Inventory control for your shop: a practical guide

Nearly every shop I know in Guatemala keeps inventory one of two ways: in the owner’s head, or in a notebook only the owner can read. Both work, and work well, up to one exact point: the day you hire someone, the day you open the second till, or the day you realise you have spent three months pushing a product that leaves you a few cents a unit.

This article is not about software. It is about the practice: how you count, what you write down, what each discrepancy means and what you do about it. If you finish it and decide your notebook is enough, you will have decided well, with the facts in hand.

What you are actually measuring

It is not one number per product. It is three:

• Physical stock — how many units are really there, on the shelf and in the back.
• Book stock — how many your records say are there.
• Cost — what each one cost you, landed in your shop.

All of inventory control is the work of keeping physical and book stock glued together, and of knowing cost precisely enough to price. Most shops start with the stock figure in memory and the cost in a vague recollection of the last supplier invoice. That is where almost every later problem is born.

Four words you are going to need, plainly:

• Shrinkage — product that leaves without being sold.
• Reorder point — the stock level at which you have to order again.
• Turnover — how fast what you are holding sells.
• Adjustment — the correction you make when what you counted does not match what you recorded.

The physical count: how to make it worth doing

There are two ways to count, and they do not compete.

Full count. You close, you count everything, you adjust everything. Once or twice a year, usually at year end or before your busy season. It is heavy, it happens at night or on a Sunday, and it exists to reset the clock to zero.

Cycle count. You count a handful of products every week, rotating, so that over a year everything gets counted several times. This is what actually holds control together, because it finds the problem while it can still be traced.

If you are only going to do one of them, do the cycle count. And if you are going to cycle count, start with the products that move the most money, not the ones with the most units. Twenty products a week is a rhythm you can sustain. Eight hundred products at once is a rhythm you abandon on the second Saturday.

The rules that separate a count worth doing from one that only tires you out:

• Freeze what you are counting. With the shop closed, or at least without selling that product while you count it. If someone rings up a crate of milk halfway through, your number describes nothing.
• Count by location, not from a list. Walk the shelf and write down what you see. If you go in holding the system’s printout, your eye confirms the number it has already read. This error is more common than theft.
• Two pairs of eyes, or two passes. One person counts and another writes; or you count, then recount only the large discrepancies before adjusting.
• Fix the unit and never change it. Twelve bottles do not mean the same thing if you buy by the case and sell by the bottle. Pick the unit you sell in and always count in that one.
• Count what is set aside, lent out, damaged or on display too. It exists, and it has been paid for.
• Adjust the same day. A count adjusted eight days later describes a shop that no longer exists.
• Keep the previous count. One discrepancy says little. The same discrepancy three times running on the same product says everything.

Physical counts per branch, from your phone

Why inventory balances on paper and not in the back room

When the number does not come out, these are the causes, roughly in the order they show up in a small shop:

• A sale that was never recorded. The tab, the quick sale to a regular, the item that walked out while someone else was being served.
• The delivery did not match the invoice. You ordered 24, 22 arrived, someone signed for 24. From that moment your records and your stockroom live in different worlds.
• Mixed units. You buy by the case of 12, sell by the unit, and someone deducted one unit while thinking of one case.
• Returns that never made it back into the records. The customer returned it, the product went back on the shelf, nobody wrote it down.
• Product the business consumed itself. The sample, the one opened to demonstrate, the one given away with a promotion.
• Damage and spillage nobody logged. It broke, it got wet, it was crushed in the stack.
• Theft. From outside and from inside. It exists, and it has to be said.
• Counting error. The most frequent of the lot, and the last one people suspect.

One rule that will save you grief: the first time something does not add up, the likeliest suspect is the record, not the person. Rule out the process before you accuse anyone. And a discrepancy you never assigned a cause to is a discrepancy that will repeat next month, because you never fixed it.

Returns that restock the shelf and leave a credit note

Shrinkage: name it, or it becomes “it went missing”

Shrinkage is any product that left without generating a sale. Its value is not a curiosity: it is cost, and it comes out of your profit exactly as if someone had bought it and never paid you.

