Guides
ATM Route Scheduling: Build a Day That Holds
ATM route scheduling uses machine needs and honest stop times to shape a day. Reserve fault slack, vary stops, and compare planned with actual time weekly.
What to take away
- Sequence by need first and geography second. A tidy loop that opens full machines is an expensive tidy loop.
- Every stop has a variable part you cannot predict, and the schedule has to hold that variance rather than assume it away.
- Predictability is a security problem. Build the variation into the plan instead of leaving it to chance.
- Faults do not fit in the schedule, so reserve capacity for them. A day with no slack turns one jam into three late stops.
- Measure the gap between planned and actual time per stop. That single number tells you whether the schedule is real.
ATM route scheduling is usually taught as a shortest-path problem. It is not. The distance is the cheap part, and the expensive parts are the stops that take longer than expected, the machine that needed nothing, and the day when a fault arrives at ten in the morning.
Build the day from a need list, not a map
Start with which machines require a visit. Build that need list from four sources:
- Cash levels and expected withdrawals before the next visit.
- Machine monitoring alerts and unresolved faults.
- Scheduled periodic looks or inspections.
- Host-reported issues that require an on-site response.
Need List to Trip Order
- Identify machines needing visits
- Cash below reorder level
- Open fault reported
- Unexplained transaction drop
- Site not physically seen
- Group into trips by shared drive
Only then apply geography, by grouping the list into trips that share a drive. The order matters: geography arranges the work, it does not select it.
Where a machine on the way needs nothing, leave it. Opening a full machine costs the same time as opening an empty one, and every open is an exposure. The exception is the periodic look, which is deliberate and scheduled rather than opportunistic.
Sequencing within a trip
Three constraints usually conflict, and something has to win.
Sequencing Conflict Resolution
Which constraint wins?
Cash burn decides -> closest to dry first
Exposure objects -> move high-cash stop
Cash burn. The machine closest to running dry goes first, because that is the one where lateness has a cost in missed withdrawals.
Venue hours and tolerance. A bar does not want you at nine at night, and a clinic does not want you at eleven in the morning. Each site has a window, and the window is often narrower than the opening hours.
Exposure. The stop where you carry the most cash should not be the one where you are most visible or most predictable.
When they conflict, cash burn decides unless exposure objects. A late load costs you withdrawals; a careless one can cost more than that, which is why the ATM route repair and security guide treats the run itself as the risk rather than the machine.
Give each stop a realistic time
Most schedules fail because the stop time is the time the load takes when nothing happens.
Timing a Stop Properly
- Time the stop several times
- Record travel from previous stop
- Record time on site until tested
- Record venue conversation time
- Keep the range, not the average
- Widen allowance for narrow windows
Time a stop properly, several times. Record three components: travel from the previous stop, time on site until the machine is closed and tested, and the venue conversation you should have anyway.
Use this copyable row for each stop: Site | visit trigger | access window | planned arrival | travel from prior stop | planned on-site time | actual time | outcome. Put a separate, unassigned fault reserve block on the day plan; do not assign it to a stop.
Keep the range, not the average. A stop that usually takes twelve minutes but occasionally forty is a forty minute stop for planning purposes at least once a week.
Sites with a narrow access window, a manager who must be found, or a machine in a back room need a wider allowance permanently. Those are not bad days. That is what the site is.
Reserve capacity for faults
A fault will arrive on a day that is already full. Plan for it by leaving an unassigned block rather than by intending to work later.
Reserve Fault Capacity
Unassigned block
Size from own records
How often faults interrupt
How long interruptions run
Calculate a baseline fault reserve as total fault-interruption time recorded across comparable route days divided by the number of those days. Review it against individual interruptions and adjust it if the block repeatedly fails to contain them. If you do not have those records, start collecting them before redesigning the schedule, because everything else here depends on them.
A day planned to capacity is a day where one jam pushes three stops into tomorrow, and tomorrow was also full. That is how a route develops a permanent backlog while the operator concludes they need more hours.
Vary the pattern on purpose
A schedule that is efficient and identical every week is a schedule anyone can learn by watching. Vary the day, the time and the order at the sites where you are exposed, and vary them in a way you decide rather than in a way you improvise.
Vary the Pattern Deliberately
- Vary the day
- Vary the time
- Vary the order at exposed sites
- Decide variation, don't improvise
- Tell one person route and finish
The tension is real: the most efficient plan is the most predictable one. Accept a little inefficiency deliberately rather than discovering the cost of the alternative. Tell one person your route and expected finish each day regardless.
Measure whether the schedule is honest
Two numbers, taken weekly.
Weekly Honesty Metrics
- Planned vs actual time per stop
- Stops moved to another day
- Reason for each move
Planned time against actual time per stop, so you can see which sites you consistently underestimate. And the number of stops that moved to another day, with the reason. A schedule where a fifth of the stops move is not a schedule, it is a preference.
Fix the estimates before you fix the route. Most operators redesign the loop when the real problem was that four sites always take twice as long as the plan says.
When the schedule stops fitting the route
The signs are consistent: unplanned trips, reconciliations done late, and hosts telling you about faults before the monitoring does. Those are covered as a service problem in the ATM route operations and workflow guide for owners, and as a growth problem in the ATM route expansion and market guide.
At that point, the options are fewer sites, tighter clustering, less frequent visits funded by larger loads, or help. Larger loads change your parked capital and may change what your insurer and bank expect.
Check that against the risk-based review in the FinCEN statement on independent ATM operators and the requirements in the ATM route licensing overview.
The systems that hold the schedule
Whether the schedule lives in a spreadsheet, a route app or a notebook, it holds site addresses, access notes, timings and machine states, which is information worth protecting. The NIST small business quick-start guides are a reasonable starting point for how a small operation approaches that, and the answer is rarely more software.
Keep the host-facing side of it current too. A venue that knows roughly when you come and can reach you between visits complains less and terminates less, which is the relationship work covered in the ATM route marketing and growth guide.
Common questions
Should every machine be on a fixed cycle?
No. Fixed cycles overserve quiet machines and starve busy ones. Set each machine's reorder trigger from its expected withdrawals until the next feasible visit, plus a reserve based on that machine's own history. For example, if its usable cash is projected to fall short before that visit, the trigger needs to prompt a visit sooner.
Is route planning software worth it?
A route-planning tool can help with the drive legs once the number of stops exceeds what you can sequence in your head. Route4Me is one named option for route planning; keep machine visit triggers and stop-time records in a spreadsheet, since the tool cannot tell you which machine needs opening.
How do I handle a site that always runs late?
Give it a longer allowance permanently and put it at the end of a trip, or change the load so it needs fewer visits. Do not keep planning it optimistically and calling the result a bad day.
What about weekends and holidays?
Plan them backwards from when settlement resumes and from when the venues are busiest. A long weekend is where cash-outs and slow settlement land together, and it is the one to schedule first rather than last.
