Distribution & Logistics Software

The most common structural failure in distribution software is stock maintained per location as separate numbers. The main warehouse has its register, each branch has its own, and van stock is somebody's notebook. Every reconciliation is then a negotiation, and the business simultaneously runs out of an item in one place while overstocking it in another.

We model stock as a single derived position across all locations, computed from recorded movements — inward, transfer, allocation, dispatch, return, adjustment — each referencing the document that authorised it. When a physical count disagrees, you can find the missing movement instead of overwriting the number and hoping. It is the same document-linked discipline described on our manufacturing pages, and it works for the same reason.

Checking a customer's outstanding balance after dispatch is checking it too late. The goods have gone, the invoice exists, and the only remaining option is a difficult conversation. Yet a great many distribution systems do exactly this, because the credit check lives in accounts and the order lives in sales.

So the check belongs at order entry, against a live outstanding balance that includes unpaid invoices and undelivered orders already in the pipeline. Where a customer is over their limit, the order is held pending an override with an approval workflow — which means someone senior takes a documented decision to extend credit, rather than the business discovering it accidentally. That single change is frequently the fastest return in a distribution project.

Delivery disputes are settled by evidence, and in most distribution businesses there is none — a signed paper challan that may or may not still exist, and a driver's recollection three weeks later. The customer says four cartons; the invoice says five; the credit note gets written because arguing is not worth the relationship.

A driver app capturing signature, photograph, timestamp and GPS location at the point of handover changes that entirely, provided it records partial delivery and refusal at line level rather than as a free-text note. It also has a second effect worth more than the disputes it wins: because delivery status is captured as it happens, the office knows about a refusal in the afternoon rather than the following week, while it can still be acted on.

Goods coming back are where stock and money most often stop agreeing. A driver brings cartons back, the customer is credited to keep them happy, and the physical goods sit in a corner without ever re-entering the system. The write-off appears at the annual count as unexplained shrinkage, and by then nobody can trace it.

We model returns as an explicit sequence — authorisation, inbound receipt, inspection outcome, restocking or write-off decision, then credit note against the original invoice. Each step is a tracked state with an owner. It is slightly more process than people expect, and it is the difference between shrinkage you can explain and shrinkage you simply absorb.