What turns shrinkage into useful information is the cause. These are the ones worth separating:

• Theft — sends you to review cameras, access and the till process.
• Damage — sends you to review how goods are stacked, moved and handled.
• Expiry — almost never a stockroom problem. It is a buying problem: you bought too much, or too early.
• Counting error — sends you to train whoever counts, or to change how counting is done.
• Supplier shortfall — sends you to review goods-in, or the supplier.
• Promotion or sample — not a loss, a marketing expense. Separate it or you will think you are being robbed.
• Returned to supplier — it left, but you will be paid or replaced.

Measure shrinkage in quetzales at cost, not in units. Fifty plastic bags and three units of an expensive cream are not the same problem, however much the unit count says they are.

In Cuadra, every stock movement that is not a sale is recorded with its cause, and the shrinkage report shows it grouped by cause and by branch, with the value lost in quetzales, the percentage it represents of cost of goods sold, and the products costing you the most.

The shrinkage report, by cause and by branch

Reorder point: the number that prevents “sorry, we’re out”

The reorder point is the level at which you have to order again so the goods arrive before you run out. The basic form:

Reorder point = (average daily sales × your supplier’s lead time in days) + safety stock

An example, with made-up numbers purely to show the mechanics: you sell 6 shampoos a day on average, your supplier takes 5 days to deliver, and you want 2 days of cushion. That is 6 × 5 = 30, plus 6 × 2 = 12. Your reorder point is 42. When you drop to 42, you order.

Safety stock is where the number is won or lost. It depends on two things: how much your daily sales vary, and how unpredictable your supplier is. A product that sells 6 every single day needs little cushion. One that sells 0 for three days and 25 on Saturday needs far more, even though the average is the same.

The formal version of that idea uses the standard deviation of daily sales rather than the bare average. Cuadra’s purchase-suggestion engine computes the mean and standard deviation of your daily sales and builds the cushion from 1.65 standard deviations times the square root of lead days plus safety days — roughly a 95% service level. You do not need to do that arithmetic by hand to understand what it says: the jumpier a product’s sales, the more of it you have to hold to avoid letting a customer down.

Two Guatemalan caveats:

• Lead times are not the same all year. December is not March, and a supplier driving up from the capital to the interior does not take the same time during the fair season. Revisit the number by season instead of fixing it once.
• The reorder point does not apply to fashion or seasonal stock. There you do not want to replenish, you want to sell what you bought and close it out. Those are different decisions.

Average cost or last cost: which one, and what for

You bought 10 units at Q8.00 and later 10 at Q11.00. You sell one. What did it cost you?

• Last cost: Q11.00. Easy to maintain, and it reflects what replacing the item costs today. Its flaw is that your profit rises and falls with every supplier price change, even when you did nothing differently.
• Weighted average cost: Q9.50. That is (10 × Q8.00 + 10 × Q11.00) ÷ 20. It smooths the jumps and better describes what the stock you are holding cost you on average.
• FIFO, first in first out: Q8.00 until the first batch runs out, then Q11.00. The most faithful to physical reality and the heaviest to keep by hand, because it demands tracking every batch.

For a shop, weighted average is almost always the right answer. But keep last cost in mind for something different: pricing. Price looks forward, at what replacing the item will cost; profit looks backward, at what the goods you actually sold cost you. If you price off a stale average while your supplier has already raised you, you will be selling cheap without noticing.

Two details people forget when working out cost:

• Freight and haulage are part of the product’s cost, not a separate expense. If you bring goods from the port or the capital, spread it across what you brought.
• Free goods change the cost. If you get 13 for the price of 12, your real unit cost is the total divided by 13.

In Cuadra the cost we hold per product is a weighted average: every purchase you receive recalculates it using the new batch’s quantity and price, and sales do not move it.

Expired stock and dead stock

Two different problems, solved in different places.

Expired. Pull it, record it as shrinkage with an expiry cause, and note its value at cost. Then ask the question that matters: why did it get that far? If the same product expires on you twice, your reorder point is too high, or the pack size your supplier sells is too big for your turnover. The fix is not in the stockroom, it is in the purchase order.

What the stockroom does solve:

• What expires first goes at the front. New deliveries go behind. It sounds obvious and it is the rule most often broken.
• Label the date on receipt, not on sale.
• Check short-life products every fortnight. A product with 20 days left, found in time, still sells at a discount; found late, it is rubbish.

