How to clean a contact list before a WhatsApp campaign

Every list you buy, scrape, or collect over time accumulates numbers that were never on WhatsApp, were reassigned, or were mistyped at signup. Sending to them wastes credits, drags down your delivery metrics, and in bulk-messaging tools can get a sending number flagged. None of that shows up by looking at the list: a bad number and a good number are indistinguishable in a spreadsheet.

This is a checklist for getting a list from "probably fine" to "actually checked" before you send to it.

Why invalid numbers pile up

Contact lists rot, and they rot for a few predictable reasons:

• Time. A number collected two years ago may belong to someone else now, or the original owner may have stopped using WhatsApp on it. Registration isn't permanent.
• Formatting drift. A CSV export from a CRM often has inconsistent formats: some rows with a country code, some without, some with local formatting (dashes, spaces, leading zeros) instead of a clean international format.
• No verification at the point of collection. Purchased lists, scraped lists, and unvalidated form fields all let through numbers that were never checked against anything.

If you're wondering how much of your own list this actually adds up to, see how many WhatsApp numbers are fake for the real decay and ban data behind that question.

None of this is visible until you try to send. That's the core problem with a dirty list: it fails silently until it fails expensively.

The checklist

1. Normalize formats first. Every number needs a country code before it can be checked meaningfully. A number without one isn't invalid. It's just unchecked, and it'll either fail the check or get validated against the wrong country. If your list has a mix of formats (some local, some international), normalize to E.164 (+ followed by country code and number, no spaces or punctuation) before anything else. See how to format a WhatsApp number by country for the country-specific exceptions (Mexico, Argentina, Brazil) that a generic strip-and-prepend formula misses.

2. Validate before you send, not after a bounce. Checking after the fact only tells you what already failed. It doesn't prevent the failure. Validating upfront turns a bounce into a filtered-out row before the send ever happens.

3. Re-validate lists older than a few months. WhatsApp registration changes over time. A list validated in January isn't guaranteed clean in July: numbers get reassigned, people switch to new SIMs, accounts get deleted. Treat validation as something a list needs periodically, not once.

4. Group related checks into a batch. Running a re-validation against a named batch keeps the results tied to that specific list, so you can come back later and see exactly which numbers passed or failed without re-running the check or losing track of which list a result belonged to.

5. Don't stop at "valid": segment by result. A validation pass gives you more than a pass/fail signal. Numbers that fail should be pulled out of the sending list entirely, not just flagged: a flagged-but-still-included number defeats the point of checking it. Keep the failed segment separate so you can decide what to do with it (drop it, try to re-collect it, or route it to a different channel) instead of it quietly re-entering the next campaign.

A concrete example

Say you're cleaning a 5,000-row CRM export before a campaign. A typical pass looks like this:

Step	What happens
Normalize	Strip formatting inconsistencies, add missing country codes
Validate	Check all 5,000 numbers against WhatsApp in one batch
Segment	~4,100 valid, ~900 invalid or unregistered
Send	Campaign goes only to the 4,100 valid numbers

Without the cleanup step, that campaign would have sent to all 5,000, generated roughly 900 failed sends, and reported a bounce rate that looks like a messaging problem when it was actually a list problem the whole time.

Doing it at scale

For a one-off list, checking numbers individually works fine. For anything larger, upload a CSV and validate the whole list in one pass instead of checking numbers one at a time through a UI. See CSV upload vs API for the tradeoffs between a CSV upload and calling the API directly, including when you'd want to check a number the moment it enters your system instead of in a bulk pass later.

The end result is the same either way: a list with only real, active WhatsApp numbers, and a sending step that isn't wasting effort on numbers that were never going to deliver. See why bounce rates run high on WhatsApp outreach for what an uncleaned list actually costs you beyond the wasted sends, and how to reduce WhatsApp broadcast blocks for how those same dead numbers feed into Meta's quality rating.

Numbers that fail the WhatsApp check aren't necessarily dead everywhere: some are still reachable over SMS. See WhatsApp vs. SMS marketing: which one actually gets delivered for how to route the segment you just pulled out instead of just discarding it.