The call came on a Saturday morning from a tire shop manager who had been patient for three months. Two of his four rows had been empty since Wednesday. Not almost empty. Empty, with the spirals turning nothing. He had called the number on the machine twice and left one message, then bought a case of soda at the gas station across the street and put it in the shop fridge for his crew.
The operator’s schedule said the machine was due Monday. The schedule was right. The machine was wrong.
Vending machine restocking frequency gets decided in one of two ways. Someone picks a day and defends it for years, or someone counts what the machine actually sells and lets that number set the clock. The first is easier to staff. It also costs more than most operators realize.
Why a fixed day drifts out of date
A two week loop is not a schedule. It is an average, and averages hide the two machines that matter. A location that starts strong in March and fades by July is still on the same loop in August, still getting visits it does not need while a machine across town runs dry.
Most operators inherit their vending machine restocking frequency from whoever trained them, and the person who trained them inherited it from someone else. Nobody ever recalculated it against the machines actually on the truck.
The drift is slow enough to miss. A plant adds a shift. A gym loses a third of its members when a competitor opens four blocks away. A school stops buying in June and nobody tells the operator, because the person who signed the agreement retired.
Published operator guides converge on a simple rule. Restock when inventory drops to roughly 30 to 40 percent of capacity, not on a named weekday. That rule needs two pieces of information before it means anything, which is why so many routes ignore it.
The numbers that set the interval
Vending machine restocking frequency is a division problem. Units sold per day, divided into the units a row holds, tells you how many days a row lasts. Everything else adjusts that answer.
The first number is sell through per row, not per machine. A machine can average 40 units a day and still have one row selling 11 while another sells nothing. Rows are the unit that runs empty.
The second number is row capacity, which depends on coil pitch and product size. Standard coils on a snack machine hold 8 to 12 bags. A beverage shelf might hold 6 to 8 bottles deep. Double depth coils change the arithmetic, and so does the fact that a fat bag of chips can eat a coil that was designed for a slim one.
The third number is the cost of the trip, and it belongs in the same conversation. A stop that costs 14 dollars in labor and fuel cannot be justified to rescue 9 dollars of sales.
| Input | Where it comes from | What it changes |
|---|---|---|
| Units per row per day | Sales report or a two week manual count | Days until the row empties |
| Row capacity | Coil pitch and product size | How much buffer a full load buys |
| Cost per stop | Wage, drive time, fuel | Whether a midweek visit pays for itself |
| Stockout cost per row | Price times units per day | What waiting another day really costs |
Operators who track all four stop arguing about calendars. Operators who track none of them restock by feel, and feel is usually a week behind.
What the traffic level actually looks like
The rough bands published by service companies are a reasonable place to start, and they hold up because they are built from the same arithmetic. Facilities with 300 or more daily users usually need service two to three times a week. Mid sized sites with 75 to 300 daily users land at once or twice a week. Sites under 75 daily users often run every one to two weeks without a single stockout.
| Site type | Daily users | Typical visits | First thing to watch |
|---|---|---|---|
| Manufacturing plant, hospital | 300 and up | 2 to 3 per week | Beverage rows, shift change timing |
| Mid sized office, school | 75 to 300 | 1 to 2 per week | Snack rows, Friday afternoons |
| Small office, clinic | Under 75 | Every 1 to 2 weeks | Date codes, slow rows going stale |
Treat those bands as a starting guess and replace them within 60 days. A hospital with 400 daily users and a machine tucked in a basement corridor will not behave like the table, because the table describes traffic near the machine, not headcount in the building.
Par levels, and the buffer that keeps you honest
Par levels are how vending machine restocking frequency turns into something a driver can act on without guessing. A par level is the count at which a row needs attention. The working formula in most operator guides is daily sales multiplied by the service interval, plus a quarter of that as a buffer for variance. A row selling 4 units a day on a three day loop needs 12 to cover the interval, plus 3 as cushion, so 15 units. If you walk up and the row holds 15 or more, it can wait. Below that, it gets filled.
The buffer is not padding. It covers the day a heat wave doubles beverage sales, the day a delivery is late, and the day a driver calls in sick.
Par levels also let a driver work without guessing, and they connect directly to how you handle inventory management between visits. The alternative is a driver who fills everything to the top every visit, which sounds diligent and quietly creates a different problem. Overfilled coils jam. Mixed sizes on one coil double vend. Product that sat through three visits comes home as waste, and waste is the number nobody puts on the route sheet.
What an empty row actually costs
The instinct is to treat a stockout as a small loss, one sale deferred. It is not deferred. A customer who wanted a specific drink and found an empty row buys water, or nothing, or walks to the gas station and stops checking your machine.
Run the arithmetic on one row. A bag of chips selling at 1.50 with a cost near 0.45 clears about a dollar of contribution. If that row sells 5 a day and sits empty for two days, the loss is 10 dollars of contribution, which is most of the profit the row would have made in a week. Across six rows on a 20 machine route, a recurring two day gap is not a rounding error. It is the difference between a route that covers its labor and one that does not.
The waste side of the same coin gets ignored because it happens quietly. Ten stale bags a month per machine is roughly 54 dollars a year in product bought and never sold. Neither failure is visible on a bank statement. Both show up as a route that should be profitable.
