Inventory management when the spreadsheet breaks
A spreadsheet stops being enough at one identifiable moment. The four things that decide it, what a stock-out really costs, and what to count how often.
The short version
- A spreadsheet fails at a specific moment: the first time one unit of stock is counted in two places and the counts disagree. Product count is not the trigger — channels, lead times and variants are.
- At a healthy gross margin, a stock-out is expensive per unit and an overstock is expensive per month, so the right caution depends on whether cash or sales is your tighter constraint.
- Record a lead time and a reorder point per product. Average daily sales times lead time in days, plus a week or two of buffer, is more useful than any dashboard.
- Count a slice of your stock weekly rather than all of it yearly. No system is more accurate than your last physical count.
You sold something you did not have. The confirmation email went out at 9:14 in the morning, the customer is delighted, and the shelf where that unit should be has been empty since last Thursday. Now you get to write the apology.
That moment is what most people are actually searching for when they type inventory management for small business into Google. The question underneath it is usually "do I need software?" — and that is the wrong question to open with.
Here is the honest answer. A spreadsheet is not a bad inventory system. It stops being enough at one identifiable moment: the first time a single unit of stock is counted in more than one place and the two counts disagree. Before that moment, software mostly adds a second thing you have to keep up to date. After that moment, no amount of personal discipline will save you, because the problem is no longer effort.
The rest of this is how to tell which side of that line you are standing on, and what to do in either case.
The four things that actually decide it
Product count is the metric everyone reaches for and it is close to useless. A store with 400 identical prints and one fulfillment location is easier to run than a store with 12 products in four sizes sold on a website, at a market stall and through two wholesale accounts.
These four questions decide it instead.
1. In how many places does a unit get counted?
One website, one shelf, one person: a spreadsheet is fine, possibly forever. Add a marketplace listing, a market stall, a retail stockist or an Instagram DM sale and every single one of those becomes a place where a unit can leave without the spreadsheet hearing about it.
Two channels is the threshold, not twenty. The arithmetic of being wrong does not care how small you are.
2. How long does it take to get more?
If you can make or reorder a unit in two days, running out is an inconvenience. If your supplier takes six weeks and ships by sea, running out is a quarter of lost sales on that product and no way to shorten it.
Long lead times are the single strongest argument for a real system, because the decision you need help with is not "how many do I have" but "what do I need to order today so I do not run out in week five".
3. How many variants, not how many products
A shirt in five sizes and three colors is fifteen things to count, not one. Variant count is the number that actually breaks manual tracking, and it is the number people leave out when they tell you their catalog is small.
4. Who else touches stock?
The moment a second person can take a unit off a shelf — a partner, a weekend assistant, a family member helping at an event — your mental model of what you own stops being accurate. Not because they are careless, but because the count now lives in two heads and one sheet.
What a stock-out and an overstock actually cost
Most advice treats these as equally bad. They are not, and which one is worse for you depends almost entirely on your gross margin and your shelf life.
All figures below are illustrative — they are there so you can redo the arithmetic with your own numbers, not because they describe any real store.
The cost of a stock-out. Say an item sells for $40 and costs you $16 landed, so each sale earns $24 gross. It sells about 20 units a month, which is roughly 0.7 a day. You are out of stock for nine days. That is about six units, or $144 of gross profit you did not earn. Then add the part you cannot calculate: some share of those six customers bought the thing from somebody else and are now that store's customer, not yours.
The cost of an overstock. Same item. You order 200 units to get the better unit price. That is $3,200 of cash on a shelf and ten months of cover at your current sales rate. If the last 50 units go stale and you clear them at half price, you take $20 each instead of $40 — you still cover the $16 cost, so the loss is the $20 of margin you gave up, about $1,000, plus ten months of not having $3,200 available for anything else.
The useful conclusion from that pair of numbers is not "overstock is worse". It is this: at a 60% gross margin, a stock-out is expensive per unit and an overstock is expensive per month. So the right level of caution depends on which resource you are shorter of.
| Your situation | Which error hurts more | Practical bias |
|---|---|---|
| High margin, long shelf life, slow supplier | Stock-out | Hold more cover than feels comfortable |
| Low margin, fast reorder | Overstock | Order little and often |
| Perishable, dated or seasonal | Overstock, badly | Under-order and accept selling out |
| Made to order | Neither — it is a lead-time problem | Manage your calendar, not your shelf |
| Cash is your binding constraint | Overstock | Narrower range, deeper on the proven items |
Match your situation to an answer
| Where you are | What to use | What to stop doing |
|---|---|---|
| One channel, under ~50 variants, you are the only person | A spreadsheet with a reorder-point column | Counting from memory |
| One channel, 50–300 variants | Your store platform's own stock tracking, as the single source of truth | Keeping a parallel spreadsheet "just in case" |
| Two or more channels | A system where every channel reads one stock number | Reconciling channels by hand each evening |
| Any channel count, 6-week lead times | The above, plus a written reorder point and lead time per product | Reordering when you notice the gap |
| Staff, multiple locations, or wholesale | A system with per-location counts and an audit trail | Trusting one person's mental total |
| Pre-launch, no orders yet | A spreadsheet. Genuinely. | Buying inventory software before you have inventory |
The pattern in that table is that the jump is triggered by channels and lead times, not size. A store doing $2,000 a month across three channels needs more system than a store doing $20,000 a month through one.
