DEEP DIVE · HALF-BUILT

Warm Ad Audiences

Most of the signals got into a Clay audience. What never shipped was the second layer: if someone in that audience was a CMO or VP of Marketing, they shouldn't only see ads — they should get resequenced into a human-activated sequence and put in front of a rep. Ads for everyone warm enough to enroll; outbound activation for the titles that matter most.

audience layer mostly built sequencing layer unfinished Clay Audiences Clay Ads HubSpot Lemlist / EmailBison LinkedIn + Meta
[ Who Gets In ]

Not just MQLs. Anyone we've already proven is worth the spend.

The audience was never supposed to be "every MQL." MQLs are a lifecycle label. This was an engagement + ICP membership list that Clay would keep current as HubSpot and the outbound tools changed underneath it.

Anyone who hit one of these would land in the Clay audience (and get removed when they stopped qualifying — customers, open opps, bad fit, unsubscribes):

Enrollment triggers (mostly piped)
Replied to an EmailBison or Lemlist message · Moved into a warm HubSpot lifecycle stage · Closed-lost deal, still in ICP · Existing ICP contact with a verified email and recent engagement · Explicit exclusion for customers, employees, and open opportunities
[ Two Layers ]

Ads for the warm list. A human sequence for the titles that matter.

Clay Audiences was the membership layer: HubSpot + outbound signals in, enrich for match rate, living sync out to LinkedIn / Meta. That part mostly existed — signals were already landing in an audience.

The unfinished half was downstream routing. Membership alone isn't a GTM motion. For senior buyers — CMO, VP of Marketing, equivalent — landing in the ad audience should have also kicked a second path: re-enroll into a human-activated sequence, route to the right rep, and treat the paid impression as air cover while a human actually worked the account.

flowchart TB
  subgraph Signals["Enrollment signals"]
    REPLY[Email / Lemlist reply]
    LIFE[HubSpot lifecycle stage]
    LOST[Closed-lost, still ICP]
  end

  subgraph Layer1["Layer 1 · mostly built"]
    SEG[Clay audience:
warm + ICP + not excluded] ADS[Sync to LinkedIn / Meta
retargeting] end subgraph Layer2["Layer 2 · never built"] TITLE{Title = CMO / VP Marketing
or equivalent?} SEQ[Human-activated sequence
+ rep routing] end REPLY --> SEG LIFE --> SEG LOST --> SEG SEG --> ADS SEG --> TITLE TITLE -->|yes| SEQ TITLE -->|no| ADS style Layer1 stroke:#10b981 style Layer2 stroke:#f59e0b style SEQ stroke:#f59e0b,stroke-width:2px

swipe sideways on mobile →

[ Why Match Rates ]

You're not uploading a scraped list. You're uploading people you've already verified.

Ad platforms match on hashed identity — mostly email, plus name and company. Cold uploaded lists of work emails alone usually match poorly, because most people signed up for LinkedIn and Meta with a personal address, not their work inbox.

The bet was that match rates would beat a typical retargeting upload because every row already had some proof of personhood before it hit Clay Ads: a verified work email from outbound, a CRM record with lifecycle history, often a reply (inbox real and monitored), plus Clay's personal-email / hashed-email-for-ads waterfall for the platform match. Same people marketing wants to retarget — with more identity signal attached.

What never got built
The title split and everything after it: detect CMO / VP Marketing (and peers) in the audience, resequence them into a human-activated sequence, and route to the owning rep. Ads air cover without a human follow-through for the titles that matter was the unfinished half.
"Getting them into the ad audience was layer one. If they were a CMO, layer two should have put a human on them — not just another impression."

Want systems like these at your org?

Contact Me 727-637-2778