Retargeting WhatsApp Campaigns by Outcome (Re-target)

Re-target is an audience mode in the ConvoFly broadcast composer that builds your next campaign's recipient list out of a previous campaign's results — you pick one past broadcast, tick one or more of five outcomes (Replied, Didn't reply, Didn't read, Wasn't delivered, Failed to send), and only the contacts matching those outcomes are messaged. It is the alternative to re-sending to your whole list and hoping the right people notice the second time.

The highest-value audience for your next send is almost always sitting inside your last one. A person who replied "how much?" and never heard back is worth more than 100 cold contacts, and a person whose message failed to deliver never had a chance to respond at all. Re-target separates those groups in about four clicks, with each contact messaged only once even if they match several outcomes.

This guide covers exactly who each outcome selects, how the "Replied" match is computed live from the conversation, the reply-keyword filter, which numbers are permanently excluded, and when an automation is the better tool.

The five outcomes, and precisely who each one selects

Find these in the composer under Audience → Re-target, as a row of chips under the heading "Which contacts", labelled "— pick one or more".

Pick more than one and the logic is a union: a contact is included if they match any outcome you selected, and each person appears only once. The composer states this under the chips: "Anyone matching any of the outcomes you pick is included (each contact once)."

Why "Didn't read" always includes people who read it

This is the one place Re-target can quietly mislead you, and it is a WhatsApp limitation rather than a ConvoFly one. Any recipient can switch read receipts off in their own privacy settings, and when they do their phone never reports a read status back — so a message they read carefully still sits at "delivered" forever. Every one of those people lands in your Didn't read audience.

The practical rule: write a "Didn't read" follow-up so that it causes no harm if the person did read the first message. "In case you missed it" works. "Why haven't you opened my message?" does not. The full explanation of what each delivery status can and cannot prove is in delivery stats and reports.

How the "Replied" match is computed

Replied is not read from a stored snapshot taken when the first campaign finished. It is worked out live, at the moment you build and send the new broadcast, by looking at the actual conversation: has this contact sent you a message dated after the original broadcast went out to them?

That matters in a specific, practical way. If you sent a campaign on Monday and build the follow-up on Thursday, every reply that arrived on Tuesday and Wednesday is included — you are not working from a stale Monday-night list. It also means the audience count you see in the composer can be slightly different from the "Replied" number on the old campaign's report page, and the composer's is the more current of the two.

Matching is done on the last 10 digits of the phone number, so contacts stored with or without a leading country code are still recognised as the same person.

The "Only if their reply contains" filter

Selecting Replied reveals a text box labelled "Only if their reply contains", described as "comma-separated words; leave blank for any reply". It arrives pre-filled with 7 common affirmatives: yes, send, please, price, haan, ji, ok.

The filter matches if any word in your list appears anywhere in the reply, case-insensitively, as part of a longer word or on its own. So "yes" matches "Yes please" and "yesss". Clear the box entirely to include everybody who replied at all, whatever they said.

Three things worth knowing before you rely on it:

• It is a plain word match, not a language model. It cannot tell "yes" from "not yes", so a reply saying "no, don't send" containing the word "send" would match. On a small list, skim the audience before sending.
• Edit the defaults for your market. The pre-filled list mixes English with Urdu/Hindi affirmatives (haan, ji). If your customers write in another language, replace them with words your customers actually use.
• Match your own call to action. If your broadcast said "reply PRICE for the list", the only word that needs to be in the box is price. A narrower list produces a cleaner audience.

Numbers ConvoFly always excludes

Two suppressions apply to every Re-target audience, and neither can be switched off:

• Numbers not on WhatsApp, and invalid numbers. Once a number has failed for either reason in any of your past broadcasts, it is removed from every future audience in your account — including from Failed to send. The composer says so directly: "Numbers that aren't on WhatsApp (or are invalid) are automatically left out — even from Failed to send — so you never re-message an unreachable number."
• Contacts who opted out. Anyone tagged as opted out is excluded from every audience mode, not just Re-target.

