The problem this solves
Retargeting everyone in the CRM with the same ad ignores the one thing the CRM knows: where each person is in the journey. An MQL needs proof; an SQL needs a reason to take the meeting; an open opportunity needs the objection handled. One audience per lifecycle stage lets the campaign speak to each group differently, and the CRM already holds the segmentation.
The manual version is an export per stage, a hashing step, an upload per audience and a note somewhere about which is which - repeated whenever the stages change. Because the work is tedious, it is done once and the audiences drift. Creating the audiences directly from HubSpot on the connected Meta account, with the sizes reported and a targeting note per audience, makes the segmentation live and the campaign structure obvious.
How the mission runs
- Read HubSpot contacts by lifecycle stage. The Audience Agent reads HubSpot contacts and groups them by lifecycle stage - MQL, SQL, opportunity - keeping the identifiers Meta matches on and dropping contacts marked do-not-contact or unsubscribed from marketing.
- Suppress customers and closed-lost where appropriate. Current customers are removed from the acquisition stages so they are not retargeted with prospect messaging, and closed-lost contacts are kept as a separate optional audience for winback if you want one.
- Stage one custom audience per stage. Each stage is staged as a Meta custom audience on your connected ad account, named with the stage and the date, hashed as Meta expects, and shown to you with counts before creation.
- Create, report sizes and write the targeting note. On approval the audiences are created, and once Meta reports the matched sizes, a Google Doc is written with one section per audience: what stage it holds, its size, the message that stage needs, and the suggested exclusions between audiences so an SQL never sees the MQL ad.
The prompt
This is the exact objective the agent receives. Swap the obvious placeholders for your own domain, segment or channel and run it as-is from the console, Slack, or the API.
What comes back
One Meta custom audience per lifecycle stage on your connected ad account - MQL, SQL and opportunity, plus an optional closed-lost audience - with matched sizes reported, and a Google Doc targeting note describing each audience, the message it should receive and the exclusions between them. HubSpot is read but never changed.
Make it yours
- Add a customer audience for expansion campaigns and exclude it from every other audience.
- Split the opportunity stage by deal size for two different messages.
- Refresh weekly so contacts move between audiences as their stage changes.
- Create the same stages as LinkedIn contact-matched audiences for a second channel.
- Add the deal owner as a field so a rep can be alerted when their opportunity contacts engage with the campaign.
Frequently asked questions
Are contacts hashed before they reach Meta?
Yes. Identifiers are hashed the way Meta specifies before upload, and only the fields needed for matching are used.
What happens to contacts who opted out of marketing?
They are excluded from every audience, and the report shows how many were removed for that reason.
Can the audiences come from Pipedrive or Attio instead?
Yes. Name the CRM and the stage field to use, and the same mission runs against it.
How large does each audience need to be?
Meta requires a minimum matched size before an audience can serve; the report flags any stage that falls below it and suggests merging adjacent stages so the campaign can still run.
Can the audiences drive exclusions automatically?
The targeting note lists the exclusions to set - opportunity excluded from MQL, customers from everything - and you apply them in the campaign; the audiences themselves are created ready to be referenced.