How to build a van sales load list that doesn't guess

A beverage distributor running nine vans out of a single depot loaded every truck the same way every morning: the driver walked the warehouse with a clipboard and pulled roughly what he'd pulled last Tuesday, adjusted up if it had been hot that week, adjusted down if a big account had cancelled an order. Nobody had written this process down. It lived entirely in the head of whoever drove that route, which was fine until the depot's two longest-serving drivers left within a month of each other.

Their replacements didn't know the rhythm of the route. One loaded too heavy and came back with a third of the truck still full, which meant warehouse staff had to unload, recount, and re-shelve stock that had spent all day riding around in the heat. The other loaded too light, ran out of the two best-selling SKUs by the fourth stop, and either turned away orders or scrambled to get a same-day top-up run out to him — burning fuel and driver hours on a delivery that shouldn't have been needed.

Why "load what you sold last time" breaks

The clipboard method works only as long as the person holding the clipboard is the same person who drove the route long enough to build the instinct for it. That instinct is real — it's just not written anywhere, not transferable, and not adjusted for anything the driver doesn't personally notice. It doesn't account for a customer who called ahead with a bigger order, a promotion that's about to move volume, or a SKU that's been trending up for three weeks running. It also doesn't distinguish between a slow week because demand actually dropped and a slow week because the driver forgot to offer the new flavor at half his stops.

The cost shows up in two directions that look opposite but come from the same root cause. Overloading ties up cash in stock that rides the van instead of turning over, adds fuel and wear for weight the truck didn't need to carry, and creates a return-and-reshelve cycle at day's end that eats warehouse labor. Underloading is worse for the top line: a stockout mid-route is a sale the rep can't make on the spot, and in van sales the on-the-spot sale is the whole model — a customer who wanted a case and got told to wait for tomorrow often just buys from whoever showed up with the case in hand.

What a load list is actually solving

A load list is the answer to one question, asked per route, per SKU, every morning: how much of this should ride on this van today? Getting that right means replacing "what I remember selling" with two things a spreadsheet or clipboard can't hold at once — a rolling par level per SKU per route, and same-day exceptions layered on top of it.

The par level is built from actual sales history on that specific route, not a warehouse-wide average. A route that covers convenience stores sells differently than one covering restaurants, and averaging them produces a number that's wrong for both. The par should move with a trailing window — the last several visits to that route, weighted toward the more recent ones — so a SKU that's been climbing shows up heavier on the list before a driver notices it by feel, and a SKU that's cooling off gets trimmed before it's riding around unsold for two weeks.

Exceptions sit on top of that baseline: a customer who phoned in a larger order for today, a promotion pushing volume on one item this week, a known closure that should pull a stop off the list entirely. None of this should require the driver to remember it — it should already be reflected in the number by the time the load starts.

Making the count match the list

A load list only holds up if what actually gets loaded matches what the list says — and that means the count at load-out has to be as disciplined as the count at day close. If drivers are trusted to eyeball the pallet and call it close enough, the list becomes a suggestion instead of a control, and the gap between "list says" and "van has" is exactly the kind of gap that turns into an argument at reconciliation. Scanning or confirming each SKU at load time, against the generated list, closes that gap before the van leaves the yard rather than discovering it at 6pm.

This also changes what "reconciliation" means at the end of the day. Instead of comparing a vague morning estimate to a vague evening count, the day closes against a number that was precise from the start: loaded quantity, minus sales, minus confirmed returns, should equal what's still on the van. When it doesn't, the variance is real — a shrinkage signal worth investigating — rather than noise from an approximate starting point.

Practical takeaway

If your load list is still built from memory, here's what to put in its place:

• Set a par level per SKU per route, from a trailing window of that route's own sales history — not a warehouse-wide average, and not one driver's memory.
• Weight recent visits more heavily than older ones so a rising or falling SKU shows up in the number within a couple of weeks, not a season.
• Layer same-day exceptions on top of par — pre-booked orders, promotions, known closures — so the list reflects today, not just the trend.
• Count against the list at load-out, by scan or confirmed tally, so the van leaves the yard with a verified quantity instead of an estimate.
• Reconcile against that same number at day close so a variance means something is actually wrong, not that the morning guess was off.
• Review par levels on a schedule, not just when a route visibly breaks — a slow drift is harder to notice than a stockout, but costs just as much over a quarter.

iotoms generates the daily load list itself, from each route's own trailing sales history plus same-day exceptions, and carries that same quantity through load-out confirmation and end-of-day reconciliation — so the number a driver loads against is the same number the day closes against.