 2.25 hours a week. That's a plan a person can keep.
And if even that doesn't fit? Cut the tail's frequency to every other year — half the tail each year — and protect the weekly tier with everything you have. Never the other way around. The bestsellers are where a missed count becomes a missed reorder; the tail is where a missed count becomes a shrug.
Tighten or loosen: two rules
Tighten the cadence when any of these happen: a count finds a variance (recount that SKU's neighbors the same week), a new staff member starts, a new location or a POS channel goes live, a velocity step-change hits (ads, a press mention), or the same SKU shows up twice on the watchlist. For a stretch after any of these, double the affected tier's frequency, then settle back.
Loosen when a tier earns it: three consecutive clean counts moves that tier down one level — weekly becomes monthly, monthly becomes quarterly. The three-clean-counts rule keeps your hours pointed at today's risk instead of last year's.
Whether the whole routine is working is measurable: the store-wide accuracy rate gives your cadence a scoreboard, and a cadence that can't move that number is just calendar clutter.
What counting fixes — and what it doesn't
A count corrects the record; it doesn't find the cause. A variance that repeats on the same SKU, same shift, or same reason code needs the discrepancy investigation workflow — otherwise you're re-correcting the same leak forever.
And accuracy is upstream of everything else you compute. Reorder math multiplies your on-hand by velocity assumptions; if a chunk of SKUs are phantom, no formula downstream saves it. Native Shopify counts make the routine easier to run: you create a count for a location and chosen products, staff record quantities by scanner or manually, and a manager reviews and approves before any quantity changes (Shopify Help Center, "Stock counts," accessed November 2026) — the approval-before-change part matters when the person counting isn't the person who owns the numbers. Our own Buying List assumes your on-hand quantities are true; a counting cadence is what keeps that assumption honest.
Common questions
How often should a small store cycle count inventory? Weekly for bestsellers and pocketable high-value SKUs, monthly for the steady middle, and once or twice a year for the tail. If you adopt only one rule, count 20–30 SKUs every week without exception — a small, unbreakable weekly habit beats an ambitious calendar that collapses in week three.
Is monthly counting enough? For mid-tier SKUs, yes. For bestsellers, monthly counting leaves up to 30 days of drift exposure on exactly the SKUs where a phantom quantity blocks a reorder — and velocity does the multiplying. Monthly is a fine floor for the middle tier; it's a bad ceiling for the critical one.
How many SKUs should one counting session cover? Whatever fits in 30–60 minutes: at a typical pace of about 20 SKUs an hour (illustrative), that's 10–20 SKUs per session. Consistent small sessions sustain themselves; sporadic three-hour marathons get postponed, and postponed counts are how drift compounds.
Should I count more often before peak season? Yes — tighten every tier in the six to eight weeks before your peak, whether that's Q4 or your category's own spike. Pre-peak counts clear out drift while suppliers can still act on your reorder, and post-peak counts catch the returns-and-promo chaos. A weekly cadence in November is worth more than a monthly one all spring.
Do I have to close the store to count? Not for cycle counts — you count a selected subset of SKUs on the live floor, between picks or during a quiet hour, and Shopify's counts let you choose the products per session. Full wall-to-wall counts are the ones that usually demand closed doors, and that trade-off is the subject of the cycle count vs physical count comparison.
Next step
This Monday, count your ten fastest SKUs — thirty minutes, no permission needed. Keep the streak for three weeks before touching any other tier. When a tier goes three clean counts in a row, step it down and move those hours up the velocity ranking. And when you're ready to run the sessions properly — sheets, scanners, approval flow — the process itself is laid out in how to run cycle counts and stock takes on Shopify.