A workflow may keep triggering while the CRM receives a contact with no phone number, source or other important field. The gap could be in the form, a stale mapping, a filter or the destination. An “on” status does not prove that the right data reached the right record.
The short answer
Choose one recent record you are authorized to inspect and trace the value through source → mapping → filter or transformation → destination. Decide whether the property was never sent, was legitimately blank, changed name or type, or disappeared later. If the value is required for the business action, pause that action or send the record for review. Do not invent a default just to make the run look green.
Zapier documents “missing value” when fields are renamed, removed or exist only in setup samples. Make documents validation errors when required input data is missing. These are possible symptoms, not a universal diagnosis: an empty field may also represent a customer who deliberately did not answer an optional question.
A field contract, not one generic fix
| Observed state | Evidence at source | Decision before the action |
|---|---|---|
| Property absent | The source did not send the property | Check form and trigger; do not remap blindly |
| Property present but empty | It exists without an answer | Check whether optional; review if essential |
| Renamed or moved | Value appears under a new path | Update mapping; test old and new records |
| Type or format changed | Text replaces a number/date or a nested object appears | Normalize by explicit rule; reject ambiguity |
| Reaches filter but not destination | Input exists; output is missing or rejected | Inspect condition, permissions and destination response |
This Pulse matrix separates five different decisions. An optional phone number can remain blank; a required order ID should not silently create a financial update. “Required” is a business rule, not a visual property of a form control.
Trace the record without exposing the customer
- Identify a specific record ID, time and version of the source form or app. Use authorized access and minimize personal data viewed.
- Inspect the trigger output. If the value is already absent, investigate the source question, conditional display or schema change. An old sample is not proof of a current payload.
- Compare the exact property mapped downstream. “Status” and “Status reason” may look similar but mean different things. Change the mapping only when the correct source field is known.
- Check filters and transformations. Whitespace, number representation and date format can change a condition without a dramatic error.
- Inspect the destination record by source ID. A completed run may still have written an incomplete record; verify the actual outcome.
Keep only the audit metadata needed: a pseudonymized ID, step where the value vanished, classification, owner and decision. Do not paste full payloads containing personal data into a public ticket or an external AI service.
Fix, pause or recover
If the field is optional, document the legitimate blank and proceed only if the destination supports it. If required, stop the write action and route it to review. For a renamed field, remap in a controlled environment using two fictional examples: old and new. For a format change, define a conversion rule and an error path. Then compare source and destination IDs for the affected period; do not replay everything without checking for duplicates and effects already applied.
A tabletop check uses five cases: complete value, optional blank, required missing, renamed field and unexpected format. It fails if a required-missing case creates an apparently complete record without evidence. These are proposed cases, not tests performed by Pulse.
For up-front field ownership, see how to map the business process. If the gap already affects service, follow the automation failure contingency plan. The first prevents design gaps; this guide diagnoses a mapping that no longer represents current data.
Sources and limits
Zapier’s filter guidance covers missing values, stale samples, renamed fields and formats. Its data transfer guidance distinguishes empty from removed properties. Make explains required-data errors. The field contract and decision flow are editorial models; verify your own rules, permissions and platform behavior before changing production.