Field notes

How to Turn WhatsApp Chats Into Better Meta Ad Audiences

Automation guide By plumcut Published 5 min read
A notebook and a phone side by side on a desk

Photo by Sancho Papa on Flickr, CC BY 2.0

Most businesses running click to WhatsApp ads give Meta one signal: someone started a chat. Then everything that matters happens inside WhatsApp, where Meta can't see it. Who bought. Who asked the price and left. Who was a supplier, a job seeker or a wrong number. Meta keeps optimizing for people who start chats, not people who buy, because that is the only thing you told it.

Your chats already contain the answer. This guide covers how to classify them, how to send that information back to Meta, and how to use it to retarget the right people and stop paying for the wrong ones.

It picks up where our guide to what happens after the click leaves off.

Why conversation count is the wrong target

When the campaign goal is conversations, Meta looks for more people like the ones who started conversations. If half your chats are price checkers who never buy, the algorithm learns to find more price checkers. Cheaper cost per conversation, fewer sales.

The fix is to tell Meta what happened next.

Step 1: Classify every chat

Before any technical setup, decide on a small set of outcomes and tag every conversation with one of them. Five is usually enough:

Buyer. Placed an order or booked.

Hot lead. Asked about a specific product, price or slot, showed intent, did not buy yet.

Price checker. Asked the price, nothing else, left.

Support. An existing customer with a question about an order.

Not a fit. Wrong area, wrong product, supplier, job seeker, spam.

Each tag maps to a different action. That is the whole point:

TagWhat to do with it in Meta
BuyerSend a purchase event. Use as a lookalike source. Exclude from acquisition ads.
Hot leadSend a lead event. Retarget with the product they asked about.
Price checkerRetarget with an offer or bundle, not the same ad again.
SupportExclude from acquisition ads.
Not a fitExclude. If this group is large, your targeting is off.

Doing this by hand works for 20 chats a day. At 200 it stops happening, which is why it is worth automating.

Step 2: Send the outcomes back to Meta

Meta has three routes for getting chat outcomes back into Ads Manager.

Labels in the WhatsApp Business app. If you use the free app, you can turn on sharing of customer activity with Meta and use the app's predefined order labels. Chats from click to WhatsApp ads labelled as an order or payment count as purchase events. Custom labels may not be attributed.

Automatic events. On the WhatsApp Business Platform, a business can opt in to automatic events, where Meta detects a lead or purchase in a chat that came from a click to WhatsApp ad and reports it as LeadSubmitted or Purchase.

Conversions API for business messaging. The most precise route. When a customer arrives from a click to WhatsApp ad, the incoming message carries a click ID (ctwa_clid). When that chat turns into an order, your system sends a purchase event to Meta with that click ID, so the sale is tied to the exact ad. If the click ID is missing, hashed customer details can be used to match instead.

Once purchases flow back, a new option opens. Meta's "Maximise number of purchases through messaging" goal for click to WhatsApp ads needs at least 10 purchase events shared in the previous 30 days, from chats that started on a click to WhatsApp ad and were marked as a purchase within 7 days of the click. Then Meta optimizes for buyers, not chatters.

Step 3: Build audiences from your tags

Events teach the algorithm. Audiences let you target directly.

Meta's customer list custom audiences match your customers by identifiers like phone number, which every WhatsApp chat already has. Build one list per tag:

  • Buyers as the source for a lookalike audience, and as an exclusion on acquisition campaigns, so you stop paying to reach people who already bought.
  • Hot leads as a retargeting audience with the product or service they asked about.
  • Price checkers as a retargeting audience for a bundle, an offer or a reason to buy now.
  • Support and not a fit as exclusions.

Two details matter. Meta marks a customer list as outdated if it hasn't been updated in more than 30 days, which can affect delivery, so refresh lists at least monthly, ideally automatically. And a list created as "only exclusions" can't be used for a lookalike later, so keep your buyer list as a normal list.

If you want to message these people on WhatsApp later with marketing templates, you need their marketing opt-in. Retargeting someone with an ad is not the same as messaging them first. Our guide on customer data safety covers what to ask before data leaves your inbox.

What to measure

  • Cost per purchase from click to WhatsApp ads, not just cost per conversation
  • Share of chats by tag. If "not a fit" grows, fix targeting before raising budget
  • Hot lead to buyer rate after retargeting
  • Purchase events sent per month, to reach and keep the 10 needed for purchase optimization

Where plumcut fits

Classifying chats, sending events and refreshing audiences is the part most teams never get to, because it means reading every conversation. That is exactly what plumcut does.

plumcut builds and runs an intelligent automation system on WhatsApp and your other channels. It answers and sells in the chat, classifies every conversation as it happens, and turns that into signals your ads can use: who bought, who is close, who to stop paying for. Your WhatsApp becomes an extra sales funnel that also makes your Meta ads smarter.

If you run ads that open WhatsApp and want them to find buyers instead of chatters, talk to us.

Sources

Checked on 25 October 2026. Meta updates these features often, so confirm current requirements before you set up a campaign.

Don't take our word for it

Have a question? Ask plum.
See it for yourself

Running ads that open WhatsApp? Ask plum how it classifies every chat and turns it into signals your Meta ads can use.