How to Track Vending Machine Routes, Inventory, and Revenue

Vending margin lives in the route. The driver who restocks 14 machines a day vs. 22 is the difference between a profitable territory and a loss leader. The location that pays you a 25% commission on $400 a month is fine; the location that pays 25% on $90 a month — after you drive 47 miles each way — is bleeding you. And the machine that throws a coil jam at 7 a.m. on Monday and doesn't get a service call until Thursday is dead revenue for three days, in a business where weekly velocity is the whole game.

Most operators I know running 50-300 machines manage all of this in a spreadsheet, a clipboard, and the route driver's memory. That works at 30 machines. It does not work at 200. By the time you scale past 100 machines, you're losing track of which slots are slow movers, which locations have stopped paying out commissions, and which drivers are skimming a few dollars off the cash collection on the way back to the depot.

This is the playbook I use to run a tight operation. Five layers: telemetry on every machine, route optimization per driver, planogram per location, location commission tracking, and driver cash reconciliation. Every layer feeds the next. Skip one and the operation gets sloppy fast.

If you're still driving to a machine to find out how empty it is, you're a generation behind. Modern vending machines support DEX (Direct Exchange) and MDB (Multi-Drop Bus) protocols that report sales by SKU, cash and cashless totals, error states, and inventory levels in near real time. A cellular telemetry module ($150-300 per machine, $5-10/mo per machine for the data plan) makes this data available in your back office.

What I want to see on the telemetry dashboard every morning, before any driver leaves the depot:

- Sales by machine, by SKU, for the last 24 hours and trailing 7 days - Current inventory level per slot (rows that are below 30% trigger a restock flag) - Machine error states: bill validator jam, coil error, cooling unit fault, refrigeration temp out of range - Cash vs. cashless split (a machine where cashless suddenly drops to zero is usually a card reader that lost connectivity) - Last successful sale timestamp (a machine with no sales in 36 hours is either dead, unplugged, or the location closed unannounced)

This becomes the source of truth that drives the route. Without it, you are guessing.

Once you know what every machine needs, the next question is: in what order does the driver hit them? The naive answer is geographic — closest first. The right answer is a weighted blend of inventory urgency, driving distance, location windows (some offices won't let you in before 9 a.m. or after 4 p.m.), and machine error priority.

My weighting roughly:

- Machines below 20% inventory on top sellers: highest priority, must hit today - Machines with active error states (bill jam, coil fault): hit before any other restock — every hour offline is lost revenue - Machines below 40% inventory: target today if route allows, otherwise tomorrow - Machines above 60%: skip, revisit in 2-3 days - Routes are then optimized as a traveling-salesman problem against this priority list, constrained by location-access windows

The driver gets a mobile app on the truck — pre-built route, machine list, inventory recommendation per slot (based on planogram + current sales velocity), and a barcode scanner for restock confirmation. The app captures restock time, slot-by-slot fills, cash collection, and any service notes. No paper, no end-of-day reconciliation drama.

A well-routed driver in a dense urban territory should hit 22-28 machines a day. A poorly routed driver in the same territory hits 12-15. That gap is your margin.

A vending machine in a high school sells different products than the same machine in a hospital ICU break room. School: sugar drinks, candy bars, hot Cheetos, $1-2 price points. Hospital: bottled water, low-sugar protein bars, sandwich items, $2-4 price points. Office park: coffee drinks, energy drinks, healthier snacks. Same hardware, completely different planogram.

The planogram is the slot-by-slot map of what product goes in which coil at what price. I run an ABC analysis per location every 60 days:

- A items (top 20% of SKUs by revenue): always stocked, premium slot positions, never sub out - B items (next 30% by revenue): stocked, can rotate seasonally - C items (bottom 50% by revenue): candidates for replacement — slow movers eat slots that A items could fill

A slot that turns less than once every 14 days is dead capital. Pull it, replace with a tested A or B item from a similar location. Track the lift over the next 30 days. If the new SKU underperforms, swap again. Vending is a perpetual A/B test where the cost of being wrong is one restock cycle of foregone revenue per slot.

Location managers also have opinions, which matter politically even when they're commercially wrong. The hospital cafeteria manager who insists you stock baked chips that nobody buys is a relationship cost you have to weigh against the revenue cost.

Most vending placements pay the location a commission — usually 10-30% of gross revenue, sometimes a fixed monthly fee, sometimes a hybrid. The contract terms vary by location, often by machine, and they have renewal cycles you cannot afford to miss.

What I track per location:

- Commission structure: percentage of gross, percentage of net (after sales tax and product cost), fixed monthly, or hybrid - Payment cadence: monthly, quarterly, or year-end reconciliation - Statement format the location wants: detailed by-SKU, summary by machine, or just a check - Contract start, end, and auto-renewal terms - Exclusivity clauses (some require you are the only vending operator on site)

Every month I generate location statements automatically from the telemetry data — gross revenue per machine, less any agreed-upon deductions, with the commission calculated and the payment attached. Locations that get a clean, on-time statement don't shop the contract. Locations that get a vague handwritten check 45 days late are reading their next operator's pitch deck.

The other use of this data is renewal triage. 90 days before a contract expires, I look at trailing-12-month revenue per machine. Machines doing under $200/month gross are renegotiation targets — either drop the commission rate, swap the machine for a smaller one, or pull it entirely. Machines doing $600+/month are protected — sometimes I'll proactively bump the commission a point or two to lock in a 24-month renewal.

Cash is where vending operations leak. Cashless transactions reconcile themselves — the processor settles into your bank account, the telemetry log matches, and shrink there is essentially zero. Cash is different. The driver opens the bill validator, pulls the stacker, counts (or doesn't), and the difference between what the machine reported and what the driver deposits is your shrink window.

My reconciliation flow per route:

- Telemetry reports cash sales total per machine since last collection - Driver scans the machine, opens the validator, drops the stacker into a tamper-evident bag with a printed manifest - Bag goes to the depot or directly to a bank deposit drop - Counting room (or armored car service) opens bags, counts, and posts actual deposit amounts back into the system - System compares telemetry-reported cash to actual deposit per machine - Variance over 2-3% on any single machine triggers an investigation flag

Most variance is innocent — bill validator counted a $5 as a $1, a customer stuck a fake bill, a coin mech rejected a quarter that fell in anyway. But a driver whose machines consistently come in 4-6% under telemetry across the route every week is a problem. The data tells you which driver, which machines, and how much. From there it's an HR conversation, not a forensic investigation.

Going increasingly cashless solves a lot of this. Machines with card readers and mobile-pay support typically run 60-80% cashless in office environments, which collapses the shrink surface. But cash will not be zero in this business for the foreseeable future, and reconciliation has to be airtight.