.webp){kind=logo}
Key takeaways:
Stock control has a reputation for being simple, and it is, in the way that going to the gym is simple. The methods are not complicated. Doing them on the same day every week, when the delivery is late and someone called in sick, is the entire difficulty.
Most guides on this topic answer a different question. They explain classification schemes and ordering formulas, which belong to the planning layer, and skip the part where a person walks into a cooler with a clipboard on a Tuesday morning.
This guide covers the routine: what the week looks like, how the count is actually run and how long it takes, the delivery check, what to do when a supplier shorts you, the rules for dated product, and where an automated system genuinely helps against where it quietly lies to you.
The usual definition says stock control is the narrower physical piece and inventory management is the broader planning piece. That is true and it does not help anyone decide what to do on Monday.
A more useful cut is by ownership and rhythm. Stock control is work somebody performs on a schedule: counting, receiving, checking dates, rotating shelves, logging waste. It has a named owner, a time slot, and an output you can point at. Inventory management is work somebody decides: what to stock, at what level, from which supplier, at what cost.
The two fail differently, which is the practical test. Control failure looks like the walk-in holding nine cases when the sheet says twelve. Management failure looks like twelve cases of something nobody orders, counted perfectly.
Small operations often run both jobs through one person and never separate them, which works until the business grows enough that the counting stops happening. That breaking point arrives earlier than most owners expect, and handling small business inventory management well past it means giving the routine an owner who is not also doing the buying.
Stock control works when it is a fixed part of the week rather than a task that gets done when there is time, because there is never time.
Different categories need different frequencies, and counting everything weekly is how programs collapse under their own weight. Perishable, fast-moving items justify a weekly count because they turn over completely between counts and they spoil. Cleaning supplies do not.
| Category | Count frequency | Why |
|---|---|---|
| Fresh produce, dairy, protein | Weekly | Fast turnover, spoilage risk, largest share of food cost |
| Frozen and dry goods | Weekly or biweekly | Slower movement, but still a large share of value |
| Beverage and alcohol | Weekly | High value per unit and the highest shrinkage exposure |
| Paper, chemicals, cleaning | Monthly | Slow movement, low value, low variance |
| Smallwares, glassware, equipment | Quarterly | Losses are breakage, which shows up without a count |
Gordon Food Service recommends roughly this split, weekly for perishable and fast-moving items and monthly or quarterly for the rest. The pattern holds across operations of very different sizes because it follows the money and the spoilage rather than the item count. Building it into a repeating shift duty rather than a reminder is what makes it survive a busy month, which is the job a restaurant checklist system does.
The mechanics of counting get treated as obvious, and they are not. Two people counting the same walk-in with different methods will produce different numbers, which is why stock inventory control has to specify a method rather than an outcome.
Run it this way:
Step five is the one that separates a real count from a ritual. Working sheet to shelf means you only ever find items your sheet already lists, so anything mis-shelved, mislabeled, or ordered by someone else stays invisible indefinitely. Working shelf to sheet finds it.
Step three explains most of the difference in how long counts take. Operations that rotate and organize the night before consistently report far faster counts, because the two jobs have been separated instead of being done at once.
A count sheet ordered by storage zone rather than alphabetically does most of this work for you, since the sheet then matches the walk. Laying out a restaurant inventory list that way is a one-time job that saves time every week afterward.
No authoritative body publishes a benchmark for this, so treat any single figure with suspicion, the ones below included. They are offered to calibrate against, not to grade yourself on.
A two-person count against $11,000 to $14,000 of ending inventory typically runs about three hours each, so six labor hours a week, with invoice entry taking separate time afterward. Around $16,000 of stock can absorb six to seven hours when it is broken into three blocks covering dry storage, freezer and cooler separately.
The same $16,000 to $20,000 can also be counted solo in about three hours, where receiving was entered ahead of count day and the storeroom was already organized. A small pizza shop holding $2,500 to $3,000 sits closer to 25 to 45 minutes.
