Bulk WhatsApp Registration Check: TXT Input, Excel QA, and CRM Import

A bulk WhatsApp registration check is more than uploading a file and downloading a result. The usable outcome depends on preserving the source list, normalizing numbers with country evidence, selecting the right task, reconciling Excel rows, and importing without overwriting permission or conversation history.

Workflow goal: use WS Registration Check to record whether the task returned a registration signal at that time. It is not an activity test and it does not grant permission to message.

Define the exact question first

Business question	Task	Excel fields	Not established
Did the number return a registration signal?	WS Registration	Phone, registration status	Activity, reply intent, permission
Are activity fields required?	WS Activity	Phone, activity time, active days, mapped number	Online now
Are all supported fields required?	WS Full Format	Ten supported field groups	Customer value or identity
Is the goal only format cleanup?	Local normalization	Error and normalized phone	Platform state

This guide stays with WS Registration. Mixing tasks in one acceptance test makes field lineage unclear and turns “not run” into an apparent negative result.

Create a batch manifest

Assign a unique batch ID, keep the source read-only, and add source row and contact ID to a working copy. Duplicate phones and changed sorting can then be reconciled.

Field	Purpose	Reason
batch_id	Unique job identifier	Joins TXT, Excel, and import log
source_row	Original row	Locates errors and duplicates
contact_id	CRM key	Avoids phone-only joins
phone_raw	Original value	Preserves evidence
country_evidence	Order or form country	Supports country code
permission_status	Separate record	Must never come from checking

Step 1: normalize with country evidence

Convert numbers to an international representation where evidence supports it. The ITU-T E.164 international numbering plan provides the numbering-plan context, but reliable country data is still required. Do not guess a country code from language or intuition.

1. Remove display spaces, parentheses, and hyphens while retaining the raw value.
2. Standardize an international 00 prefix to plus notation.
3. Confirm country code from orders, forms, or governed master data.
4. Quarantine unknown country, abnormal length, and alphabetic values.
5. Preserve duplicate provenance and create a separate deduplicated task view.

Step 2: generate traceable TXT input

The AppShai WhatsApp (WA/WS) Number Checker accepts TXT input only and exports Excel. Use UTF-8 with one phone number per line; do not upload Excel or include names, labels, headers, or comma-separated columns.

Control	Pass condition	Failure action
Encoding	UTF-8 opens correctly	Re-export
Line structure	One phone per line	Remove headers and notes
Blank lines	None	Clean and recount
Row count	Matches manifest task view	Reconcile omissions
Filename	Includes batch and version	Do not overwrite

Step 3: run a representative sample

Include normal, duplicate, multi-country, and boundary-format rows. Confirm WS Registration is selected and the output schema matches Phone and Registration Status before processing the full batch. A small sample validates the pipeline; it is not a population market-share estimate.

Step 4: accept the Excel output

QA measure	Definition	Warning
Unique input	Valid deduplicated TXT rows	Manifest mismatch
Output rows	Excel data rows	Unexplained loss or gain
Status coverage	Rows with explicit status / output	Many blanks
Join rate	Rows matched to manifest / output	Unmatched records
Output duplicates	Duplicate phone groups	Needs review

Do not convert blank, unknown, or format errors into Not Registered. Unknown means insufficient result; a negative status is an explicit task outcome. Preserve that distinction.

Step 5: import through staging

CRM field	Value	Rule
wa_registration_status	Registration status	Only from WS Registration
wa_checked_at	Completion time	Update with status
wa_check_batch_id	Batch ID	Supports rollback
wa_check_exception	Unknown or join error	Never force to No
permission_whatsapp	Existing evidence	Never overwrite
last_reply_at	Observed conversation	Never overwrite

Load a staging table first. Prefer contact ID, use normalized phone only as a governed fallback, and create inserted, updated, conflicted, and unmatched reports before merging.

Step 6: preserve rollback and rerun evidence

Retain the source, normalization manifest, final TXT, original Excel, mapping file, and CRM diff. Roll back only fields written by the batch ID; restoring the whole contact table could erase later replies or opt-outs.

Acceptance sampling should cover positive, negative, unknown, duplicate, and unmatched rows—not only successful results. Compare raw phone, normalized phone, contact ID, and destination field. Where one contact owns multiple phones, define primary and alternate-number rules before any merge.

Production acceptance checklist

• The task matches the business question.
• TXT is UTF-8 with one phone per line.
• Input, output, exception, and duplicate counts reconcile.
• The original Excel is unchanged.
• Status, check time, and batch ID are written together.
• Permission, opt-out, and interaction fields remain intact.
• The diff has been sampled and rollback is available.

Frequently asked questions

Can the exported Excel become a send list?

No. It is task output, not permission. Source, opt-in, opt-out, content, and platform rules still govern messaging.

Should a not-registered number be deleted?

Review format, country evidence, and retention rules first. It may be excluded from the WhatsApp path without deleting the master record.

How often should the job run?

There is no universal interval. Define triggers from data change and use case, and always retain checked_at.

A production-grade bulk check delivers more than a yes/no column: it creates evidence that explains source, task, time, exceptions, and every CRM change.