Understanding Data Trust

TaktSignal is only as reliable as the operational data it can see. Data Trust is how TaktSignal tells you what it can see, how fresh it is, and where the numbers disagree with each other — before you rely on an alert.

You first meet it at the end of setup (Before using operational alerts, let's check your data). Afterwards it is on the dashboard (Can we trust the data? and the freshness line) and on the Reconciliation page.

Completeness and freshness

TaktSignal tracks four kinds of data separately: inventory, production, customer orders and items & BOMs. For each it shows the age of the last complete read.

• Stale synchronization — if a kind of data has not been read completely for longer than the limit (Settings → Data is stale after), alerts that depend on it drop to Low confidence. Check the connection (Sync and freshness).
• Unknown age is shown as unknown — never as fresh.
• Degraded sources — when a recommended document type cannot be read, the affected kind of data stays incomplete: it is not shown as fresh, dependent alerts get lower confidence, and those alerts are not resolved automatically while it lasts.

Typical gaps and what they do

Gap	Effect
Job Card not readable	Operation progress comes from work orders only and may lag the floor; timing confidence is lower.
Stock Entry not readable	Material issued versus used per work order cannot be compared (material reconciliation unavailable).
Purchase Receipt / Invoice not readable	Supplier deliveries and invoiced quantities cannot be checked.
Warehouse not readable	Warehouse roles are guessed; stock in unrecognised warehouses is not counted as usable.
Missing purchasing data	Incoming supply is unknown, so shortages may be overstated.
Assumed calendars	Work centers without shifts use a default calendar; capacity and timing alerts say so.
Unsupported statuses or customisations	Records TaktSignal cannot interpret safely are quarantined and the data is marked incomplete.
Incomplete material movements	Issues, returns, scrap and offcuts that are not recorded in ERPNext show up as unexplained variance.

Reconciliation — do the numbers agree?

The Reconciliation page compares what ERPNext believes with what physically happened:

• Work-order material variance — expected material (BOM × good output) versus what was accounted for (issued − returned − reusable offcuts), split into known variance (scrap, rework, explained) and unexplained variance.
• Supplier receipts — ordered versus received in ERPNext versus invoiced versus physically verified.
• Stock counts — physical counts (recorded in ERPNext or in TaktSignal) versus system quantity.
• Inventory trust per material — low when counts or variances disagree. Shortage alerts on low-trust materials get lower confidence and an Inventory trust alert explains why.

TaktSignal never corrects ERPNext. People record counts, receipt checks and explanations in TaktSignal; whether and how to correct ERPNext is your decision.

How Data Trust affects alerts

Every alert lists the data it depends on. When that data is stale, incomplete or doubtful, the alert's confidence drops and the reason is written on the alert. A clean Data Trust picture does not make an alert right — it means the evidence is as good as TaktSignal can see.