The spread across those cases is not explained by inventory value, and it is not explained by headcount. It is explained by whether the prep work happened first. Counts that run fast are counts where shelves were rotated and organized on an earlier shift, and counts that drag are counts doing both jobs at once.
That is worth costing out. Six labor hours a week is roughly 300 hours a year, so moving the tidying to a different shift and halving the count returns about 150 hours without buying anything.
Stock level control is the part of the routine that decays quietest. Whatever levels you set, they will be wrong within a season, and nothing in the system will tell you.
Levels drift for ordinary reasons, which is why stock level control has to be a scheduled review rather than a reaction. A menu changes and an item's usage halves. A supplier's lead time slips from two days to four and the buffer no longer covers it. A promotion triples demand for six weeks and nobody resets it afterward.
Each of these is invisible day to day, because the level still works well enough to avoid an obvious stockout while quietly holding too much or too little.
The routine fix is to review levels against the last quarter's actual usage once a season, rather than waiting for a stockout to reveal the problem. Pull usage per item, compare against the level currently set, and look for the two failure signatures: items that never dropped near their reorder point, which are over-stocked, and items that hit zero between deliveries, which are under-stocked.
Watch for one non-obvious case. An item that never triggers a reorder may not be over-stocked at all, and may instead be one somebody is topping up manually outside the system, which hides both the real usage and the real level.
Orders placed outside the system are precisely what a dedicated ordering channel removes, and VoiceOrder Solutions sits on the supplier's end of one. Field sales platforms such as SimplyDepo close the same gap from the other direction, capturing the order on a rep's tablet during the store visit rather than leaving it as a note somebody keys in later.
Software built for restaurant stock management will surface the first pattern automatically, and the second needs a person to notice.
The delivery check is the highest-value ten minutes in the stock control week, and it is routinely delegated to whoever happens to be near the door.
The standard is straightforward: confirm that the items and quantities being dropped off match the delivery invoice before signing anything, then have the receiver initial and date the invoice so there is a record of who accepted what. Gordon Food Service makes both points, and the second one matters more than it looks, because an unsigned dispute has no owner.
Checking after the truck leaves is a materially weaker position. Sign for a delivery you were not present to receive, find discolored product in the box an hour later, and you are arguing about a signature you already gave. The blunt version of the rule is that it pays to be the person who checks in your own orders.
Three things go wrong at receiving and each needs a different response. A short is a quantity that did not arrive, which should be marked on the invoice before signing.
A substitution is a different product sent in place of what was ordered, which you can refuse at the door and cannot easily refuse later. Damage or short-dated product is a quality rejection, which needs to be photographed at the moment of discovery.
Errors that originate upstream in the order itself are a different problem with a different fix, covered in our guide to reducing order errors.
Noticing a problem and being paid for it are separate achievements, and the second one has a process.
The method that works is unremarkable, and most of its power is in doing it the same day: photograph the defect, send the photo to your sales rep, and ask for a credit or a replacement on the next delivery. Independent operators generally report little trouble getting credits handled that way. Buyers on large chain accounts sometimes report the opposite, refused a credit despite ordering far more volume.
That contrast is the useful part. Credit outcomes track the strength of the rep relationship more than they track the size of the account or the wording of the policy, which means the person who talks to your rep should be someone who talks to them often.
Log every credit request and whether it was honored. A supplier that reliably issues credits is absorbing the cost of its own errors, which is what you are paying for. A supplier that quietly does not is charging you for product you threw away, and three months of a simple log makes the difference visible in a way that a single argument never will.
For anything ready to eat, stock rotation stops being an efficiency practice and becomes a regulated one.
The FDA's Food Code requires that ready-to-eat food requiring time and temperature control for safety, prepared in-house and held more than 24 hours, be date-marked with the day by which it must be consumed, sold or discarded. Held at 41 degrees Fahrenheit or below, that limit is seven days, and the day of preparation counts as day one. The purpose is controlling the growth of Listeria monocytogenes.