What an extra stop actually costs
The mirror image of a stockout is a visit that was not needed, and it is the second half of any honest look at vending machine restocking frequency. This one is easier to measure because it appears in payroll. A stop is not the eight minutes at the machine. It is the drive, the parking, the walk, the load, the count, and the walk back. Realistically 20 to 30 minutes door to door for a single machine in a strip center.
Published figures put a full time driver at 12 to 18 machines a day when stock is pre-picked by location and routing is dynamic. The same driver using traditional methods, carrying bulk stock and deciding at each stop what to load, manages 6 to 10. The gap is not work ethic. It is the difference between arriving with a tote built for that machine and arriving with a truck full of everything.
At a 3,250 dollar monthly cost for a driver, 25 machines works out to 130 dollars per machine. Fifteen machines works out to 217. That number is the reason route density matters more than almost any other decision in a small operation, and it is also why a midweek visit to a half full machine is a real expense rather than a harmless habit.
Consolidation, and the honest limits of telemetry
The cheapest improvement to vending machine restocking frequency is not a new schedule. It is fewer, better loaded visits. Three moves do most of the work.
Pre-pick by location at the warehouse. Pack a tote per machine from the last sales report, load in reverse delivery order, and the truck carries no surplus. This is the single change most often credited with doubling stops per day, and it starts with how the warehouse is laid out.
Cluster by geography and set a minimum. If a machine will not clear the cost of a stop, it waits for the next clustered run. Published guidance suggests giving a weak location 60 to 90 days before moving the equipment, which is fair, but the test has to be about product mix and price rather than drive time.
Use telemetry to skip, not to justify. A cashless reader with remote monitoring reports sales as they happen, and operators who move from a calendar route to a sales triggered one commonly remove 20 to 30 percent of visits. The savings are real. What telemetry cannot tell you is whether the stock was rotated or the glass was wiped, so the visit it removes still has to happen on some rhythm.
When the season changes the answer
Summer is the one period where vending machine restocking frequency genuinely has to shorten, and not only because people drink more. Compressor load rises with ambient heat, so the machine that ran fine in April is working harder in July and failing more often. A route that visits twice a week in winter often needs three in July, and the extra visit doubles as a coil cleaning.
The reverse happens in January at outdoor sites. Sales fall, product sits, and date codes become the risk rather than stockouts. On slow rows, pull anything within two weeks of its date code and log it, because wastage by row is the fastest way to find a merchandising mistake. Cold weather also changes what sells, and a plan built for June will leave the wrong rows full.
Five signs the interval is wrong
The number is wrong if customers tell you before the report does. Watch for these, in this order.
A manager mentions an empty row, even politely. That is a failure that has already repeated.
The same three rows are empty on consecutive visits while the rest of the machine is full. That is a merchandising problem wearing a scheduling costume. Move the fast seller to a bigger coil.
The driver arrives to find a machine at 80 percent. You are paying for a visit that did not earn its cost.
Sales per machine drop for two months with no price change. Customers have stopped trying.
Cash and card totals diverge from the telemetry count. Something is being vended that is not recorded, and the answer is usually a jam rather than a person.
Where to start this week
Pick your three highest volume machines and count what leaves each row for 14 days. No software, no reader, a clipboard and a phone photo of the row counts. Divide units by days, divide row capacity by that number, and you have the true vending machine restocking frequency for each machine.
Then compare it to what you are actually doing. Most operators find one machine being visited twice as often as needed and another running empty for a day and a half every cycle. Fixing those two is usually worth more than a new machine, and it costs a schedule change.
If you are still choosing equipment for a site, or deciding whether a location can carry a route at all, the conversation is worth having before the machine lands rather than after the third empty weekend. Our team works through volume estimates with operators regularly, and the contact page is the fastest way in. For the wider picture on how machines, routes, and margins fit together, start with the business ROI breakdown.
Preguntas frecuentes
How often should a vending machine be restocked
Vending machine restocking frequency is driven by sell through rather than by the calendar. A high traffic site with 300 or more daily users usually needs two to three visits a week, a mid sized office once or twice a week, and a site under 75 daily users every one to two weeks. Restocking when a machine falls to 30 to 40 percent of capacity is a more reliable trigger than a fixed weekday.
Is a weekly visit enough for an office of 40 people
Usually yes, with one condition. The mix has to match the crowd. Forty people who all want the same two drinks will empty those rows in three days while the rest of the machine sits. Fewer rows, deeper coils, and a product mix that reflects the actual building will hold a week comfortably.
What is a par level and how do I set one
A par level is the count at which a row needs filling. Multiply daily sales for that row by the number of days between visits, then add about 25 percent as a buffer. A row selling 4 units a day on a three day loop gets a par level of 15.
Can telemetry replace scheduled visits
It replaces the guesswork, not the visit. Remote monitoring tells you which machines need product and lets you skip the ones that do not, and operators commonly cut 20 to 30 percent of stops that way. Someone still has to rotate stock, check date codes, and clean the machine on a rhythm.
Does restocking frequency change with the season
At outdoor and warm climate sites it does. Beverage and frozen sales climb in summer while the compressor works harder, so the same route often needs an extra visit and a coil cleaning. Winter reverses it, and the risk shifts from empty rows to stale product.
How many machines can one driver restock in a day
Between 12 and 18 with pre-picked totes and sensible routing, according to published operator figures. The same driver using bulk stock and deciding loads on site typically manages 6 to 10. The difference is preparation at the warehouse, not speed at the machine.