What to count, and how often
Nobody with a job should be counting their whole catalog every month. The technique that works at small scale is cycle counting: you count a slice of your stock often, instead of all of it rarely.
- Weekly, five minutes: your top ten sellers by unit volume. These are where errors accumulate fastest and where a stock-out costs the most.
- Monthly: anything you have sold or received since the last count, plus any product whose recorded number you have personally doubted.
- Quarterly: everything, including the box in the hallway. This is the count that finds the shrinkage, the damaged units and the returns that never went back into stock.
- Immediately, always: anything you discover is wrong. The expensive habit is noticing a discrepancy and planning to fix it later.
Two things to record alongside the quantity, because they are what turn a count into a decision: the lead time for that product in days, and the reorder point — the stock level at which you place the order. A workable reorder point is your average daily sales multiplied by your lead time in days, plus a buffer of a week or two of sales for the supplier being late.
For the $40 item above: 0.7 a day, 42-day lead time, so 29 units, plus ten days of buffer, so order when you hit about 36. That number is more valuable than any dashboard.
Three cases where the obvious answer is wrong
"I only sell on one website, so I am fine." Not if you also sell at events, or ship from two places, or let people order through DMs. Count the ways a unit can leave the building, not the number of websites.
"I have hundreds of products, so I clearly need software." If every one of them is a print-on-demand or digital item, you have no inventory at all and a stock system is a solution to a problem you do not have. Check whether you actually hold units before you buy anything to count them with.
"Software will stop this happening." It will stop arithmetic errors and channel conflicts. It will not stop the unit that fell behind the shelf, the return that was never processed, or the five you gave away as samples. Every system is only as accurate as the last physical count, which is why cycle counting is not optional at any scale.
What inventory software will not fix
It will not tell you what to buy. Forecasting on a small store is mostly judgment, because the sample is tiny — twenty sales a month cannot distinguish a trend from a quiet fortnight, and any tool that claims it can is reading noise.
It will not fix a supplier who is unreliable about dates. A reorder point accounts for lead time; it cannot account for a lead time that is secretly random.
It will not reconcile your stock with your accounts. Inventory value, cost of goods sold and shrinkage are bookkeeping questions, and a stock count is an input to them, not an answer.
And it will not make you do the count. That remains the whole job.
Where this sits in mAutomate
mAutomate includes inventory, sales and analytics as part of its business operations, so the stock number behind your storefront is the same number you manage — there is no second system to reconcile. On top of that, the AI drafts and recommends, and nothing goes live until you approve it, which matters here specifically: a reorder suggestion you have not read is just a purchase you did not decide to make.
How to price a product covers the margin side of the same arithmetic, and the plans page lists what each tier includes.
Questions people ask
When does a small store actually need inventory software?
The practical trigger is selling the same stock through two or more channels, or working with supplier lead times long enough that you must order before you run low. Until one of those is true, a spreadsheet with a reorder-point column is adequate. Product count matters far less than people assume — hundreds of single-variant items are easier to track than twenty items in fifteen variants each.
How often should I do a physical stock count?
Count your ten fastest-moving items weekly, anything you have received or sold monthly, and the entire catalog once a quarter. This is called cycle counting, and it works because errors concentrate in fast-moving stock. A full count once a year tells you how wrong you were, not how wrong you are.
Is it worse to run out of stock or to hold too much?
They fail differently. Running out costs you the gross profit on each sale you could not make, plus some customers permanently. Holding too much ties up cash every month and often ends in a markdown. High-margin, long-shelf-life goods justify holding more cover; perishable, seasonal or low-margin goods justify ordering little and often.
How do I calculate a reorder point?
Multiply your average daily sales by your supplier's lead time in days, then add a safety buffer of roughly one to two weeks of sales for late deliveries. An item selling 0.7 units a day with a 42-day lead time needs about 29 units of cover plus buffer, so you reorder around 36 units. Recalculate when either sales or lead times shift.
Can inventory software predict what I should order?
Not reliably at small volumes. Forecasting needs enough sales history to separate a real trend from a quiet month, and a product selling twenty units a month does not provide it. Treat any suggested order quantity as a starting point for your own judgment about supplier reliability, cash and shelf life.
- Inventory
- Operations
- Running a store
Ready to start your store?
Let the AI build, market, and run it — you ship the product.
Get started free