The Code does not mandate a particular label format. A calendar date, a day of the week, or a color-coded mark all satisfy it as long as the seven-day parameter is respected, which means you can pick whichever system your staff will actually use.
The operational consequence is that shelf rotation has to run on expiry rather than on receipt date. In practice that is a placement and labeling job rather than a policy one, because whoever reaches into the walk-in mid-service takes whatever is nearest and does not read dates.
So the case closest to its date has to be the case closest to their hand. Where a delivery turns up with less life left on it than the stock already in there, putting it behind by habit is how it quietly times out.
Capturing dates and lot codes at receiving is what makes any of this checkable later, and it is the same data a recall depends on, which our guide to food traceability software covers in depth.
Waste is the part of stock control that gets discussed constantly and measured almost never.
The scale is genuinely large. A USDA Economic Research Service study estimated that in 2010, 31% of the 430 billion pounds of food available at the retail and consumer levels went uneaten, or 133 billion pounds, worth about $161.6 billion at retail prices. That is 2010 data published in 2014, so treat it as an order-of-magnitude figure rather than a current one.
The practical instrument is unglamorous: a running log of every item lost, stolen, soiled or damaged, with its price, kept daily rather than reconstructed at month end. It takes seconds per entry and it does two things at once.
First, it converts an invisible cost into a number you can act on, and the pattern in that log usually points at a specific station, shift or item rather than at general carelessness. Second, discarded stock that is properly recorded is frequently deductible, and operations that throw away expired product without logging it as a bookkeeping loss are paying for it twice.
Dead stock is the slower version of the same problem: product that is not spoiling but is not moving either. It shows up as items appearing on every count at the same quantity, which is a pattern you only notice if you keep the old counts rather than overwriting them.
Across several sites the same item is often dead in one location and moving in another, so the fix for multi-site restaurants and food service operators is usually a transfer rather than a write-off.
At warehouse scale the routine changes character, because the limiting factor stops being the counting and becomes finding.
Three physical disciplines do most of the work. Every location gets a label and every item gets a code, so an item's position is a data point rather than institutional memory.
Incoming goods are checked, labeled and logged before they enter the storage area rather than after, since a pallet put away unlogged is functionally lost. And fast-moving items sit closest to the point of use, with slow movers pushed to the back, which shortens the walk on every single pick.
Bin accuracy is the metric that captures whether these are working. It asks not whether the total quantity is right but whether the quantity in each specific location is right, and it is always worse than total accuracy.
A total that reconciles while individual bins do not is the signature of stock being moved between locations without being recorded, and it produces the failure where the system says an item is in stock and the picker cannot find it.
The scenario that follows is common enough to plan for: a product shows as available, an order is promised against it, and a whole case turns out to be physically missing because a previous delivery arrived short and was never checked properly at the door. Every step of that chain is a routine that was skipped rather than a system that failed.
Almost every article on this subject treats automation as an unambiguous upgrade. It is not, and knowing the failure modes before you buy is worth more than another list of benefits.
What an automated stock control system genuinely changes is the transcription step. A scanned item cannot be written down wrong, transposed, or entered against the wrong SKU, and that eliminates a large share of everyday discrepancy at a stroke. It also makes reorder triggers automatic rather than dependent on somebody noticing.
| Where automation genuinely helps | Where it quietly fails |
|---|---|
| Removing manual transcription between shelf and system | Anything received or moved without being scanned stays invisible |
| Flagging items that breach a level, without anyone checking | A mislabeled or duplicated barcode makes the wrong item authoritative |
| Recording who did what, and when | Staff who find the scan step slow will work around it under pressure |
| Reconciling counts and reporting variance | The number stays confidently wrong until somebody physically counts |
| Tracking lots and expiry dates automatically | Dates not captured at receiving cannot be tracked at all |
The pattern in that right-hand column is consistent: automation records what it is told and cannot detect what it was not told. Theft, breakage, a case put on the wrong rack and a delivery signed for unchecked all move physical stock without generating a transaction, so a perpetual system will keep reporting a number that stopped being true days ago.
