September 25, 2026
Integration dashboards: how to measure sales, stock, orders and invoicing in real time
Which indicators an integration dashboard should show: sales, stock, failed orders, pending invoices, syncs and alerts.

Many companies have sales reports, stock reports and invoicing reports, each generated from a different system, reviewed at different moments, by different teams. What almost none of them has is a single view bringing those indicators together in real time to answer the most basic question: is the operation working right now?
This article reviews what an integration dashboard designed for that should show: immediate operational visibility, not a historical report put together at month end.
Quick summary
- A report answers what happened; an integration dashboard answers what is happening and what to do about it.
- Every indicator should enable a concrete action. If it only informs, it belongs in a report.
- The most common mistake is showing everything: a saturated dashboard stops being looked at within two weeks.
- Different profiles need different views over the same data.
The difference between a report and an integration dashboard
A traditional report is generated with data from a moment in the past, by exporting and consolidating information from different systems, often by hand. An integration dashboard is fed directly from the data flows between systems, so it reflects what is happening at the moment it is consulted.
The difference is not just one of frequency, it is one of purpose. The report is for analysing and deciding in the medium term. The dashboard is for acting today: checking a stalled order, restocking, escalating a sync error before it reaches the customer.
Which indicators it should include
| Indicator | What it shows | Action it enables |
|---|---|---|
| Sales per channel | Volume and evolution in real time | Detecting anomalous drops in a channel |
| Stock per channel | Real availability and stockout alerts | Restocking or pausing publication |
| Stalled orders | Orders with no progress beyond the expected time | Unblocking before the complaint |
| Pending invoices | Documents unissued or unlinked | Correcting before the close |
| Sync status | Health of each active integration | Escalating a failure |
| Response times | Latency per system | Detecting degradation early |
| Stock discrepancy | Difference between channel and real stock | Adjusting before overselling |
The right-hand column is the acid test. An indicator that enables no concrete action takes up space and competes for attention with the ones that do.
Who it is useful for and what each profile should see
One single dashboard rarely works for everyone. The usual approach is a common base with views per profile:
- CTO or IT lead: technical health of the integrations, recent errors with their cause, latency and retries.
- COO or operations manager: stalled orders, bottlenecks per stage, stock discrepancies.
- Ecommerce manager: sales per channel for the day, availability of campaign products, orders at risk of missing the delivery promise.
- Finance: pending invoices, unissued credit notes, marketplace reconciliation.
- Customer support: status of a specific order or return, with direct search.
That last view is usually the one with the most day-to-day impact and the one least considered when designing the dashboard.
How to keep the dashboard from becoming noise
The most common mistake is wanting to show everything. An effective dashboard prioritises a few well-chosen indicators that make it possible to take a decision, instead of a screen full of charts nobody finishes looking at.
Four design criteria that sustain its use over time:
- The five-second rule. On opening it, it should be immediately clear whether there is anything to attend to. If it needs interpreting, it is badly designed.
- One indicator, one action. Everything shown has to enable a concrete decision.
- Visual thresholds, not just numbers. The team should not have to remember which value is normal.
- Access to the detail in one click. Seeing that there are twelve stalled orders is of little use if finding out which ones requires opening another system.
And a maintenance rule: periodically review which indicators nobody consults and remove them. A dashboard that only grows ends up abandoned.
How it relates to technical monitoring
There is a difference of audience between the two. Technical integration monitoring (metrics and alerts) is the instrumentation for the team that resolves issues: latency, error rate, retries and alerts. The integration dashboard translates that information into operational and business terms: the sales, stock, orders and invoicing KPIs this article covers.
Both are fed from the same source. Without per-transaction traceability in the integration layer, neither can be built with trustworthy data: you end up consolidating information from several systems by hand, which is exactly the problem you wanted to solve.
Frequently asked questions
What is the difference between an integration dashboard and a BI one? The BI one is designed for analysis: it aggregates historical data, allows exploration and answers trend questions. The integration one is designed for operating: it shows the current state of the flows and enables immediate actions. They complement each other, but they solve different needs.
Can it be built with the tools we already have? It depends on whether a common data source exists. If each system maintains its own information with no layer recording the transactions between them, the dashboard ends up being a manual consolidation under another name. In retail, that common source is built by solving how to integrate physical store, ecommerce and ERP in a single data flow.
How many indicators should it have? The ones the team consults and acts on. As a practical reference, between five and eight per profile view is usually the limit before attention scatters.
How often should it be updated? It depends on the indicator. Stalled orders and sync status justify real time or near real time. Sales per channel tolerates minutes. Pending invoices can update hourly without losing usefulness.
Who should be responsible for the dashboard? It is best for it to have an identified owner who periodically reviews whether the indicators are still the right ones. Without that role, the dashboard goes stale as the operation changes and stops reflecting what matters.
Checklist to build your integration dashboard
- Is the dashboard fed from real-time data, not from historical reports?
- Does every indicator enable a concrete action?
- Does it include visible alerts for stalled orders and pending invoices?
- Does it show the health status of the active integrations?
- Can you move from the indicator to the detail without opening another system?
- Do different profiles find there what they need?
- Is there someone responsible for reviewing and pruning the indicators periodically?
You may also be interested in reading:
• “Integration monitoring: technical metrics and alerts” • “OMS in retail: what data it must sync with ERP, POS and the online store to avoid stalled orders” As part of its retail integrations, Weavee centralises sales, stock, orders and invoicing in a real-time integration dashboard, with per-transaction traceability and views per profile, so the operation can be monitored without waiting for the month-end report.