This is a deliberate guardrail rather than a limitation. Meta's Business Messaging Policy requires that you message people who want to hear from you, and repeatedly pushing messages at unreachable or unwilling numbers is precisely what erodes a sender's standing (WhatsApp Business Messaging Policy). Since Meta raises your messaging limit only when you send high-quality messages across your numbers and templates (Meta, WhatsApp messaging limits), refusing to retry dead numbers protects your ability to grow.

How to build a re-target broadcast

1. Go to Broadcasts and create a new broadcast as normal.
2. In the audience section, click the Re-target mode button (it sits alongside "By segment", "CSV list" and "Google Sheet").
3. Under Past broadcast, choose the campaign whose results you want to use from the "— Select a broadcast —" dropdown. If you have not sent any broadcasts yet, this section will tell you there is nothing to re-target.
4. Under Which contacts, tick the outcomes you want. Didn't reply is already selected — untick it if you want something else.
5. If you picked Replied, edit or clear the "Only if their reply contains" word list.
6. Choose the template for this new message. It does not have to be the same template as the original campaign, and usually should not be.
7. Check the recipient count, send yourself a test, then send or schedule.

The same Re-target picker is also available when you build a drip campaign, so a past broadcast's outcomes can seed a whole multi-step sequence rather than a single follow-up.

Turning "yes" into a sale: a worked example

You sent "Want our new price list? Reply YES" to 800 contacts. Over the next two days 62 people replied, of whom most said some form of yes and a handful asked something else or said no.

1. Create a new broadcast and choose Re-target.
2. Pick the price-list campaign as the source.
3. Select only Replied, and untick the default Didn't reply.
4. In the keyword box, keep yes, send, please and price, and delete the rest.
5. Choose your price-list template and send.

Only the people who said yes receive it, once each. Everyone who replied "no thanks" is left alone, which is the entire point — and because Replied is computed live, anyone who replied this morning is included too.

For replies that arrive after this send, Re-target is the wrong tool: it is a one-off, manual action, not a standing rule. Set up an automation that responds to every future reply so late responders are handled without you watching the inbox.

Re-target versus an automation versus a segment

Most stores need all three. Use a segment to send the first campaign, an automation to handle replies as they land, and Re-target a day or two later to pick up whoever the first two missed.

Why this beats re-blasting everyone

• It protects your quality rating. The people most likely to block you are the ones who already saw the message and ignored it deliberately. Re-target lets you avoid re-sending to people who read it and chose not to act, which whole-list re-sends cannot.
• It respects your daily cap. Messaging limits count unique recipients per rolling 24 hours, starting at 250 for a new number and stepping to 2,000. A 200-person re-target fits comfortably inside a cap that an 800-person re-blast would consume entirely — see messaging limits and tier caps.
• It costs less. Under Meta's per-message pricing you are charged on every delivered template message, so sending to 200 relevant people instead of 800 indiscriminate ones costs roughly a quarter as much (Meta, WhatsApp pricing).
• It reads as relevant. "Still interested?" to someone who never replied is a reasonable message. The same words to someone who bought yesterday are not.

What Re-target cannot do

Being clear about the edges saves you planning a campaign that will not work:

• One source campaign at a time. You pick a single past broadcast, not a range of dates or a combination of several campaigns.
• No click outcome. The five outcomes cover replies and delivery only. Even with click tracking enabled, "people who clicked" is not a re-targetable audience.
• It is a snapshot at send time, not a standing rule. Anyone who replies after you send the follow-up is not included; that is what automations are for.
• Permanently unreachable numbers stay excluded. Even ticking "Failed to send" will not bring back a number flagged as not-on-WhatsApp or invalid.

A simple two-step play to start with

1. Send your campaign to a segment as normal.
2. Wait 24 to 48 hours so replies and delivery statuses settle.
3. Create a new broadcast, choose Re-target, pick that campaign, keep Didn't reply selected, and send a follow-up with a genuinely different angle — a deadline, an objection answered, or a smaller ask.

Do not send a third. Two touches on a non-responder is persistence; three is the reason people block business numbers. That exact scenario has its own step-by-step walkthrough in how to follow up with people who didn't reply.