Dead. A product with no sales in 90 days is sleeping capital taking the space of something that does sell. Your options are to cut the price until it moves, bundle it with something that does sell, return it to the supplier if your terms allow, or give it away as a courtesy and record it as promotional shrinkage. The only bad option is leaving it there another year because it will sell someday.

The number that tells you whether this is happening to you is turnover: cost of goods sold in the period, divided by average inventory at cost. Compare it across your own categories before comparing it to any outside benchmark: normal turnover for groceries and for hardware look nothing alike.

In Cuadra, when you pull expired stock you record it as shrinkage with an expiry cause and the report tells you what it cost you. The system does not track each batch’s expiry date: that still lives on the product label and in your fortnightly check.

What to do if your inventory is already out

Do not try to reconstruct the past. It cannot be done, and the attempt is what makes people give up.

Week 1. Pick 20 products: the ones that move the most money, not the most units. Count them with the shop closed. That count is your starting line, not an accusation about what happened before.

Week 1, same day. Write half a page on how a purchase comes in, how a sale goes out, how damage gets logged and who is allowed to adjust. Stick it up in the stockroom. A process that lives only in your head is not a process.

Weeks 2, 3 and 4. Count those same 20 products every week. Write down the discrepancy, and write down the cause. When you do not know the cause, write “don’t know” — that is information, and you will want to know how many of them you have.

Day 30. Look at your causes. The one that repeats most is your real problem, and in my experience it is almost never the one you assumed at the start. Fix it before you widen the net.

Day 31 onward. Add the next 50 products.

Twenty products sustained for a year beat eight hundred products abandoned in week two. Every time.

And when the notebook stops being enough

The notebook is enough for as long as you are the only person recording, the only one buying and the only one counting. It stops being enough when any of these happens:

• Someone else sells, and you need to know what left without asking them.
• You have two branches and want to move goods between them without losing track.
• You want margin per product, not total sales.
• Discrepancies appear and you cannot trace when they started.

Cuadra is the point-of-sale and inventory system we build in Guatemala, and this is the inventory half of it, plainly: stock falls on its own with every sale at the counter; the physical count is done from the phone by picking a branch, seeing what the system says and typing what you counted, and it is recorded as a recount movement for every line that came out different; every non-sale movement carries its cause; the shrinkage report groups by cause and by branch in quetzales; every product can carry a reorder point, and purchase suggestions are built from your own sales velocity; and cost is a weighted average. The Pro plan is Q249.00 a month and there is a 14-day trial.

The point of sale that moves stock as you charge

How to choose a system for your shop in Guatemala

And if after all of this your conclusion is that counting 20 products every Saturday in the notebook solves your problem — do that. Inventory control is a practice before it is a program, and the practice is the part you cannot buy.

Frequently asked questions

• How often should I do a physical count in my shop? A full count once or twice a year, plus weekly cycle counts over a small, rotating group of products. The cycle counts are what actually keep inventory straight, because they find the problem while you can still trace where it came from.
• What counts as shrinkage, and how do I record it? Shrinkage is any product that leaves without being sold: theft, damage, expiry, samples, counting errors or supplier shortfalls. Always record it with its cause and value it at cost, in quetzales. Without a cause, shrinkage is a number that lets you fix nothing.
• How do I work out a product’s reorder point? Multiply your average daily sales by the days your supplier takes to deliver, then add safety stock. The more that product’s daily sales swing around, the bigger that cushion has to be.
• Should I use average cost or last cost? Weighted average cost to measure your profit, because it describes what the goods you sold actually cost you. Last cost as your reference when setting prices, because that is what it costs to replace the item today. The two figures answer different questions.
• My inventory does not add up. Am I being robbed? Possibly, but theft is the last suspect, not the first. The most frequent causes are unrecorded sales, deliveries that did not match the invoice, mixed units of measure and counting errors. Rule out the process before you accuse a person.
• Do I need a system to control inventory? Not at first. A notebook with cycle counts and written-down causes works for as long as you are the only person recording anything. A system starts to earn its keep when someone else sells, when you have more than one branch, or when you need margin per product and total sales are no longer enough.