That is why the count survives automation rather than being replaced by it. The system's job is to be right between counts, and the count's job is to prove that it was. Any candidate for best food inventory software should therefore be judged on how well it supports counting, not on whether it promises to end it.
Everything above ends in the same place: a decision that you are short of something and need more.
That decision is stock control. Acting on it is a separate transaction with a supplier, and it is where accuracy built over an hour in a cooler can still be lost. A count done properly and then read aloud over the phone to a rep who types it into another system arrives as whatever survived that handoff, which is why a careful operation can still receive the wrong quantities.
That gap is where VoiceOrder Solutions operates. It is sold to the distributor, who issues the app to the restaurants and stores on its books, loaded with the items and prices that account has been quoted.
Somebody walks the storeroom, says what is missing, and the distributor receives structured data rather than a message to transcribe. An order spoken at eleven at night is held and queued, instead of decaying in a voicemail box until somebody plays it back.
None of that is stock control. Nothing in it counts anything, no shelf quantity is stored inside it, and it has no view of what is physically in your building. What a distributor sees there moves because an order or a receipt was processed, never because a person walked a storeroom with a clipboard.
So order taking belongs after the count in the sequence, and a tool that improves it cannot substitute for the count that decides what to order.
If the routine does not currently exist, build it in one order: pick the day, pick the owner, then pick the method.
The day matters most and is the easiest to get wrong. Count the day before your largest delivery, at a time when the operation is closed, and put it on the schedule as a shift duty with a named person rather than as a task that floats. A routine without an owner is a routine that stops in the first busy week.
Add the delivery check second, because it protects everything the count measures. One person, present at the door, comparing the invoice against what came off the truck before signing.
The count and the check, held together for a quarter, will do more than any software purchase made before them.
If the reorder is where accuracy leaks out of your week, that is the step to look at first, and it is the one closest to being fixable without changing anything on your shelves. Expect a quote rather than a rate card.
Book a VoiceOrder Solutions demo
Stock control is the routine work of keeping physical stock and its record in agreement: counting on a schedule, checking deliveries against the invoice, rotating product by date, and logging waste. It is performed work with an owner and a time slot.
It stops short of deciding what to stock or at what level, which belongs to the planning layer above it. The distinction is useful because a stock control problem is fixed with discipline and a planning problem is fixed with analysis.
Match the frequency to how fast the category moves and how much it is worth. Perishable and fast-moving items justify a weekly count, frozen and dry goods weekly or biweekly, cleaning and paper supplies monthly, and equipment or smallwares quarterly.
The specific day matters as much as the frequency. Counting the day before a large delivery means less product to count, and counting outside service hours means nobody is removing stock while you work.
Stock control is the recurring physical routine: counting, receiving, rotating and logging. Inventory management is the decision layer: what to carry, at what level, from whom, and what it is worth. Stock inventory control is used loosely for both, which is where most of the confusion starts.
They fail in ways you can tell apart. If the shelf and the system disagree, that is stock control. If they agree and you are still holding the wrong products, that is inventory management.
No, and treating it that way is the most common way these systems lose credibility. An automated system updates on transactions, so it accurately reflects everything that was scanned and is blind to everything that was not.
Theft, breakage, mis-shelving and an unchecked delivery all change physical stock without creating a transaction. Counting is how you find those, and automation reduces how often you need to count rather than eliminating it.
Operators counting $15,000 to $20,000 of ending inventory commonly report three to seven labor hours a week, and the range depends far more on preparation than on the amount of stock.
Counts run fastest where shelves have been rotated and organized on a previous shift and deliveries have already been entered, which turns counting day into counting rather than tidying. Separating those two jobs is the single change that most reliably halves the time.
