nohold vs splitting orders manually in Brightpearl.

The tax problem is the most expensive. Brightpearl prorates tax at the order level. When you split manually and adjust totals by typing them in, you can get tax to the right number per Sales Order, but the line-level rounding rarely matches what Shopify reported. Three months later, your accountant flags a small discrepancy across 50 orders. You spend a Friday on a reconciliation that wouldn’t exist if the splits had landed correctly the first time.

Shipping has the same shape. Discounts are worse, because Shopify-side discount codes can be percentage or flat fee, and the manual splitter usually applies them with a flat division. Returns processed against manually-split orders have a noticeable reconciliation error rate, somewhere in the 12 to 18% range based on what we’ve seen in audits. None of those costs show up on a single invoice, which is why they’re easy to ignore.

The split runs on the Shopify webhook before either system has a queue entry. The two Brightpearl Sales Orders land already split, with prorated tax, shipping, and discounts on each half, with the right status codes, with audit notes pointing back to the source order.

No one on ops opens anything. No one types totals. No one chases the warehouse for a release once stock arrives, because the auto-release flips the held SO the moment Brightpearl shows stock on the pre-order line. The two SOs look like any other Shopify-to-Brightpearl order. Your team works the orders the way they always did. The manual workflow goes away.