Vending stockouts: test refill decisions without treating zero sales as zero demand

A low weekly sales total can describe weak demand or a product that sold out early. Before reducing its stock allocation, reconstruct when customers could actually buy it. Use the resulting evidence to test a refill or capacity change, while treating estimated missed purchases as a hypothesis rather than recorded revenue.

Nayax's inventory documentation describes stock levels as estimates supported by the machine's product map and sales reporting. It also distinguishes live and scheduled reporting sources and notes that sales accuracy depends on reporting. That is a useful reminder to check the underlying configuration and visit records before interpreting a dashboard as a physical stock count.

Confirm the product mapping, usable selection capacity, actual fill quantity and last service time. A wrong mapping, a missed refill entry or a machine fault can resemble a demand problem. Reconcile a sample with physical observations before basing route changes on a highly precise but unreliable stock estimate.

Consider a hypothetical machine in a location open for seven comparable trading days each week. One product starts with twenty-four units and sells its last unit at the end of day four. Assume the visit and sales records support that availability history and the selection remains empty for days five through seven.

The ordinary weekly average is twenty-four divided by seven, or about 3.43 units per trading day. The observed rate during four fully stocked trading days is six units per day. Using the lower figure as the next week's demand assumption can reproduce the shortage because it treats the empty period as a normal selling opportunity.

The six-unit rate is not proof of forty-two units of weekly demand. The later days may attract different customers, a promotion may have ended, or purchases may shift to another selection. Estimate availability by the relevant trading hours where possible, and use comparable periods rather than assuming every calendar day is equally productive.

If six units per day were sustainable across all seven days, the hypothetical weekly requirement would be forty-two units, eighteen above the original fill. That is a planning scenario based on observed available-period sales, not a measurement of eighteen customers who were definitely turned away.

A second visit at the start of day four could add eighteen units in this example. After three days at six sales per day, six units remain; adding eighteen restores the selection to its twenty-four-unit capacity. Those twenty-four units would cover the remaining four days at the assumed rate.

Preserve the physical capacity check in the plan. A refill quantity cannot exceed the approved usable selection space, and products must remain suitable through the intended selling period. A high modeled demand rate does not justify overfilling the mechanism or ignoring the product's applicable handling requirements.

Assume each additional sale of the refilled product contributes an illustrative $1.20 after its relevant unit costs. Eighteen additional sales would contribute $21.60 before visit costs. However, suppose six of those buyers would otherwise have purchased another machine product contributing $0.80 each. That alternative contribution totals $4.80.

The estimated increase in machine contribution is then $16.80, not $21.60. If the extra visit adds $22 of actual labor and travel cost, the modeled weekly change is negative $5.20. The refill can increase the target product's sales and still reduce the route's contribution under these assumptions.

The substitution count is uncertain and should be tested, not asserted as a fact about customers. Review total machine contribution and the affected selections alongside the promoted product. A product-level sales gain can partly reflect movement within the same machine rather than additional spending at the location.

If the equipment and product configuration permit it, a second selection could increase capacity without adding the same service trip. In the example, two twenty-four-unit selections offer forty-eight units of capacity, enough for the forty-two-unit scenario. That is a capacity comparison, not a recommendation to fill every position to its maximum.

The new selection can displace an existing product. Suppose the removed product would have sold eight units at an illustrative $0.70 contribution, and those sales are genuinely lost rather than transferred elsewhere. Its $5.60 contribution must be deducted from the proposed change. Configuration work and any other affected costs also belong in the decision.

Under the earlier substitution assumptions, the $16.80 contribution opportunity becomes $11.20 before those setup costs. Compare the recurring result over a stated period with the actual changeover effort. Keep the machine's product range and the location agreement in view; a small calculated gain does not automatically satisfy every service commitment.

Change a defined set of comparable periods and record the stock plan, prices, trading conditions and machine status. Measure whether the product remained available and whether total machine contribution improved. A trial that still stocks out cannot reveal full demand, while a trial during a special event may not represent ordinary weeks.

Keep physical count differences, faults and reporting gaps separate from demand outcomes. If the machine stopped reporting, an apparent flat sales period should not be classified as zero demand without checking what happened. Resolve the data problem before tightening a replenishment rule around it.

Use the result to choose a supported fill quantity, facing allocation or visit schedule. Retain the uncertainty where the evidence remains limited. The strongest decision is not the highest estimated missed-sales number; it is a feasible change that improves contribution while making the product available when customers actually want to buy it.