Inventory Management Without Losing Sales: The Minimal System That Works

There are two ways to lose money on inventory and both are silent. The first: your best seller runs out on a Friday and nobody notices until Monday — three days of sales gifted to the competition. The second: you bought 200 units of something that turns at 5 a month — capital buried in the stockroom that you can't use to buy what actually sells.

Inventory management sounds like a big-company problem, but it's the reverse: the smaller the business, the more every stockout and every immobilised dollar hurts. The good news is that 80% of the benefit comes from a system that fits in a well-kept spreadsheet. This article builds it.

The two mistakes, and which one is costing you more

Quick diagnosis: look at your ten best sellers over the last three months. How many days were they sold out? Every day is measurable lost revenue. Now look at the stockroom: what share of the inventory didn't sell a single unit last month? That's your dead capital. Almost always one of the two numbers scares you more — start there.

The reorder point: the only formula you need

For a small store, the whole apparatus of inventory management reduces to answering one question per product: at what quantity do I need to reorder? The formula is simple:

• Daily sales come from data, not memory: units sold in 90 days ÷ 90. Memory always exaggerates the good days.
• Replenishment days are real, from confirmed order to goods on the shelf — including the time you take to place the order.
• The buffer depends on the supplier: 30% for the reliable one, 50% or more for the one who "sometimes runs late". Imported goods: add the customs margin.
• Only for the products that matter. Applying this to the 20 products that make 80% of your sales already captures nearly all the value.

Not every product deserves the same attention

The classic operational mistake is treating the whole catalogue the same. ABC classification organises the effort: a few products concentrate the sales, and those are the ones that can never be missing.

• A — the ones that pay the bills (the ~20% of the catalogue making ~70-80% of sales): calculated reorder point, stock reviewed weekly, stockouts unacceptable.
• B — the complements (~30% of the catalogue, ~15-25% of sales): approximate reorder point, fortnightly or monthly review.
• C — the long tail (~50% of the catalogue, ~5-10% of sales): replenished when they run out, or not at all. Clearance candidates if they don't turn.

This classification also disciplines purchasing: the supplier's volume discount only makes sense on A products. Buying 100 units of a C product because "it worked out cheaper" is the overstock factory.

The sold-out product: losing the sale without losing the customer

The stockout will happen eventually anyway. The difference between losing a sale and losing a customer lies in how the store presents it:

• Sold-out products stay visible, not deleted. The page keeps ranking on Google and keeps informing the customer.
• With a "notify me" button: it captures the email and turns the stockout into a deferred sale — and tells you how much demand you're missing.
• With an estimated date if you know it: "available from the 15th" retains whoever can wait.
• With a visible alternative: "meanwhile, look at these similar ones" recovers part of today's sale.
• Never sell what you don't have without saying so: taking the order and announcing "3 weeks' delay" after payment is the most avoidable negative review there is.

Counts: because the system lies

System stock and stockroom stock drift apart on their own: packing errors, breakage, theft, badly re-entered returns. If you never count, the error accumulates until you sell something online that doesn't exist.

1. 01Cycle counts, not annual ones: every week you count a small slice (this month's A products, one shelf, one category). Twenty minutes a week avoids the two-day full inventory.
2. 02A products get counted monthly; B quarterly; C once or twice a year.
3. 03Every discrepancy gets fixed on the spot and its probable cause noted. Three identical discrepancies signal a broken process, not bad luck.
4. 04Returns have their own process: receive, inspect, re-enter or write off — the limbo of "boxes to check" is ghost stock.

Several channels, one stock

If you sell through the online store, the physical shop and WhatsApp at once, the danger is triple: selling the same unit twice. The rule is a single source of truth — one system (even the spreadsheet at first) that every channel reads from and deducts against.

• The shop deducts immediately, not "at the end of the day". The 11 am counter sale can collide with the 11:30 online sale.
• WhatsApp sales get recorded like all the others. The informal channel is where the mismatches are born.
• Reserve stock for paid orders, not carts: an abandoned cart can't be allowed to block inventory.
• When the spreadsheet starts failing — two people editing, crossed sales — that's the moment for a system connected to the store. How to evaluate that integration is in integrations and APIs for your business.

Clear without fear: dead stock doesn't improve with age

The product that didn't turn in six months won't turn in the seventh. Every month you keep it costs space, ties up capital and — in fashion or tech — loses value. Clearing isn't failing: it's converting a past mistake into cash to buy what does sell.

• Define the rule before you need it: no sales in X months → stepped discount → bundle with an A product → donation or write-off.
• The deep, short discount works better than the eternal 10%: it genuinely clears, and it doesn't train customers to wait for sales.
• The bundle rescues margin: the C product free or discounted alongside an A product raises the ticket and cleans the stockroom at once.
• Record what you cleared and why you bought it: the list of past clearances is your best filter for future purchases.