Keep the control. Hand off the vigilance. Vista Watch: early accessVista Watch, now in early access
Integration triage
Pick the symptom that matches what you are seeing in your NetSuite, Celigo, and commerce stack. We will tell you what is most likely happening and exactly where to start, no form first, no sales pitch.
The diagnosis
Pick what you are seeing on the left. We will tell you what is most likely happening and where to start, no form first.
A starting point, not a diagnosis of record. When the problem is bigger than a config change: a parallel run, a cutover, a close that will not tie. That is the right time to bring someone in.
Every symptom, in full
The same diagnoses the tool surfaces, laid out to read straight through.
A marketplace or processor deposit is a netted number: your sales minus the fees, refunds, reserves, and adjustments the channel already took out (on marketplaces, the facilitator tax too). Until each of those components lands on its own GL line, the difference hides in a clearing account and the reconciliation never ties. The fix is a deterministic decomposition that foots to the deposit, with genuine exceptions routed to a person instead of swept away.
This is almost always a match-key or idempotency problem. A blank or non-unique item key drops the order on a failed lookup; a mismatched externalId format between old and new flows, or a replayed at-least-once webhook, creates the same order twice. Each has a precise, repeatable fix once you can see which one it is.
The books often cannot close until every channel's payout is reconciled, so when that work is manual, reconciliation, not the close itself, is usually the bottleneck. Reconcile the settlements as the files land and the matched cash, isolated fees, and posted entries are ready when the close opens, instead of being the thing that holds it up.
Stalled projects usually trace to scope drift, a thin integration design, or a handoff that left no clear owner. The move is not to add more scope. It is to triage first: stabilize the highest-risk breakpoints, get the system defensible again, and only then decide what is worth building next.
A common cause of silent failures is an execution-context gap: a NetSuite script or workflow filtered to one context, a CSV import or the UI, never runs on records an integration creates through web services. There is no error; the validation or auto-populate just does not happen, so the cost shows up later in the data, not in an alert.
These are source-modeling errors. Marketplace-facilitator tax that the channel is liable for (channel_liable) double-reports when it is not carried into NetSuite; discounts synced as item price adjustments instead of discount line items hide revenue, pollute the GL, and break refund and margin reporting. Both are fixable at the mapping, not with a workaround.
This is post-go-live drift: the system that fit at launch now needs more channels and far less hand-keying. The answer is targeted optimization and governed automation that adds capacity without adding headcount, not a full re-implementation, which trades a working system for risk.
Tell us the symptom and what you were doing when it started. We will tell you what it means and what it takes to fix, whether that is a config change, a reconciliation, or a rescue.