How to Plan a Vending Route and Pre-Kit the Truck
Planning a vending route well comes down to two decisions made the day before: which machines are worth a stop, and what goes on the truck for each of them. Operators who make both in the warehouse, from numbers, run fewer miles, carry less product and spend the visit filling instead of guessing. This is how to get there with or without telemetry.
Start with what each machine actually needs
The default route for most small operators is a fixed loop: the same machines every Tuesday whether they are empty or not. It is simple, and it wastes the two scarcest things you have, a driver's hours and the space on the truck. A full machine visited on schedule is a stop that cost you forty minutes and earned nothing; an empty machine not due until next week is a location that is quietly looking for a new vendor.
The alternative is a dynamic route: each day's stops chosen from the machines that need one. To do that you need, for every machine, an estimate of its current stock per selection. That estimate comes from one of three places, in descending order of reliability:
- Telemetry from a card reader or telemetry device that pulls DEX reads on a schedule. If you have it, you know this morning's stock within a vend or two.
- The last DEX file pulled at the previous visit, minus nothing, since you have no sales since. This tells you what was left when you closed the door, and you project from there.
- The last count and a sales rate. If the machine sold an average of 14 units a day over the last six weeks and you filled it to 180 nine days ago, you have about 54 left. Crude, but far better than a fixed loop, and it gets better with every visit's count.
However you get it, convert the estimate to a single number per machine: days until it drops under your threshold, where the threshold is the fill level below which the location starts to look bare. Many operators use 30 to 40 percent. A machine at 2 days goes on tomorrow's route; a machine at 9 days does not.
Order the stops by reason, not by map
Map order matters, but it comes second. Decide who gets a visit first, then drive the shortest path through them. A practical ranking for which machines make the cut:
| Priority | Reason | Typical trigger |
|---|---|---|
| 1 | Under threshold now | Estimated fill below your floor, or a sold-out event |
| 2 | Alert on the machine | Bill jam, coin mech error, door open, temperature on a food machine |
| 3 | Empty soon | Will cross the threshold before the next scheduled day in that area |
| 4 | Cash at cap | Estimated cash in the box near the amount you are willing to leave on the street |
| 5 | Too long since a visit | No visit in two weeks, whatever the numbers say; locations notice absence |
| 6 | Unknown | No estimate at all; a new machine or a broken reader. Visit to establish one |
Everything that does not make the list stays off the route and gets re-evaluated tomorrow. Keep the omitted machines visible with a one-line reason, because a driver who sees "full, 11 days" trusts the system far more than one who just sees a shorter list.
Pre-kitting: build the load before the truck leaves
Pre-kitting (also called kitting or pre-picking) means picking every machine's fill in the warehouse the day before, from the stock estimate, so the driver loads exactly what each stop needs and nothing else. The alternative, the rolling store, where the truck carries a bit of everything and the driver shops from it at the machine, is how most routes start and why most trucks are overloaded with product that turns twice a year.
The pre-kitting process in its simplest form:
- For each machine on tomorrow's route, compute par minus estimated stock for every selection. Par is the quantity a full coil holds, or the level you have chosen to fill to for that selection.
- Round each line up to what the coil can take and down to what you actually want to put in it. A slow seller with 2 left and a par of 10 may only deserve 4.
- Sum the lines across machines into a pick list by product, then convert to whole cases, because that is how the warehouse is shelved. The remainder from an opened case is tomorrow's first pick.
- Pick by product into totes or bins, one per machine, labelled. Scanning each item as it goes in catches the wrong size or flavour before the truck does.
- Mark the kit as loaded and take it out of warehouse stock at that moment, not when the machine is filled. Otherwise your warehouse count is always a day wrong.
The pick list by machine is for the driver; the pick list by product is for whoever picks. You want both views of the same numbers.
The visit itself
When the planning is done in the warehouse, the visit becomes short and repeatable. The driver opens the machine with a list of what to add per selection already filled in, confirms or corrects the count found, adds the kit, pulls anything expired or dead with a reason, pulls the cash and notes the amount, pulls the DEX file if the machine produces one, and closes up. A count only matters where the estimate was wrong, and those corrections are what sharpen the next estimate.
Two habits make the difference between a dynamic route that holds up and one that drifts:
- Record a skipped stop as a visit with a reason. "Location closed", "could not park", "machine moved". A skip that leaves no trace just makes the machine look visited.
- Finish the visit on the phone before driving off. Counts entered at the end of the day from memory are the main way estimates go bad.
What to measure after a month
Compare the month before and after. Stops per day should fall and units per stop should rise. Truck load at the end of the day should be close to zero. Sold-outs seen on arrival should drop. If any of those went the wrong way, the threshold is set too low, the sales rate window is too short, or the counts are not being entered. All three are fixable; none of them is a reason to go back to the loop.
How Vending Route does this. Vending Route computes every machine's fill and days-to-threshold from the best source it has (a Nayax sync, the last DEX file, or the last count and sales rate), and builds the day's route from the stops worth a visit, each with its reason, fill, source and cash estimate; full machines are listed below with why they were left off. The Kit screen gives the load by product in whole cases or by machine, for today or tomorrow, with tick-by-scan, and "Mark kitted" takes the stock out of the warehouse. On the visit the add quantities are pre-filled, counts and pulls take a reason, cash pulled is checked against the machine, and it all works with no signal and syncs later. Drivers use the Android app; you plan in the web app. More about Vending Route.
More guides: all Adelyons guides